diff --git "a/wandb/run-20240514_011900-d0zhpk4t/files/output.log" "b/wandb/run-20240514_011900-d0zhpk4t/files/output.log"
deleted file mode 100644--- "a/wandb/run-20240514_011900-d0zhpk4t/files/output.log"
+++ /dev/null
@@ -1,21858 +0,0 @@
-Overriding model.yaml nc=80 with nc=10
-WARNING ⚠️ no model scale passed. Assuming scale='x'.
-Transferred 1226/1241 items from pretrained weights
-
-
-train: Scanning /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/labels... 6471 images, 0 backgrounds, 0 corrupt: 100%|██████████| 6471/6471 [00:06<00:00, 1003.56it/s]
-val: Scanning /kaggle/working/datasets/VisDrone/VisDrone2019-DET-val/labels... 90 images, 0 backgrounds, 0 corrupt:  16%|█▋        | 90/548 [00:00<00:00, 893.89it/s]
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/0000137_02220_d_0000163.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/0000140_00118_d_0000002.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/9999945_00000_d_0000114.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/9999987_00000_d_0000049.jpg: 1 duplicate labels removed
-train: New cache created: /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/labels.cache
-albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8))
-
-val: Scanning /kaggle/working/datasets/VisDrone/VisDrone2019-DET-val/labels... 548 images, 0 backgrounds, 0 corrupt: 100%|██████████| 548/548 [00:00<00:00, 951.20it/s]
-Plotting labels to runs/detect/RTDETR/labels.jpg...
-optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically...
-optimizer: SGD(lr=0.01, momentum=0.9) with parameter groups 193 weight(decay=0.0), 256 weight(decay=0.0005), 276 bias(decay=0.0)
-  0%|          | 0/809 [00:00<?, ?it/s]
-TensorBoard: model graph visualization added ✅
-Image sizes 640 train, 640 val
-Using 4 dataloader workers
-Logging results to runs/detect/RTDETR
-Starting training for 300 epochs...
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      1/300      14.7G      1.787     0.7231     0.5982        445        640: 100%|██████████| 809/809 [16:42<00:00,  1.24s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:18<00:00,  3.87it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:19<00:00,  3.52it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      2/300      13.5G      1.597     0.4111     0.3238        490        640: 100%|██████████| 809/809 [16:22<00:00,  1.21s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.80it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.86it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      3/300        14G      1.638     0.3526     0.3015        104        640: 100%|██████████| 809/809 [16:18<00:00,  1.21s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-                   all        548      38759      0.564     0.0731     0.0745     0.0327
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-      4/300      13.7G      1.842     0.2821      0.311        331        640:   0%|          | 1/809 [00:01<16:14,  1.21s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      4/300      13.7G      1.631     0.3374     0.2848        100        640: 100%|██████████| 809/809 [16:18<00:00,  1.21s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.86it/s]
-                   all        548      38759      0.437      0.103     0.0834     0.0331
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      5/300      13.6G      1.601     0.3361     0.2621        545        640: 100%|██████████| 809/809 [16:14<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]
-                   all        548      38759       0.49      0.121        0.1     0.0398
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      6/300      13.7G      1.582     0.3385     0.2455        158        640: 100%|██████████| 809/809 [16:14<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.79it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      7/300      13.7G      1.519     0.3526     0.2223        440        640: 100%|██████████| 809/809 [16:12<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.96it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      8/300      13.5G      1.482     0.3622     0.2125        466        640: 100%|██████████| 809/809 [16:12<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.87it/s]
-                   all        548      38759      0.584      0.146      0.153     0.0731
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-      9/300      14.2G      1.461     0.3676     0.2068        223        640: 100%|██████████| 809/809 [16:13<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-                   all        548      38759       0.53       0.18      0.192     0.0951
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     10/300      13.3G      1.397     0.3962     0.1931        353        640: 100%|██████████| 809/809 [16:12<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-                   all        548      38759      0.352      0.165      0.168     0.0774
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     11/300      13.8G      1.358     0.4101     0.1878        210        640: 100%|██████████| 809/809 [16:12<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.92it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     12/300      13.9G      1.353     0.4073     0.1926        102        640: 100%|██████████| 809/809 [16:13<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.99it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     13/300      13.8G      1.345     0.4095     0.1801        206        640: 100%|██████████| 809/809 [16:07<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.94it/s]
-                   all        548      38759      0.528      0.186      0.198     0.0931
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-     14/300      13.6G       1.18     0.4311     0.1195        183        640:   0%|          | 1/809 [00:01<15:48,  1.17s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     14/300      13.6G      1.281     0.4276     0.1728        112        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.00it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.95it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     15/300      13.5G      1.323     0.4074     0.1774        161        640: 100%|██████████| 809/809 [16:08<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]
-                   all        548      38759      0.495      0.214       0.23      0.116
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     16/300      13.6G      1.331     0.4066     0.1761        174        640: 100%|██████████| 809/809 [16:07<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  4.02it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     17/300      13.5G       1.28     0.4216     0.1696        316        640: 100%|██████████| 809/809 [16:03<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:01,  3.96it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     18/300      13.6G       1.26     0.4248      0.163        349        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.97it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.94it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     19/300      13.8G      1.292     0.4135     0.1681         87        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:16<00:00,  3.99it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.94it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     20/300      14.8G      1.264     0.4223     0.1654        129        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]
-                   all        548      38759      0.436      0.257      0.256      0.129
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     21/300      13.7G       1.29     0.4161     0.1667        306        640: 100%|██████████| 809/809 [16:07<00:00,  1.20s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:17<00:00,  3.98it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     22/300      13.5G      1.241     0.4279     0.1578        187        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.94it/s]
-                   all        548      38759      0.412      0.263      0.262      0.129
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     23/300      13.7G      1.257     0.4226     0.1549        299        640: 100%|██████████| 809/809 [16:06<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.95it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     24/300      13.7G      1.237      0.433     0.1539        271        640: 100%|██████████| 809/809 [16:03<00:00,  1.19s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.94it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-     25/300      13.8G      1.119     0.4936     0.1346        338        640:   0%|          | 3/809 [00:03<15:56,  1.19s/it]
-     25/300      13.8G      1.327     0.4231     0.1919        497        640:   1%|          | 5/809 [00:06<16:33,  1.24s/it]
-     25/300      13.8G       1.28     0.4407     0.1762        166        640:   1%|          | 6/809 [00:07<16:10,  1.21s/it]
-     25/300      13.8G      1.246     0.4477     0.1613        162        640:   1%|          | 8/809 [00:09<15:51,  1.19s/it]
-     25/300      13.8G        1.2     0.4479     0.1536        347        640:   1%|          | 10/809 [00:12<16:12,  1.22s/it]
-     25/300      13.8G      1.216     0.4426     0.1574        171        640:   1%|▏         | 11/809 [00:13<16:13,  1.22s/it]
-     25/300      13.8G      1.198      0.445     0.1548        164        640:   2%|▏         | 13/809 [00:15<15:52,  1.20s/it]
-     25/300      13.8G      1.212      0.436      0.152        415        640:   2%|▏         | 15/809 [00:18<15:57,  1.21s/it]
-     25/300      13.8G      1.213     0.4358     0.1496        259        640:   2%|▏         | 16/809 [00:19<15:49,  1.20s/it]
-     25/300      13.8G       1.21      0.435     0.1491        132        640:   2%|▏         | 18/809 [00:21<15:47,  1.20s/it]
-     25/300      13.8G      1.201     0.4379     0.1455        197        640:   2%|▏         | 20/809 [00:24<15:42,  1.19s/it]
-     25/300      13.8G      1.198     0.4359     0.1439        179        640:   3%|▎         | 22/809 [00:26<15:32,  1.18s/it]
-     25/300      13.8G      1.193     0.4366     0.1429        158        640:   3%|▎         | 23/809 [00:27<15:24,  1.18s/it]
-     25/300      13.8G      1.198     0.4355     0.1438        357        640:   3%|▎         | 25/809 [00:30<15:46,  1.21s/it]
-     25/300      13.8G      1.171     0.4449     0.1406        126        640:   3%|▎         | 27/809 [00:32<15:40,  1.20s/it]
-     25/300      13.8G      1.183      0.442     0.1428        410        640:   3%|▎         | 28/809 [00:33<15:32,  1.19s/it]
-     25/300      13.8G      1.172     0.4432     0.1436        489        640:   4%|▎         | 30/809 [00:36<15:37,  1.20s/it]
-     25/300      13.8G      1.184      0.442     0.1443        330        640:   4%|▍         | 32/809 [00:38<15:32,  1.20s/it]
-     25/300      13.8G      1.205     0.4358     0.1518        655        640:   4%|▍         | 33/809 [00:39<16:27,  1.27s/it]
-     25/300      13.8G      1.188     0.4407      0.149        246        640:   4%|▍         | 35/809 [00:42<15:41,  1.22s/it]
-     25/300      13.8G      1.194     0.4388     0.1498        438        640:   4%|▍         | 36/809 [00:43<15:52,  1.23s/it]
-     25/300      13.8G      1.183     0.4417     0.1476        283        640:   5%|▍         | 38/809 [00:45<15:30,  1.21s/it]
-     25/300      13.8G      1.174      0.443     0.1448        234        640:   5%|▍         | 40/809 [00:48<15:10,  1.18s/it]
-     25/300      13.8G      1.163     0.4479     0.1428        312        640:   5%|▌         | 42/809 [00:50<15:10,  1.19s/it]
-     25/300      13.8G      1.167     0.4471     0.1435        388        640:   5%|▌         | 43/809 [00:51<15:09,  1.19s/it]
-     25/300      13.8G      1.162     0.4482     0.1421        155        640:   6%|▌         | 45/809 [00:54<15:30,  1.22s/it]
-     25/300      13.8G      1.161     0.4481     0.1417        277        640:   6%|▌         | 46/809 [00:55<15:17,  1.20s/it]
-     25/300      13.8G       1.17     0.4451     0.1413        350        640:   6%|▌         | 48/809 [00:57<15:05,  1.19s/it]
-     25/300      13.8G      1.169     0.4446     0.1413        362        640:   6%|▌         | 50/809 [01:00<15:05,  1.19s/it]
-     25/300      13.8G      1.155     0.4477     0.1399        129        640:   6%|▋         | 52/809 [01:02<14:48,  1.17s/it]
-     25/300      13.8G      1.147     0.4495     0.1388        135        640:   7%|▋         | 53/809 [01:03<14:39,  1.16s/it]
-     25/300      13.8G      1.146     0.4504     0.1392        235        640:   7%|▋         | 55/809 [01:05<14:44,  1.17s/it]
-     25/300      13.8G      1.145     0.4495     0.1388        135        640:   7%|▋         | 57/809 [01:08<15:02,  1.20s/it]
-     25/300      13.8G      1.152     0.4475     0.1396        445        640:   7%|▋         | 58/809 [01:09<14:55,  1.19s/it]
-     25/300      13.8G      1.164     0.4431     0.1404        291        640:   7%|▋         | 60/809 [01:12<14:58,  1.20s/it]
-     25/300      13.8G      1.159      0.445      0.141         91        640:   8%|▊         | 62/809 [01:14<14:36,  1.17s/it]
-     25/300      13.8G      1.155     0.4456     0.1402        172        640:   8%|▊         | 63/809 [01:15<14:35,  1.17s/it]
-     25/300      13.8G      1.153     0.4469     0.1394         86        640:   8%|▊         | 65/809 [01:17<15:01,  1.21s/it]
-     25/300      13.8G      1.161     0.4448     0.1413        564        640:   8%|▊         | 67/809 [01:20<14:56,  1.21s/it]
-     25/300      13.8G      1.166     0.4432     0.1426        296        640:   8%|▊         | 68/809 [01:21<14:48,  1.20s/it]
-     25/300      13.8G      1.175     0.4403      0.143        385        640:   9%|▊         | 70/809 [01:23<14:49,  1.20s/it]
-     25/300      13.8G      1.177     0.4399     0.1421        220        640:   9%|▉         | 72/809 [01:26<14:35,  1.19s/it]
-     25/300      13.8G      1.178     0.4393     0.1418        134        640:   9%|▉         | 73/809 [01:27<14:42,  1.20s/it]
-     25/300      13.8G      1.181     0.4393     0.1422        220        640:   9%|▉         | 75/809 [01:29<14:24,  1.18s/it]
-     25/300      13.8G      1.181     0.4408     0.1417        306        640:  10%|▉         | 77/809 [01:32<14:14,  1.17s/it]
-     25/300      13.8G      1.176     0.4438     0.1402        110        640:  10%|▉         | 79/809 [01:34<14:34,  1.20s/it]
-     25/300      13.8G      1.179     0.4425     0.1412        321        640:  10%|▉         | 80/809 [01:35<14:40,  1.21s/it]
-     25/300      13.8G      1.181     0.4424     0.1406        220        640:  10%|█         | 82/809 [01:38<14:36,  1.20s/it]
-     25/300      13.8G      1.186     0.4408     0.1419        500        640:  10%|█         | 83/809 [01:39<14:40,  1.21s/it]
-     25/300      13.8G      1.193     0.4383     0.1433        215        640:  11%|█         | 85/809 [01:41<14:30,  1.20s/it]
-     25/300      13.8G       1.19     0.4386     0.1425        192        640:  11%|█         | 87/809 [01:44<14:10,  1.18s/it]
-     25/300      13.8G      1.188     0.4394     0.1419        129        640:  11%|█         | 89/809 [01:46<14:22,  1.20s/it]
-     25/300      13.8G      1.191     0.4385     0.1421        308        640:  11%|█         | 90/809 [01:47<14:19,  1.20s/it]
-     25/300      13.8G      1.192     0.4382     0.1417        360        640:  11%|█▏        | 92/809 [01:50<14:12,  1.19s/it]
-     25/300      13.8G      1.188     0.4393     0.1409        244        640:  12%|█▏        | 94/809 [01:52<14:01,  1.18s/it]
-     25/300      13.8G      1.188     0.4395      0.141        285        640:  12%|█▏        | 95/809 [01:53<13:57,  1.17s/it]
-     25/300      13.8G      1.186     0.4402     0.1407        218        640:  12%|█▏        | 97/809 [01:56<14:09,  1.19s/it]
-     25/300      13.8G      1.184     0.4404     0.1401        315        640:  12%|█▏        | 99/809 [01:58<13:55,  1.18s/it]
-     25/300      13.8G      1.181     0.4417     0.1399        236        640:  12%|█▏        | 101/809 [02:00<13:45,  1.17s/it]
-     25/300      13.8G       1.18     0.4422     0.1395        225        640:  13%|█▎        | 102/809 [02:01<13:42,  1.16s/it]
-     25/300      13.8G      1.183     0.4409     0.1404        185        640:  13%|█▎        | 104/809 [02:04<13:44,  1.17s/it]
-     25/300      13.8G      1.182     0.4411     0.1403        109        640:  13%|█▎        | 106/809 [02:06<13:56,  1.19s/it]
-     25/300      13.8G      1.183      0.441     0.1402        286        640:  13%|█▎        | 107/809 [02:07<13:48,  1.18s/it]
-     25/300      13.8G      1.182     0.4418     0.1406        390        640:  13%|█▎        | 109/809 [02:10<13:49,  1.19s/it]
-     25/300      13.8G      1.181     0.4421     0.1403        234        640:  14%|█▎        | 111/809 [02:12<13:36,  1.17s/it]
-     25/300      13.8G      1.179      0.443     0.1407        166        640:  14%|█▍        | 112/809 [02:13<13:32,  1.17s/it]
-     25/300      13.8G      1.181     0.4427      0.141        287        640:  14%|█▍        | 114/809 [02:16<13:34,  1.17s/it]
-     25/300      13.8G      1.183     0.4415     0.1407        250        640:  14%|█▍        | 116/809 [02:18<13:34,  1.17s/it]
-     25/300      13.8G      1.186     0.4405     0.1407        302        640:  15%|█▍        | 118/809 [02:20<13:38,  1.19s/it]
-     25/300      13.8G      1.188     0.4395     0.1406        227        640:  15%|█▍        | 119/809 [02:22<13:58,  1.21s/it]
-     25/300      13.8G      1.187     0.4403     0.1406        117        640:  15%|█▍        | 121/809 [02:24<13:52,  1.21s/it]
-     25/300      13.8G      1.183     0.4415       0.14        207        640:  15%|█▌        | 123/809 [02:26<13:29,  1.18s/it]
-     25/300      13.8G      1.181     0.4416     0.1397        208        640:  15%|█▌        | 124/809 [02:27<13:33,  1.19s/it]
-     25/300      13.8G      1.181     0.4423     0.1397        206        640:  16%|█▌        | 126/809 [02:30<13:22,  1.18s/it]
-     25/300      13.8G      1.178     0.4441      0.139        100        640:  16%|█▌        | 128/809 [02:32<13:10,  1.16s/it]
-     25/300      13.8G      1.175     0.4448     0.1386        129        640:  16%|█▌        | 129/809 [02:33<13:18,  1.17s/it]
-     25/300      13.8G      1.175     0.4447     0.1383        339        640:  16%|█▌        | 131/809 [02:36<13:29,  1.19s/it]
-     25/300      13.8G      1.177     0.4441     0.1384        300        640:  16%|█▋        | 133/809 [02:38<13:11,  1.17s/it]
-     25/300      13.8G      1.181      0.443     0.1406        532        640:  17%|█▋        | 134/809 [02:39<13:39,  1.21s/it]
-     25/300      13.8G      1.182     0.4426     0.1404        267        640:  17%|█▋        | 136/809 [02:42<13:20,  1.19s/it]
-     25/300      13.8G      1.181     0.4428     0.1403        264        640:  17%|█▋        | 138/809 [02:44<13:15,  1.19s/it]
-     25/300      13.8G      1.181     0.4427     0.1402        217        640:  17%|█▋        | 139/809 [02:45<13:12,  1.18s/it]
-     25/300      13.8G       1.18     0.4427       0.14        221        640:  17%|█▋        | 141/809 [02:48<13:02,  1.17s/it]
-     25/300      13.8G      1.179     0.4426     0.1395        155        640:  18%|█▊        | 143/809 [02:50<12:59,  1.17s/it]
-     25/300      13.8G      1.176     0.4439     0.1389        237        640:  18%|█▊        | 145/809 [02:52<13:17,  1.20s/it]
-     25/300      13.8G      1.175      0.444     0.1386        259        640:  18%|█▊        | 146/809 [02:53<13:06,  1.19s/it]
-     25/300      13.8G       1.18     0.4422     0.1395        475        640:  18%|█▊        | 148/809 [02:56<13:10,  1.20s/it]
-     25/300      13.8G      1.179     0.4431      0.139         90        640:  19%|█▊        | 150/809 [02:58<12:55,  1.18s/it]
-     25/300      13.8G      1.177     0.4435     0.1387        128        640:  19%|█▊        | 151/809 [02:59<12:48,  1.17s/it]
-     25/300      13.8G      1.178     0.4433     0.1392        243        640:  19%|█▉        | 153/809 [03:02<13:02,  1.19s/it]
-     25/300      13.8G      1.177      0.443     0.1388        190        640:  19%|█▉        | 155/809 [03:04<12:50,  1.18s/it]
-     25/300      13.8G       1.18     0.4423     0.1401        455        640:  19%|█▉        | 156/809 [03:05<13:02,  1.20s/it]
-     25/300      13.8G      1.184     0.4413     0.1402        309        640:  20%|█▉        | 158/809 [03:08<12:58,  1.20s/it]
-     25/300      13.8G      1.184      0.441     0.1403        315        640:  20%|█▉        | 160/809 [03:10<12:49,  1.18s/it]
-     25/300      13.8G      1.185     0.4408     0.1404        207        640:  20%|█▉        | 161/809 [03:11<13:06,  1.21s/it]
-     25/300      13.8G      1.189     0.4395     0.1414        216        640:  20%|██        | 163/809 [03:14<12:53,  1.20s/it]
-     25/300      13.8G       1.19      0.439     0.1412        172        640:  20%|██        | 165/809 [03:16<13:02,  1.22s/it]
-     25/300      13.8G      1.189     0.4396     0.1409         79        640:  21%|██        | 166/809 [03:17<12:50,  1.20s/it]
-     25/300      13.8G      1.194     0.4379     0.1414        273        640:  21%|██        | 168/809 [03:20<12:46,  1.20s/it]
-     25/300      13.8G      1.198     0.4367     0.1421        267        640:  21%|██        | 170/809 [03:22<12:51,  1.21s/it]
-     25/300      13.8G        1.2      0.436     0.1426        478        640:  21%|██        | 171/809 [03:23<12:59,  1.22s/it]
-     25/300      13.8G      1.201     0.4354     0.1423        252        640:  21%|██▏       | 173/809 [03:26<12:51,  1.21s/it]
-     25/300      13.8G      1.204      0.434      0.143        656        640:  22%|██▏       | 175/809 [03:28<12:55,  1.22s/it]
-     25/300      13.8G      1.205     0.4336     0.1429        295        640:  22%|██▏       | 176/809 [03:29<12:46,  1.21s/it]
-     25/300      13.8G      1.205     0.4337      0.143        196        640:  22%|██▏       | 178/809 [03:32<12:38,  1.20s/it]
-     25/300      13.8G      1.206     0.4337     0.1434        342        640:  22%|██▏       | 180/809 [03:34<12:38,  1.21s/it]
-     25/300      13.8G      1.205     0.4339     0.1432        230        640:  22%|██▏       | 181/809 [03:35<12:28,  1.19s/it]
-     25/300      13.8G      1.206     0.4333     0.1432        252        640:  23%|██▎       | 183/809 [03:38<12:27,  1.19s/it]
-     25/300      13.8G      1.205     0.4334      0.143        110        640:  23%|██▎       | 185/809 [03:40<12:32,  1.21s/it]
-     25/300      13.8G      1.204     0.4342     0.1427        160        640:  23%|██▎       | 186/809 [03:41<12:18,  1.19s/it]
-     25/300      13.8G      1.205     0.4339     0.1435        231        640:  23%|██▎       | 188/809 [03:44<12:19,  1.19s/it]
-     25/300      13.8G      1.206     0.4334     0.1436        276        640:  23%|██▎       | 190/809 [03:46<12:13,  1.18s/it]
-     25/300      13.8G      1.204     0.4337     0.1435        230        640:  24%|██▎       | 192/809 [03:49<12:07,  1.18s/it]
-     25/300      13.8G      1.205     0.4336     0.1435        252        640:  24%|██▍       | 193/809 [03:50<12:22,  1.20s/it]
-     25/300      13.8G      1.204     0.4338     0.1434        223        640:  24%|██▍       | 195/809 [03:52<12:06,  1.18s/it]
-     25/300      13.8G      1.207     0.4329     0.1441        522        640:  24%|██▍       | 196/809 [03:53<12:17,  1.20s/it]
-     25/300      13.8G      1.209     0.4324      0.145        230        640:  24%|██▍       | 198/809 [03:56<12:17,  1.21s/it]
-     25/300      13.8G      1.208     0.4326     0.1448        266        640:  25%|██▍       | 200/809 [03:58<12:11,  1.20s/it]
-     25/300      13.8G      1.207     0.4325     0.1446        148        640:  25%|██▍       | 202/809 [04:01<12:03,  1.19s/it]
-     25/300      13.8G      1.206     0.4331     0.1445        119        640:  25%|██▌       | 203/809 [04:02<11:55,  1.18s/it]
-     25/300      13.8G      1.206      0.433     0.1445        279        640:  25%|██▌       | 205/809 [04:04<11:48,  1.17s/it]
-     25/300      13.8G      1.209     0.4324     0.1456        326        640:  26%|██▌       | 207/809 [04:07<12:04,  1.20s/it]
-     25/300      13.8G      1.208     0.4326     0.1455        256        640:  26%|██▌       | 208/809 [04:08<11:57,  1.19s/it]
-     25/300      13.8G      1.207     0.4331     0.1454        121        640:  26%|██▌       | 210/809 [04:10<11:44,  1.18s/it]
-     25/300      13.8G      1.207     0.4329     0.1452        275        640:  26%|██▌       | 212/809 [04:12<11:39,  1.17s/it]
-     25/300      13.8G      1.208     0.4326     0.1456        329        640:  26%|██▋       | 213/809 [04:14<11:41,  1.18s/it]
-     25/300      13.8G      1.208     0.4327     0.1455        319        640:  27%|██▋       | 215/809 [04:16<11:35,  1.17s/it]
-     25/300      13.8G       1.21      0.432     0.1461        647        640:  27%|██▋       | 217/809 [04:18<12:04,  1.22s/it]
-     25/300      13.8G      1.208     0.4325     0.1459        198        640:  27%|██▋       | 218/809 [04:20<11:54,  1.21s/it]
-     25/300      13.8G      1.208     0.4324     0.1458        198        640:  27%|██▋       | 220/809 [04:22<11:41,  1.19s/it]
-     25/300      13.8G      1.208     0.4321     0.1456        290        640:  27%|██▋       | 222/809 [04:24<11:31,  1.18s/it]
-     25/300      13.8G      1.208     0.4323     0.1455        284        640:  28%|██▊       | 224/809 [04:27<11:31,  1.18s/it]
-     25/300      13.8G      1.206      0.433     0.1453        123        640:  28%|██▊       | 225/809 [04:28<11:49,  1.21s/it]
-     25/300      13.8G      1.207     0.4325     0.1455        194        640:  28%|██▊       | 227/809 [04:30<11:29,  1.19s/it]
-     25/300      13.8G       1.21     0.4315     0.1461        224        640:  28%|██▊       | 229/809 [04:33<11:24,  1.18s/it]
-     25/300      13.8G      1.211     0.4311      0.146        201        640:  28%|██▊       | 230/809 [04:34<11:18,  1.17s/it]
-     25/300      13.8G       1.21     0.4309     0.1459        179        640:  29%|██▊       | 232/809 [04:36<11:11,  1.16s/it]
-     25/300      13.8G      1.212     0.4298     0.1459        362        640:  29%|██▉       | 234/809 [04:38<11:19,  1.18s/it]
-     25/300      13.8G       1.21     0.4304     0.1457        171        640:  29%|██▉       | 235/809 [04:40<11:10,  1.17s/it]
-     25/300      13.8G       1.21     0.4302     0.1456        351        640:  29%|██▉       | 237/809 [04:42<11:09,  1.17s/it]
-     25/300      13.8G      1.212     0.4294     0.1458        287        640:  30%|██▉       | 239/809 [04:44<11:22,  1.20s/it]
-     25/300      13.8G      1.212     0.4292     0.1459        310        640:  30%|██���       | 240/809 [04:46<11:18,  1.19s/it]
-     25/300      13.8G      1.211     0.4294     0.1458        426        640:  30%|██▉       | 242/809 [04:48<11:19,  1.20s/it]
-     25/300      13.8G      1.208     0.4306     0.1458        141        640:  30%|███       | 244/809 [04:50<11:03,  1.17s/it]
-     25/300      13.8G      1.206     0.4309     0.1454         93        640:  30%|███       | 246/809 [04:53<10:57,  1.17s/it]
-     25/300      13.8G      1.206     0.4308     0.1455        219        640:  31%|███       | 247/809 [04:54<10:53,  1.16s/it]
-     25/300      13.8G      1.204     0.4311     0.1451         53        640:  31%|███       | 249/809 [04:56<10:59,  1.18s/it]
-     25/300      13.8G      1.203     0.4308     0.1448        284        640:  31%|███       | 251/809 [04:58<11:03,  1.19s/it]
-     25/300      13.8G      1.205     0.4303     0.1455        311        640:  31%|███       | 252/809 [05:00<11:12,  1.21s/it]
-     25/300      13.8G      1.203     0.4309     0.1453        128        640:  31%|███▏      | 254/809 [05:02<10:54,  1.18s/it]
-     25/300      13.8G      1.203      0.431     0.1452        280        640:  32%|███▏      | 256/809 [05:04<10:50,  1.18s/it]
-     25/300      13.8G      1.205     0.4307     0.1464        443        640:  32%|███▏      | 257/809 [05:06<11:17,  1.23s/it]
-     25/300      13.8G      1.205     0.4306     0.1465        274        640:  32%|███▏      | 259/809 [05:08<11:06,  1.21s/it]
-     25/300      13.8G      1.206     0.4305     0.1468        348        640:  32%|███▏      | 261/809 [05:10<10:56,  1.20s/it]
-     25/300      13.8G      1.206     0.4302     0.1469        424        640:  32%|███▏      | 262/809 [05:12<10:54,  1.20s/it]
-     25/300      13.8G      1.206     0.4305     0.1469        188        640:  33%|███▎      | 264/809 [05:14<10:45,  1.18s/it]
-     25/300      13.8G      1.208       0.43     0.1475        327        640:  33%|███▎      | 266/809 [05:16<10:52,  1.20s/it]
-     25/300      13.8G      1.211     0.4294     0.1485        571        640:  33%|███▎      | 267/809 [05:18<11:16,  1.25s/it]
-     25/300      13.8G      1.211     0.4297     0.1488        311        640:  33%|███▎      | 269/809 [05:20<10:56,  1.22s/it]
-     25/300      13.8G       1.21       0.43     0.1485        181        640:  33%|███▎      | 271/809 [05:23<10:53,  1.21s/it]
-     25/300      13.8G       1.21     0.4299     0.1485        291        640:  34%|███▎      | 272/809 [05:24<10:58,  1.23s/it]
-     25/300      13.8G       1.21       0.43     0.1485        438        640:  34%|███▍      | 274/809 [05:26<10:48,  1.21s/it]
-     25/300      13.8G       1.21     0.4301     0.1485        151        640:  34%|███▍      | 276/809 [05:29<10:35,  1.19s/it]
-     25/300      13.8G      1.212     0.4295     0.1486        276        640:  34%|███▍      | 277/809 [05:30<10:30,  1.19s/it]
-     25/300      13.8G      1.211     0.4299     0.1483        270        640:  34%|███▍      | 279/809 [05:32<10:39,  1.21s/it]
-     25/300      13.8G      1.213     0.4295     0.1489        135        640:  35%|███▍      | 281/809 [05:35<10:37,  1.21s/it]
-     25/300      13.8G      1.212     0.4297     0.1488         63        640:  35%|███▍      | 282/809 [05:36<10:27,  1.19s/it]
-     25/300      13.8G      1.209     0.4306     0.1484        182        640:  35%|███▌      | 284/809 [05:38<10:23,  1.19s/it]
-     25/300      13.8G      1.208     0.4312      0.148        122        640:  35%|███▌      | 286/809 [05:40<10:09,  1.17s/it]
-     25/300      13.8G      1.207     0.4318     0.1476        202        640:  36%|███▌      | 288/809 [05:43<10:06,  1.16s/it]
-     25/300      13.8G      1.207     0.4316     0.1478        311        640:  36%|███▌      | 289/809 [05:44<10:16,  1.18s/it]
-     25/300      13.8G      1.206     0.4321     0.1475        153        640:  36%|███▌      | 291/809 [05:46<10:04,  1.17s/it]
-     25/300      13.8G      1.205     0.4323     0.1474        257        640:  36%|███▌      | 293/809 [05:49<10:03,  1.17s/it]
-     25/300      13.8G      1.205     0.4322     0.1473        220        640:  36%|███▋      | 294/809 [05:50<10:02,  1.17s/it]
-     25/300      13.8G      1.205     0.4324     0.1472        164        640:  37%|███▋      | 296/809 [05:52<09:59,  1.17s/it]
-     25/300      13.8G      1.203     0.4331      0.147        211        640:  37%|███▋      | 298/809 [05:55<10:04,  1.18s/it]
-     25/300      13.8G      1.202     0.4333     0.1469        180        640:  37%|███▋      | 299/809 [05:56<09:58,  1.17s/it]
-     25/300      13.8G      1.202     0.4332     0.1467        178        640:  37%|███▋      | 301/809 [05:58<09:56,  1.17s/it]
-     25/300      13.8G      1.203     0.4329     0.1468        295        640:  37%|███▋      | 303/809 [06:00<09:58,  1.18s/it]
-     25/300      13.8G      1.204     0.4327     0.1469        504        640:  38%|███▊      | 304/809 [06:02<09:57,  1.18s/it]
-     25/300      13.8G      1.205     0.4326     0.1471        220        640:  38%|███▊      | 306/809 [06:04<10:07,  1.21s/it]
-     25/300      13.8G      1.204     0.4332     0.1468        129        640:  38%|███▊      | 308/809 [06:06<09:49,  1.18s/it]
-     25/300      13.8G      1.203     0.4332     0.1465        196        640:  38%|███▊      | 310/809 [06:09<09:41,  1.16s/it]
-     25/300      13.8G      1.203     0.4332     0.1464        240        640:  38%|███▊      | 311/809 [06:10<09:39,  1.16s/it]
-     25/300      13.8G      1.202     0.4336     0.1462         70        640:  39%|███▊      | 313/809 [06:12<09:46,  1.18s/it]
-     25/300      13.8G      1.203     0.4335     0.1462        255        640:  39%|███▉      | 315/809 [06:15<09:40,  1.18s/it]
-     25/300      13.8G      1.204     0.4331     0.1466        199        640:  39%|███▉      | 316/809 [06:16<09:42,  1.18s/it]
-     25/300      13.8G      1.206     0.4327     0.1472        513        640:  39%|███▉      | 318/809 [06:18<09:52,  1.21s/it]
-     25/300      13.8G      1.207     0.4326     0.1477        508        640:  40%|███▉      | 320/809 [06:21<10:02,  1.23s/it]
-     25/300      13.8G      1.207     0.4329     0.1476        157        640:  40%|███▉      | 321/809 [06:22<10:02,  1.23s/it]
-     25/300      13.8G      1.207     0.4328     0.1474        273        640:  40%|███▉      | 323/809 [06:24<09:40,  1.20s/it]
-     25/300      13.8G       1.21      0.432     0.1477        295        640:  40%|████      | 325/809 [06:27<09:35,  1.19s/it]
-     25/300      13.8G       1.21     0.4318     0.1476        232        640:  40%|████      | 326/809 [06:28<09:31,  1.18s/it]
-     25/300      13.8G       1.21     0.4318     0.1475        268        640:  41%|████      | 328/809 [06:30<09:27,  1.18s/it]
-     25/300      13.8G      1.211     0.4314     0.1473        267        640:  41%|████      | 330/809 [06:33<09:26,  1.18s/it]
-     25/300      13.8G      1.211     0.4315     0.1472        163        640:  41%|████      | 331/809 [06:34<09:19,  1.17s/it]
-     25/300      13.8G       1.21     0.4319      0.147        179        640:  41%|████      | 333/809 [06:36<09:19,  1.18s/it]
-     25/300      13.8G      1.209     0.4323     0.1467        104        640:  41%|████▏     | 335/809 [06:39<09:17,  1.18s/it]
-     25/300      13.8G      1.209     0.4323     0.1466        172        640:  42%|████▏     | 337/809 [06:41<09:21,  1.19s/it]
-     25/300      13.8G      1.209     0.4323     0.1465        243        640:  42%|████▏     | 338/809 [06:42<09:13,  1.18s/it]
-     25/300      13.8G      1.211     0.4317     0.1466        457        640:  42%|████▏     | 340/809 [06:44<09:17,  1.19s/it]
-     25/300      13.8G      1.213     0.4309     0.1471        481        640:  42%|████▏     | 342/809 [06:47<09:19,  1.20s/it]
-     25/300      13.8G      1.214     0.4306     0.1472        365        640:  42%|████▏     | 343/809 [06:48<09:16,  1.19s/it]
-     25/300      13.8G      1.216     0.4298     0.1472        317        640:  43%|████▎     | 345/809 [06:50<09:22,  1.21s/it]
-     25/300      13.8G      1.215     0.4303     0.1471         60        640:  43%|████▎     | 347/809 [06:53<09:09,  1.19s/it]
-     25/300      13.8G      1.214     0.4305     0.1471        199        640:  43%|████▎     | 348/809 [06:54<09:04,  1.18s/it]
-     25/300      13.8G      1.214     0.4307     0.1473        152        640:  43%|████▎     | 350/809 [06:56<08:58,  1.17s/it]
-     25/300      13.8G      1.213     0.4311     0.1472        244        640:  44%|████▎     | 352/809 [06:59<08:50,  1.16s/it]
-     25/300      13.8G      1.214     0.4309     0.1475        430        640:  44%|████▍     | 354/809 [07:01<09:00,  1.19s/it]
-     25/300      13.8G      1.214     0.4309     0.1474        188        640:  44%|████▍     | 355/809 [07:02<08:58,  1.19s/it]
-     25/300      13.8G      1.215     0.4305     0.1475        358        640:  44%|████▍     | 357/809 [07:05<08:58,  1.19s/it]
-     25/300      13.8G      1.214      0.431     0.1473        113        640:  44%|████▍     | 359/809 [07:07<08:48,  1.18s/it]
-     25/300      13.8G      1.213     0.4312     0.1472        130        640:  44%|████▍     | 360/809 [07:08<08:47,  1.18s/it]
-     25/300      13.8G      1.215     0.4307     0.1474        393        640:  45%|████▍     | 362/809 [07:11<08:52,  1.19s/it]
-     25/300      13.8G      1.215     0.4307     0.1473        358        640:  45%|████▍     | 364/809 [07:13<08:45,  1.18s/it]
-     25/300      13.8G      1.215     0.4304     0.1473        240        640:  45%|████▌     | 365/809 [07:14<08:41,  1.18s/it]
-     25/300      13.8G      1.216     0.4303     0.1473        157        640:  45%|████▌     | 367/809 [07:16<08:40,  1.18s/it]
-     25/300      13.8G      1.214     0.4311     0.1471        181        640:  46%|████▌     | 369/809 [07:19<08:38,  1.18s/it]
-     25/300      13.8G      1.213     0.4314      0.147        123        640:  46%|████▌     | 370/809 [07:20<08:36,  1.18s/it]
-     25/300      13.8G      1.216     0.4306     0.1475        396        640:  46%|████▌     | 372/809 [07:22<08:39,  1.19s/it]
-     25/300      13.8G      1.215     0.4312     0.1474        221        640:  46%|████▌     | 374/809 [07:25<08:29,  1.17s/it]
-     25/300      13.8G      1.213     0.4317     0.1473        167        640:  46%|████▋     | 376/809 [07:27<08:27,  1.17s/it]
-     25/300      13.8G      1.215     0.4313     0.1483        460        640:  47%|████▋     | 377/809 [07:28<08:49,  1.22s/it]
-     25/300      13.8G      1.216     0.4309     0.1484        294        640:  47%|████▋     | 379/809 [07:31<08:37,  1.20s/it]
-     25/300      13.8G      1.217     0.4308     0.1492        450        640:  47%|████▋     | 381/809 [07:33<08:39,  1.21s/it]
-     25/300      13.8G      1.217     0.4308     0.1491        375        640:  47%|████▋     | 382/809 [07:34<08:35,  1.21s/it]
-     25/300      13.8G      1.216     0.4311      0.149        229        640:  47%|████▋     | 384/809 [07:37<08:22,  1.18s/it]
-     25/300      13.8G      1.216     0.4309     0.1489        337        640:  48%|████▊     | 386/809 [07:39<08:33,  1.21s/it]
-     25/300      13.8G      1.218     0.4305     0.1494        453        640:  48%|████▊     | 387/809 [07:40<08:37,  1.23s/it]
-     25/300      13.8G      1.218     0.4306     0.1495        201        640:  48%|████▊     | 389/809 [07:43<08:31,  1.22s/it]
-     25/300      13.8G      1.218     0.4305     0.1497        427        640:  48%|████▊     | 390/809 [07:44<08:26,  1.21s/it]
-     25/300      13.8G      1.219     0.4303     0.1498        236        640:  48%|████▊     | 392/809 [07:46<08:20,  1.20s/it]
-     25/300      13.8G      1.219     0.4304     0.1496        131        640:  49%|████▊     | 394/809 [07:49<08:11,  1.19s/it]
-     25/300      13.8G      1.221       0.43     0.1501        192        640:  49%|████▉     | 396/809 [07:51<08:12,  1.19s/it]
-     25/300      13.8G      1.221     0.4299     0.1501        301        640:  49%|████▉     | 397/809 [07:52<08:09,  1.19s/it]
-     25/300      13.8G      1.221       0.43     0.1499        230        640:  49%|████▉     | 399/809 [07:55<08:05,  1.18s/it]
-     25/300      13.8G      1.222     0.4301     0.1502        231        640:  50%|████▉     | 401/809 [07:57<08:06,  1.19s/it]
-     25/300      13.8G      1.222       0.43     0.1502        228        640:  50%|████▉     | 402/809 [07:58<08:02,  1.19s/it]
-     25/300      13.8G      1.223     0.4299     0.1503        367        640:  50%|████▉     | 404/809 [08:01<08:00,  1.19s/it]
-     25/300      13.8G      1.224     0.4296      0.151        353        640:  50%|█████     | 406/809 [08:03<08:02,  1.20s/it]
-     25/300      13.8G      1.224     0.4297     0.1509        198        640:  50%|█████     | 407/809 [08:04<07:58,  1.19s/it]
-     25/300      13.8G      1.225     0.4295     0.1513        305        640:  51%|█████     | 409/809 [08:07<08:03,  1.21s/it]
-     25/300      13.8G      1.224     0.4298     0.1511        116        640:  51%|█████     | 411/809 [08:09<07:51,  1.18s/it]
-     25/300      13.8G      1.224     0.4299      0.151        227        640:  51%|█████     | 412/809 [08:10<07:51,  1.19s/it]
-     25/300      13.8G      1.223     0.4302     0.1509        192        640:  51%|█████     | 414/809 [08:13<07:53,  1.20s/it]
-     25/300      13.8G      1.223     0.4303     0.1509        398        640:  51%|█████▏    | 416/809 [08:15<07:48,  1.19s/it]
-     25/300      13.8G      1.224     0.4299     0.1512        354        640:  52%|█████▏    | 417/809 [08:16<07:56,  1.22s/it]
-     25/300      13.8G      1.226     0.4297     0.1514        100        640:  52%|█████▏    | 419/809 [08:19<07:42,  1.19s/it]
-     25/300      13.8G      1.227     0.4292     0.1515        202        640:  52%|█████▏    | 421/809 [08:21<07:40,  1.19s/it]
-     25/300      13.8G      1.227     0.4291     0.1515         90        640:  52%|█████▏    | 422/809 [08:22<07:34,  1.17s/it]
-     25/300      13.8G      1.229     0.4286     0.1516        361        640:  52%|█████▏    | 424/809 [08:24<07:40,  1.20s/it]
-     25/300      13.8G      1.227     0.4289     0.1513         84        640:  53%|█████▎    | 426/809 [08:27<07:32,  1.18s/it]
-     25/300      13.8G      1.227     0.4291     0.1513         98        640:  53%|█████▎    | 428/809 [08:29<07:27,  1.17s/it]
-     25/300      13.8G      1.226     0.4293     0.1512         83        640:  53%|█████▎    | 429/809 [08:30<07:23,  1.17s/it]
-     25/300      13.8G      1.225     0.4296      0.151        206        640:  53%|█████▎    | 431/809 [08:33<07:18,  1.16s/it]
-     25/300      13.8G      1.225     0.4297     0.1511        207        640:  54%|█████▎    | 433/809 [08:35<07:32,  1.20s/it]
-     25/300      13.8G      1.224       0.43     0.1509        201        640:  54%|█████▎    | 434/809 [08:36<07:24,  1.19s/it]
-     25/300      13.8G      1.225       0.43     0.1514        391        640:  54%|█████▍    | 436/809 [08:39<07:23,  1.19s/it]
-     25/300      13.8G      1.225     0.4301     0.1517        405        640:  54%|█████▍    | 438/809 [08:41<07:30,  1.21s/it]
-     25/300      13.8G      1.224     0.4303     0.1517        141        640:  54%|█████▍    | 439/809 [08:42<07:27,  1.21s/it]
-     25/300      13.8G      1.223     0.4307     0.1515        336        640:  55%|█████▍    | 441/809 [08:45<07:25,  1.21s/it]
-     25/300      13.8G      1.222     0.4311     0.1515        131        640:  55%|█████▍    | 443/809 [08:47<07:13,  1.18s/it]
-     25/300      13.8G      1.222     0.4312     0.1514        268        640:  55%|█████▍    | 444/809 [08:48<07:14,  1.19s/it]
-     25/300      13.8G      1.222     0.4314     0.1515        141        640:  55%|█████▌    | 446/809 [08:51<07:27,  1.23s/it]
-     25/300      13.8G      1.222     0.4318     0.1517        268        640:  55%|█████▌    | 448/809 [08:53<07:13,  1.20s/it]
-     25/300      13.8G      1.222     0.4315     0.1518        176        640:  56%|█████▌    | 449/809 [08:54<07:18,  1.22s/it]
-     25/300      13.8G      1.223     0.4311     0.1522        396        640:  56%|█████▌    | 451/809 [08:57<07:09,  1.20s/it]
-     25/300      13.8G      1.222     0.4313      0.152        174        640:  56%|█████▌    | 453/809 [08:59<07:01,  1.18s/it]
-     25/300      13.8G      1.223     0.4311     0.1521        502        640:  56%|█████▌    | 454/809 [09:00<07:03,  1.19s/it]
-     25/300      13.8G      1.222     0.4311      0.152        226        640:  56%|█████▋    | 456/809 [09:03<06:57,  1.18s/it]
-     25/300      13.8G      1.221     0.4314     0.1519        185        640:  57%|█████▋    | 458/809 [09:05<06:54,  1.18s/it]
-     25/300      13.8G      1.221     0.4314      0.152        523        640:  57%|█████▋    | 459/809 [09:06<06:56,  1.19s/it]
-     25/300      13.8G      1.221     0.4316     0.1519        200        640:  57%|█████▋    | 461/809 [09:09<06:51,  1.18s/it]
-     25/300      13.8G      1.221     0.4315     0.1519        214        640:  57%|█████▋    | 463/809 [09:11<06:48,  1.18s/it]
-     25/300      13.8G      1.221     0.4313     0.1519        382        640:  57%|█████▋    | 465/809 [09:13<06:54,  1.20s/it]
-     25/300      13.8G      1.221     0.4313     0.1519        145        640:  58%|█████▊    | 466/809 [09:15<06:52,  1.20s/it]
-     25/300      13.8G      1.221     0.4312     0.1518        185        640:  58%|█████▊    | 468/809 [09:17<06:45,  1.19s/it]
-     25/300      13.8G      1.221     0.4314     0.1517        178        640:  58%|█████▊    | 470/809 [09:19<06:38,  1.17s/it]
-     25/300      13.8G       1.22     0.4316     0.1516        192        640:  58%|█████▊    | 471/809 [09:20<06:36,  1.17s/it]
-     25/300      13.8G      1.221      0.431     0.1519        399        640:  58%|█████▊    | 473/809 [09:23<06:46,  1.21s/it]
-     25/300      13.8G      1.222     0.4306     0.1521        410        640:  59%|█████▊    | 475/809 [09:25<06:41,  1.20s/it]
-     25/300      13.8G      1.223     0.4304     0.1523        446        640:  59%|█████▉    | 476/809 [09:26<06:41,  1.21s/it]
-     25/300      13.8G      1.226     0.4296     0.1538        317        640:  59%|█████▉    | 478/809 [09:29<06:48,  1.23s/it]
-     25/300      13.8G      1.226     0.4295      0.154        313        640:  59%|█████▉    | 480/809 [09:31<06:38,  1.21s/it]
-     25/300      13.8G      1.226     0.4296      0.154        321        640:  59%|█████▉    | 481/809 [09:33<06:49,  1.25s/it]
-     25/300      13.8G      1.226     0.4294      0.154        296        640:  60%|█████▉    | 483/809 [09:35<06:34,  1.21s/it]
-     25/300      13.8G      1.226     0.4295     0.1539        142        640:  60%|█████▉    | 485/809 [09:37<06:24,  1.19s/it]
-     25/300      13.8G      1.225     0.4298     0.1539        234        640:  60%|██████    | 486/809 [09:39<06:19,  1.18s/it]
-     25/300      13.8G      1.225     0.4297     0.1538        372        640:  60%|██████    | 488/809 [09:41<06:17,  1.18s/it]
-     25/300      13.8G      1.226     0.4292     0.1541        258        640:  61%|██████    | 490/809 [09:43<06:25,  1.21s/it]
-     25/300      13.8G      1.226     0.4293     0.1541        324        640:  61%|██████    | 491/809 [09:45<06:21,  1.20s/it]
-     25/300      13.8G      1.227     0.4291     0.1545        536        640:  61%|██████    | 493/809 [09:47<06:30,  1.24s/it]
-     25/300      13.8G      1.227     0.4291     0.1545        232        640:  61%|██████    | 494/809 [09:48<06:22,  1.21s/it]
-     25/300      13.8G      1.226     0.4295     0.1544        153        640:  61%|██████▏   | 496/809 [09:51<06:25,  1.23s/it]
-     25/300      13.8G      1.226     0.4295     0.1545        236        640:  62%|██████▏   | 498/809 [09:53<06:15,  1.21s/it]
-     25/300      13.8G      1.226     0.4295     0.1544        137        640:  62%|██████▏   | 499/809 [09:54<06:15,  1.21s/it]
-     25/300      13.8G      1.225     0.4297     0.1543        195        640:  62%|██████▏   | 501/809 [09:57<06:15,  1.22s/it]
-     25/300      13.8G      1.226     0.4292     0.1543        432        640:  62%|██████▏   | 503/809 [09:59<06:11,  1.21s/it]
-     25/300      13.8G      1.226     0.4292     0.1542        221        640:  62%|██████▏   | 504/809 [10:00<06:04,  1.19s/it]
-     25/300      13.8G      1.227     0.4291     0.1543        166        640:  63%|██████▎   | 506/809 [10:03<06:01,  1.19s/it]
-     25/300      13.8G      1.226     0.4291     0.1541        137        640:  63%|██████▎   | 508/809 [10:05<06:07,  1.22s/it]
-     25/300      13.8G      1.227     0.4289     0.1542        249        640:  63%|██████▎   | 509/809 [10:06<06:01,  1.21s/it]
-     25/300      13.8G      1.228     0.4284     0.1542        432        640:  63%|██████▎   | 511/809 [10:09<05:57,  1.20s/it]
-     25/300      13.8G      1.229     0.4282     0.1542        255        640:  63%|██████▎   | 513/809 [10:11<06:05,  1.23s/it]
-     25/300      13.8G       1.23     0.4279     0.1552        443        640:  64%|██████▎   | 514/809 [10:13<06:09,  1.25s/it]
-     25/300      13.8G      1.228     0.4283     0.1551         79        640:  64%|██████▍   | 516/809 [10:15<05:51,  1.20s/it]
-     25/300      13.8G      1.228     0.4285     0.1549        220        640:  64%|██████▍   | 518/809 [10:17<05:47,  1.19s/it]
-     25/300      13.8G      1.228     0.4283      0.155        314        640:  64%|██████▍   | 519/809 [10:19<05:50,  1.21s/it]
-     25/300      13.8G      1.228     0.4282     0.1549        251        640:  64%|██████▍   | 521/809 [10:21<05:52,  1.22s/it]
-     25/300      13.8G      1.228     0.4283     0.1547        196        640:  65%|██████▍   | 523/809 [10:23<05:40,  1.19s/it]
-     25/300      13.8G      1.228     0.4283     0.1546        206        640:  65%|██████▍   | 524/809 [10:25<05:41,  1.20s/it]
-     25/300      13.8G      1.229     0.4281     0.1545        231        640:  65%|██████▌   | 526/809 [10:27<05:43,  1.22s/it]
-     25/300      13.8G       1.23     0.4279     0.1549        564        640:  65%|██████▌   | 528/809 [10:29<05:45,  1.23s/it]
-     25/300      13.8G       1.23     0.4279     0.1548        142        640:  65%|██████▌   | 529/809 [10:31<05:46,  1.24s/it]
-     25/300      13.8G      1.229      0.428     0.1546        150        640:  66%|██████▌   | 531/809 [10:33<05:31,  1.19s/it]
-     25/300      13.8G      1.229     0.4282     0.1546        149        640:  66%|██████▌   | 533/809 [10:35<05:25,  1.18s/it]
-     25/300      13.8G      1.229     0.4281     0.1546        164        640:  66%|██████▌   | 534/809 [10:37<05:22,  1.17s/it]
-     25/300      13.8G      1.229     0.4283     0.1544        402        640:  66%|██████▋   | 536/809 [10:39<05:21,  1.18s/it]
-     25/300      13.8G      1.228     0.4286     0.1541        177        640:  67%|██████▋   | 538/809 [10:41<05:20,  1.18s/it]
-     25/300      13.8G      1.227     0.4286     0.1541        279        640:  67%|██████▋   | 540/809 [10:44<05:15,  1.17s/it]
-     25/300      13.8G      1.227     0.4287      0.154        229        640:  67%|██████▋   | 541/809 [10:45<05:14,  1.17s/it]
-     25/300      13.8G      1.227     0.4287      0.154        187        640:  67%|██████▋   | 543/809 [10:47<05:12,  1.18s/it]
-     25/300      13.8G      1.228     0.4285      0.154        255        640:  67%|██████▋   | 545/809 [10:50<05:19,  1.21s/it]
-     25/300      13.8G      1.228     0.4285      0.154        150        640:  67%|██████▋   | 546/809 [10:51<05:12,  1.19s/it]
-     25/300      13.8G      1.229     0.4281     0.1541        320        640:  68%|██████▊   | 548/809 [10:53<05:10,  1.19s/it]
-     25/300      13.8G      1.231     0.4277     0.1544        271        640:  68%|██████▊   | 550/809 [10:56<05:08,  1.19s/it]
-     25/300      13.8G      1.231     0.4276     0.1545        301        640:  68%|██████▊   | 551/809 [10:57<05:06,  1.19s/it]
-     25/300      13.8G      1.232     0.4274     0.1545        312        640:  68%|██████▊   | 553/809 [10:59<05:11,  1.22s/it]
-     25/300      13.8G      1.233     0.4272     0.1546        212        640:  69%|██████▊   | 555/809 [11:02<05:06,  1.21s/it]
-     25/300      13.8G      1.233     0.4272     0.1546        223        640:  69%|██████▊   | 556/809 [11:03<05:03,  1.20s/it]
-     25/300      13.8G      1.233      0.427     0.1546        243        640:  69%|██████▉   | 558/809 [11:05<04:59,  1.19s/it]
-     25/300      13.8G      1.234     0.4266     0.1545        442        640:  69%|██████▉   | 560/809 [11:08<04:57,  1.19s/it]
-     25/300      13.8G      1.234     0.4266     0.1546        270        640:  69%|██████▉   | 561/809 [11:09<05:00,  1.21s/it]
-     25/300      13.8G      1.235     0.4264      0.155        366        640:  70%|██████▉   | 563/809 [11:11<05:01,  1.22s/it]
-     25/300      13.8G      1.235     0.4265      0.155        129        640:  70%|██████▉   | 565/809 [11:14<04:51,  1.19s/it]
-     25/300      13.8G      1.236     0.4263     0.1555        593        640:  70%|██████▉   | 566/809 [11:15<04:55,  1.22s/it]
-     25/300      13.8G      1.238      0.426     0.1558        326        640:  70%|███████   | 568/809 [11:17<04:55,  1.23s/it]
-     25/300      13.8G      1.239     0.4257     0.1564        431        640:  70%|███████   | 569/809 [11:19<05:04,  1.27s/it]
-     25/300      13.8G      1.239     0.4257     0.1565        260        640:  71%|███████   | 571/809 [11:21<04:50,  1.22s/it]
-     25/300      13.8G      1.238     0.4258     0.1564        262        640:  71%|███████   | 573/809 [11:23<04:40,  1.19s/it]
-     25/300      13.8G      1.238     0.4258     0.1563        207        640:  71%|███████   | 574/809 [11:25<04:38,  1.18s/it]
-     25/300      13.8G       1.24     0.4256      0.157        264        640:  71%|███████   | 576/809 [11:27<04:43,  1.22s/it]
-     25/300      13.8G       1.24     0.4256      0.157        265        640:  71%|███████▏  | 578/809 [11:29<04:38,  1.20s/it]
-     25/300      13.8G       1.24     0.4255      0.157        335        640:  72%|███████▏  | 579/809 [11:31<04:35,  1.20s/it]
-     25/300      13.8G       1.24     0.4256     0.1568        335        640:  72%|███████▏  | 581/809 [11:33<04:29,  1.18s/it]
-     25/300      13.8G       1.24     0.4257     0.1567        300        640:  72%|███████▏  | 583/809 [11:35<04:25,  1.17s/it]
-     25/300      13.8G       1.24     0.4258     0.1566        295        640:  72%|███████▏  | 584/809 [11:38<04:25,  1.18s/it]
-     25/300      13.8G       1.24     0.4258     0.1566        144        640:  72%|███████▏  | 586/809 [11:39<04:23,  1.18s/it]
-     25/300      13.8G       1.24      0.426     0.1566        174        640:  73%|███████▎  | 588/809 [11:41<04:21,  1.18s/it]
-     25/300      13.8G       1.24     0.4257     0.1567        153        640:  73%|███████▎  | 590/809 [11:44<04:16,  1.17s/it]
-     25/300      13.8G      1.241     0.4255     0.1572        312        640:  73%|███████▎  | 591/809 [11:45<04:19,  1.19s/it]
-     25/300      13.8G      1.242     0.4254     0.1573        433        640:  73%|███████▎  | 593/809 [11:47<04:20,  1.21s/it]
-     25/300      13.8G      1.242     0.4254     0.1573        335        640:  74%|███████▎  | 595/809 [11:50<04:17,  1.20s/it]
-     25/300      13.8G      1.242     0.4255     0.1573        420        640:  74%|███████▎  | 596/809 [11:51<04:15,  1.20s/it]
-     25/300      13.8G      1.241     0.4257      0.157        110        640:  74%|███████▍  | 598/809 [11:53<04:15,  1.21s/it]
-     25/300      13.8G      1.242     0.4257     0.1571        231        640:  74%|███████▍  | 600/809 [11:56<04:08,  1.19s/it]
-     25/300      13.8G      1.242     0.4256     0.1572        429        640:  74%|███████▍  | 601/809 [11:57<04:15,  1.23s/it]
-     25/300      13.8G      1.242     0.4256     0.1571        225        640:  75%|███████▍  | 603/809 [11:59<04:06,  1.20s/it]
-     25/300      13.8G      1.242     0.4255      0.157        135        640:  75%|███████▍  | 605/809 [12:02<04:01,  1.18s/it]
-     25/300      13.8G      1.241     0.4256     0.1569        204        640:  75%|███████▍  | 606/809 [12:03<03:58,  1.18s/it]
-     25/300      13.8G      1.241     0.4255     0.1569        410        640:  75%|███████▌  | 608/809 [12:05<03:58,  1.18s/it]
-     25/300      13.8G      1.241     0.4254     0.1568        279        640:  75%|███████▌  | 610/809 [12:08<03:57,  1.20s/it]
-     25/300      13.8G      1.241     0.4255     0.1568        143        640:  76%|███████▌  | 611/809 [12:09<03:54,  1.18s/it]
-     25/300      13.8G      1.241     0.4255     0.1571        185        640:  76%|███████▌  | 613/809 [12:11<03:52,  1.19s/it]
-     25/300      13.8G      1.242     0.4253     0.1572        248        640:  76%|███████▌  | 615/809 [12:13<03:50,  1.19s/it]
-     25/300      13.8G      1.243     0.4251     0.1575        361        640:  76%|███████▌  | 616/809 [12:15<03:53,  1.21s/it]
-     25/300      13.8G      1.244     0.4248     0.1576        296        640:  76%|███████▋  | 618/809 [12:17<03:52,  1.22s/it]
-     25/300      13.8G      1.244     0.4248     0.1575        178        640:  77%|███████▋  | 620/809 [12:20<03:44,  1.19s/it]
-     25/300      13.8G      1.244     0.4247     0.1575        282        640:  77%|███████▋  | 621/809 [12:21<03:42,  1.18s/it]
-     25/300      13.8G      1.246     0.4242     0.1582        426        640:  77%|███████▋  | 623/809 [12:23<03:47,  1.22s/it]
-     25/300      13.8G      1.246     0.4242      0.158        160        640:  77%|███████▋  | 625/809 [12:26<03:46,  1.23s/it]
-     25/300      13.8G      1.246     0.4243     0.1579        170        640:  77%|███████▋  | 626/809 [12:27<03:40,  1.20s/it]
-     25/300      13.8G      1.246     0.4242     0.1579        176        640:  78%|███████▊  | 628/809 [12:29<03:34,  1.19s/it]
-     25/300      13.8G      1.247     0.4239      0.158        188        640:  78%|███████▊  | 630/809 [12:32<03:37,  1.22s/it]
-     25/300      13.8G      1.247     0.4239      0.158        416        640:  78%|███████▊  | 631/809 [12:33<03:44,  1.26s/it]
-     25/300      13.8G      1.247     0.4238     0.1579        150        640:  78%|███████▊  | 633/809 [12:35<03:35,  1.22s/it]
-     25/300      13.8G      1.246     0.4241     0.1577        134        640:  78%|███████▊  | 635/809 [12:38<03:25,  1.18s/it]
-     25/300      13.8G      1.246     0.4241     0.1577        257        640:  79%|███████▊  | 636/809 [12:39<03:24,  1.18s/it]
-     25/300      13.8G      1.246     0.4241     0.1576        153        640:  79%|███████▉  | 638/809 [12:41<03:26,  1.21s/it]
-     25/300      13.8G      1.246     0.4242     0.1575        200        640:  79%|███████▉  | 640/809 [12:44<03:20,  1.19s/it]
-     25/300      13.8G      1.245     0.4243     0.1574        165        640:  79%|███████▉  | 641/809 [12:45<03:20,  1.19s/it]
-     25/300      13.8G      1.245     0.4244     0.1573        309        640:  79%|███████▉  | 643/809 [12:47<03:25,  1.24s/it]
-     25/300      13.8G      1.245     0.4246     0.1572        367        640:  80%|███████▉  | 645/809 [12:50<03:20,  1.22s/it]
-     25/300      13.8G      1.245     0.4245     0.1573        301        640:  80%|███████▉  | 646/809 [12:51<03:17,  1.21s/it]
-     25/300      13.8G      1.244     0.4247     0.1571        165        640:  80%|████████  | 648/809 [12:53<03:13,  1.20s/it]
-     25/300      13.8G      1.244     0.4247     0.1571        242        640:  80%|████████  | 650/809 [12:56<03:08,  1.19s/it]
-     25/300      13.8G      1.244     0.4245     0.1571        450        640:  80%|████████  | 651/809 [12:57<03:12,  1.22s/it]
-     25/300      13.8G      1.245     0.4243     0.1575        266        640:  81%|████████  | 653/809 [12:59<03:08,  1.21s/it]
-     25/300      13.8G      1.245     0.4243     0.1575        485        640:  81%|████████  | 655/809 [13:02<03:04,  1.20s/it]
-     25/300      13.8G      1.244     0.4244     0.1574        220        640:  81%|████████  | 656/809 [13:03<03:03,  1.20s/it]
-     25/300      13.8G      1.245     0.4242     0.1574        290        640:  81%|████████▏ | 658/809 [13:05<03:00,  1.20s/it]
-     25/300      13.8G      1.244     0.4243     0.1574        173        640:  82%|████████▏ | 660/809 [13:08<02:55,  1.18s/it]
-     25/300      13.8G      1.244     0.4243     0.1575        345        640:  82%|████████▏ | 661/809 [13:09<02:55,  1.18s/it]
-     25/300      13.8G      1.244     0.4242     0.1575        329        640:  82%|████████▏ | 663/809 [13:11<02:53,  1.19s/it]
-     25/300      13.8G      1.244     0.4243     0.1573        177        640:  82%|████████▏ | 665/809 [13:14<02:51,  1.19s/it]
-     25/300      13.8G      1.244     0.4242     0.1573        383        640:  82%|████████▏ | 666/809 [13:15<02:49,  1.19s/it]
-     25/300      13.8G      1.245     0.4238     0.1574        240        640:  83%|████████▎ | 668/809 [13:17<02:48,  1.20s/it]
-     25/300      13.8G      1.245     0.4239     0.1574        333        640:  83%|████████▎ | 670/809 [13:20<02:45,  1.19s/it]
-     25/300      13.8G      1.244     0.4241     0.1574        171        640:  83%|████████▎ | 672/809 [13:22<02:40,  1.17s/it]
-     25/300      13.8G      1.245     0.4239     0.1574        226        640:  83%|████████▎ | 673/809 [13:23<02:43,  1.20s/it]
-     25/300      13.8G      1.244      0.424     0.1573        156        640:  83%|████████▎ | 675/809 [13:26<02:39,  1.19s/it]
-     25/300      13.8G      1.243     0.4242     0.1571        187        640:  84%|████████▎ | 677/809 [13:28<02:38,  1.20s/it]
-     25/300      13.8G      1.243     0.4245     0.1571        163        640:  84%|████████▍ | 678/809 [13:29<02:36,  1.20s/it]
-     25/300      13.8G      1.242     0.4247      0.157        299        640:  84%|████████▍ | 680/809 [13:32<02:31,  1.18s/it]
-     25/300      13.8G      1.243     0.4242     0.1573        304        640:  84%|████████▍ | 682/809 [13:34<02:33,  1.21s/it]
-     25/300      13.8G      1.244     0.4239     0.1573        337        640:  84%|████████▍ | 683/809 [13:35<02:32,  1.21s/it]
-     25/300      13.8G      1.246     0.4236     0.1578        397        640:  85%|████████▍ | 685/809 [13:38<02:32,  1.23s/it]
-     25/300      13.8G      1.247     0.4233     0.1583        519        640:  85%|████████▍ | 686/809 [13:39<02:34,  1.25s/it]
-     25/300      13.8G      1.248     0.4228     0.1586        651        640:  85%|████████▌ | 688/809 [13:42<02:33,  1.27s/it]
-     25/300      13.8G      1.249     0.4228     0.1585        138        640:  85%|████████▌ | 690/809 [13:44<02:27,  1.24s/it]
-     25/300      13.8G      1.248     0.4228     0.1585        267        640:  85%|████████▌ | 691/809 [13:45<02:23,  1.22s/it]
-     25/300      13.8G      1.249     0.4227     0.1585        171        640:  86%|████████▌ | 693/809 [13:47<02:18,  1.19s/it]
-     25/300      13.8G      1.249     0.4226     0.1586        384        640:  86%|████████▌ | 695/809 [13:50<02:17,  1.21s/it]
-     25/300      13.8G      1.249     0.4227     0.1585        203        640:  86%|████████▌ | 696/809 [13:51<02:15,  1.20s/it]
-     25/300      13.8G      1.249     0.4225     0.1589        425        640:  86%|████████▋ | 698/809 [13:54<02:16,  1.23s/it]
-     25/300      13.8G      1.249     0.4225     0.1588        184        640:  87%|████████▋ | 700/809 [13:56<02:09,  1.19s/it]
-     25/300      13.8G       1.25     0.4224     0.1589        388        640:  87%|████████▋ | 701/809 [13:57<02:08,  1.19s/it]
-     25/300      13.8G       1.25     0.4222      0.159        251        640:  87%|████████▋ | 703/809 [14:00<02:07,  1.20s/it]
-     25/300      13.8G      1.251     0.4219     0.1592        429        640:  87%|████████▋ | 705/809 [14:02<02:08,  1.24s/it]
-     25/300      13.8G      1.251     0.4218     0.1592        400        640:  87%|████████▋ | 706/809 [14:03<02:05,  1.22s/it]
-     25/300      13.8G      1.251      0.422      0.159        205        640:  88%|████████▊ | 708/809 [14:06<02:01,  1.20s/it]
-     25/300      13.8G       1.25     0.4221      0.159        175        640:  88%|████████▊ | 710/809 [14:08<01:56,  1.17s/it]
-     25/300      13.8G      1.251      0.422     0.1591        455        640:  88%|████████▊ | 711/809 [14:09<01:56,  1.19s/it]
-     25/300      13.8G      1.251     0.4217     0.1593        310        640:  88%|████████▊ | 713/809 [14:12<01:56,  1.21s/it]
-     25/300      13.8G      1.251     0.4219     0.1592        172        640:  88%|████████▊ | 715/809 [14:14<01:50,  1.17s/it]
-     25/300      13.8G      1.251     0.4219     0.1592        257        640:  89%|████████▊ | 716/809 [14:15<01:49,  1.18s/it]
-     25/300      13.8G      1.251     0.4219     0.1591        229        640:  89%|████████▉ | 718/809 [14:17<01:46,  1.17s/it]
-     25/300      13.8G       1.25      0.422      0.159        174        640:  89%|████████▉ | 720/809 [14:20<01:43,  1.16s/it]
-     25/300      13.8G       1.25     0.4223     0.1589        108        640:  89%|████████▉ | 722/809 [14:22<01:40,  1.16s/it]
-     25/300      13.8G      1.249     0.4223     0.1589        156        640:  89%|████████▉ | 723/809 [14:23<01:39,  1.15s/it]
-     25/300      13.8G       1.25     0.4222      0.159        370        640:  90%|████████▉ | 725/809 [14:26<01:38,  1.17s/it]
-     25/300      13.8G      1.252     0.4217     0.1598        229        640:  90%|████████▉ | 727/809 [14:28<01:37,  1.19s/it]
-     25/300      13.8G      1.253     0.4215     0.1601        307        640:  90%|████████▉ | 728/809 [14:29<01:36,  1.19s/it]
-     25/300      13.8G      1.253     0.4212       0.16        245        640:  90%|█████████ | 730/809 [14:32<01:35,  1.21s/it]
-     25/300      13.8G      1.254      0.421     0.1603        287        640:  90%|█████████ | 732/809 [14:34<01:33,  1.21s/it]
-     25/300      13.8G      1.254     0.4212     0.1602        165        640:  91%|█████████ | 733/809 [14:35<01:30,  1.20s/it]
-     25/300      13.8G      1.253     0.4213     0.1601        312        640:  91%|█████████ | 735/809 [14:38<01:27,  1.19s/it]
-     25/300      13.8G      1.253     0.4215       0.16        224        640:  91%|█████████ | 737/809 [14:40<01:26,  1.20s/it]
-     25/300      13.8G      1.253     0.4214       0.16        296        640:  91%|█████████ | 738/809 [14:41<01:24,  1.19s/it]
-     25/300      13.8G      1.253     0.4213     0.1599        241        640:  91%|█████████▏| 740/809 [14:43<01:21,  1.18s/it]
-     25/300      13.8G      1.251     0.4218     0.1597         73        640:  92%|█████████▏| 742/809 [14:46<01:19,  1.18s/it]
-     25/300      13.8G      1.252     0.4216     0.1603        332        640:  92%|█████████▏| 743/809 [14:47<01:19,  1.21s/it]
-     25/300      13.8G      1.252     0.4216     0.1603        127        640:  92%|█████████▏| 745/809 [14:50<01:17,  1.22s/it]
-     25/300      13.8G      1.252     0.4216     0.1602        357        640:  92%|█████████▏| 747/809 [14:52<01:14,  1.20s/it]
-     25/300      13.8G      1.252     0.4217     0.1602        183        640:  92%|█████████▏| 748/809 [14:53<01:13,  1.20s/it]
-     25/300      13.8G      1.251     0.4218     0.1602        213        640:  93%|█████████▎| 750/809 [14:56<01:10,  1.19s/it]
-     25/300      13.8G       1.25     0.4221     0.1601        132        640:  93%|█████████▎| 752/809 [14:58<01:06,  1.17s/it]
-     25/300      13.8G       1.25     0.4223       0.16        272        640:  93%|█████████▎| 754/809 [15:00<01:04,  1.18s/it]
-     25/300      13.8G      1.251     0.4221     0.1602        273        640:  93%|█████████▎| 755/809 [15:01<01:04,  1.19s/it]
-     25/300      13.8G       1.25     0.4223     0.1601        167        640:  94%|█████████▎| 757/809 [15:04<01:02,  1.20s/it]
-     25/300      13.8G       1.25     0.4224       0.16        132        640:  94%|█████████▍| 759/809 [15:06<00:59,  1.19s/it]
-     25/300      13.8G       1.25     0.4224     0.1599        207        640:  94%|█████████▍| 760/809 [15:08<00:59,  1.22s/it]
-     25/300      13.8G       1.25     0.4224     0.1599        190        640:  94%|█████████▍| 762/809 [15:10<00:58,  1.24s/it]
-     25/300      13.8G       1.25     0.4225     0.1598        166        640:  94%|█████████▍| 763/809 [15:11<00:55,  1.22s/it]
-     25/300      13.8G      1.249     0.4225     0.1597        105        640:  95%|█████████▍| 765/809 [15:14<00:54,  1.25s/it]
-     25/300      13.8G      1.249     0.4226     0.1597        188        640:  95%|█████████▍| 766/809 [15:15<00:52,  1.22s/it]
-     25/300      13.8G      1.249     0.4225     0.1597        293        640:  95%|█████████▍| 768/809 [15:17<00:49,  1.22s/it]
-     25/300      13.8G       1.25     0.4223     0.1597        199        640:  95%|█████████▌| 770/809 [15:20<00:47,  1.21s/it]
-     25/300      13.8G       1.25     0.4221     0.1598        399        640:  95%|█████████▌| 771/809 [15:21<00:46,  1.22s/it]
-     25/300      13.8G       1.25     0.4221     0.1598        260        640:  96%|█████████▌| 773/809 [15:23<00:43,  1.20s/it]
-     25/300      13.8G      1.249     0.4225     0.1597        160        640:  96%|█████████▌| 775/809 [15:26<00:40,  1.18s/it]
-     25/300      13.8G      1.249     0.4226     0.1596        272        640:  96%|█████████▌| 776/809 [15:27<00:38,  1.18s/it]
-     25/300      13.8G      1.249     0.4228     0.1595        121        640:  96%|█████████▌| 778/809 [15:29<00:36,  1.18s/it]
-     25/300      13.8G      1.248     0.4228     0.1595        360        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.18s/it]
-     25/300      13.8G      1.248      0.423     0.1593        116        640:  97%|█████████▋| 782/809 [15:34<00:31,  1.17s/it]
-     25/300      13.8G      1.248     0.4231     0.1593        236        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.17s/it]
-     25/300      13.8G      1.248     0.4231     0.1592        176        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.19s/it]
-     25/300      13.8G      1.246     0.4237     0.1591        117        640:  97%|█████████▋| 787/809 [15:40<00:25,  1.17s/it]
-     25/300      13.8G      1.246     0.4237      0.159        382        640:  97%|█████████▋| 788/809 [15:41<00:24,  1.17s/it]
-     25/300      13.8G      1.246     0.4239     0.1589        152        640:  98%|█████████▊| 790/809 [15:43<00:21,  1.16s/it]
-     25/300      13.8G      1.245     0.4241     0.1587        204        640:  98%|█████████▊| 792/809 [15:46<00:19,  1.16s/it]
-     25/300      13.8G      1.246     0.4239     0.1589        512        640:  98%|█████████▊| 793/809 [15:47<00:19,  1.22s/it]
-     25/300      13.8G      1.246      0.424     0.1588        225        640:  98%|█████████▊| 795/809 [15:49<00:16,  1.21s/it]
-     25/300      13.8G      1.246      0.424     0.1588        231        640:  99%|█████████▊| 797/809 [15:52<00:14,  1.18s/it]
-     25/300      13.8G      1.245     0.4241     0.1587        175        640:  99%|█████████▉| 799/809 [15:54<00:11,  1.17s/it]
-     25/300      13.8G      1.245     0.4242     0.1587         79        640:  99%|█████████▉| 800/809 [15:55<00:10,  1.17s/it]
-     25/300      13.8G      1.245     0.4241     0.1586        160        640:  99%|█████████▉| 802/809 [15:58<00:08,  1.19s/it]
-     25/300      13.8G      1.245     0.4243     0.1586        208        640:  99%|█████████▉| 804/809 [16:00<00:05,  1.18s/it]
-     25/300      13.8G      1.245     0.4243     0.1585        181        640: 100%|█████████▉| 805/809 [16:01<00:04,  1.17s/it]
-     25/300      13.8G      1.244     0.4244     0.1584        192        640: 100%|█████████▉| 807/809 [16:03<00:02,  1.16s/it]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   0%|          | 0/69 [00:00<?, ?it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  12%|█▏        | 8/69 [00:02<00:15,  3.96it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  23%|██▎       | 16/69 [00:04<00:13,  3.87it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  35%|███▍      | 24/69 [00:06<00:11,  3.92it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  46%|████▋     | 32/69 [00:08<00:09,  3.98it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  58%|█████▊    | 40/69 [00:10<00:07,  4.00it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  68%|██████▊   | 47/69 [00:12<00:06,  3.48it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  80%|███████▉  | 55/69 [00:14<00:03,  3.85it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|█████████▏| 63/69 [00:16<00:01,  3.91it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]
-                   all        548      38759      0.421      0.274      0.266      0.138
-     26/300      13.4G      1.259     0.4192    0.09375        316        640:   0%|          | 1/809 [00:01<15:38,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.4G      1.113     0.4848     0.1031        144        640:   0%|          | 3/809 [00:03<15:29,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.4G      1.076      0.503     0.1068        152        640:   1%|          | 5/809 [00:05<15:26,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.4G      1.107     0.4948     0.1108        240        640:   1%|          | 6/809 [00:06<15:29,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.046     0.5188     0.1129        178        640:   1%|          | 8/809 [00:09<16:04,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.024     0.5184      0.109        154        640:   1%|          | 10/809 [00:11<15:42,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.018     0.5152     0.1061        163        640:   1%|▏         | 11/809 [00:12<15:52,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.037     0.5079     0.1051        228        640:   2%|▏         | 13/809 [00:15<15:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.123     0.4841     0.1314        356        640:   2%|▏         | 15/809 [00:17<16:02,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.097     0.4915     0.1289        127        640:   2%|▏         | 16/809 [00:19<16:06,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.153     0.4693     0.1353        294        640:   2%|▏         | 18/809 [00:21<15:53,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.189     0.4612     0.1542        353        640:   2%|▏         | 20/809 [00:23<15:52,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.189     0.4607     0.1521        171        640:   3%|▎         | 21/809 [00:24<15:45,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.208      0.452     0.1525        339        640:   3%|▎         | 23/809 [00:27<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.203     0.4519     0.1498        239        640:   3%|▎         | 25/809 [00:29<15:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.205     0.4519     0.1497        315        640:   3%|▎         | 27/809 [00:32<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.217     0.4486     0.1529        358        640:   3%|▎         | 28/809 [00:33<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.215     0.4494     0.1499        208        640:   4%|▎         | 30/809 [00:35<15:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.208     0.4519     0.1495        214        640:   4%|▍         | 32/809 [00:38<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G        1.2     0.4549     0.1479        177        640:   4%|▍         | 33/809 [00:39<15:25,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.198     0.4547     0.1451        271        640:   4%|▍         | 35/809 [00:41<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.207     0.4548     0.1552        234        640:   5%|▍         | 37/809 [00:44<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.203     0.4542     0.1524        157        640:   5%|▍         | 39/809 [00:46<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.206     0.4531     0.1524        304        640:   5%|▍         | 40/809 [00:47<15:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.209     0.4504     0.1506        267        640:   5%|▌         | 42/809 [00:49<14:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.204     0.4506     0.1493        105        640:   5%|▌         | 44/809 [00:52<15:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.196     0.4527     0.1484        155        640:   6%|▌         | 45/809 [00:53<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.203     0.4502      0.149        287        640:   6%|▌         | 47/809 [00:55<15:25,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.216     0.4467     0.1523        226        640:   6%|▌         | 49/809 [00:58<15:23,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.212     0.4471     0.1514        187        640:   6%|▌         | 50/809 [00:59<15:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4428     0.1511        374        640:   6%|▋         | 52/809 [01:01<15:05,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4399     0.1509        369        640:   7%|▋         | 54/809 [01:04<15:02,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4383     0.1514        202        640:   7%|▋         | 55/809 [01:05<14:48,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4406     0.1496        132        640:   7%|▋         | 57/809 [01:07<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4378     0.1504        318        640:   7%|▋         | 59/809 [01:10<14:39,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4368     0.1493        276        640:   8%|▊         | 61/809 [01:12<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4362     0.1489        385        640:   8%|▊         | 62/809 [01:13<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4345     0.1481        231        640:   8%|▊         | 64/809 [01:16<14:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4344     0.1494        408        640:   8%|▊         | 66/809 [01:18<14:55,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233      0.436     0.1495        253        640:   8%|▊         | 67/809 [01:19<14:47,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4396      0.148        178        640:   9%|▊         | 69/809 [01:22<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4386     0.1506        245        640:   9%|▉         | 71/809 [01:24<14:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4374     0.1524        430        640:   9%|▉         | 72/809 [01:25<15:13,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4387     0.1515        236        640:   9%|▉         | 74/809 [01:28<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4388     0.1506        248        640:   9%|▉         | 76/809 [01:30<14:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4379     0.1519        300        640:  10%|▉         | 77/809 [01:31<14:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4397      0.151        164        640:  10%|▉         | 79/809 [01:34<14:32,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4392      0.151        165        640:  10%|█         | 81/809 [01:36<14:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4387     0.1508        224        640:  10%|█         | 82/809 [01:37<14:53,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4398      0.151        497        640:  10%|█         | 84/809 [01:40<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4399     0.1505         81        640:  11%|█         | 85/809 [01:41<14:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4374     0.1528        531        640:  11%|█         | 87/809 [01:43<14:52,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4361     0.1535        213        640:  11%|█         | 89/809 [01:46<14:47,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.244     0.4343     0.1538        341        640:  11%|█         | 90/809 [01:47<14:46,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.25     0.4326     0.1593        139        640:  11%|█▏        | 92/809 [01:50<14:50,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249     0.4335      0.159        219        640:  12%|█▏        | 94/809 [01:52<14:23,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.25     0.4328     0.1587        356        640:  12%|█▏        | 95/809 [01:53<14:17,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.248     0.4343     0.1581        268        640:  12%|█▏        | 97/809 [01:56<14:17,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249     0.4338     0.1582        224        640:  12%|█▏        | 99/809 [01:58<14:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.246     0.4342     0.1575        162        640:  12%|█▏        | 100/809 [01:59<13:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241     0.4348     0.1565        167        640:  13%|█▎        | 102/809 [02:02<13:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241     0.4345     0.1561        174        640:  13%|█▎        | 104/809 [02:04<14:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4346     0.1559        207        640:  13%|█▎        | 105/809 [02:05<13:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.243     0.4337     0.1563        267        640:  13%|█▎        | 107/809 [02:08<13:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.245     0.4331     0.1565        310        640:  13%|█▎        | 109/809 [02:10<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.242     0.4349     0.1559        100        640:  14%|█▎        | 110/809 [02:11<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4361     0.1548        233        640:  14%|█▍        | 112/809 [02:13<13:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4373     0.1551        105        640:  14%|█▍        | 114/809 [02:16<13:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4376     0.1543        235        640:  14%|█▍        | 116/809 [02:18<13:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4375     0.1543        335        640:  14%|█▍        | 117/809 [02:19<13:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4371     0.1538        128        640:  15%|█▍        | 119/809 [02:22<13:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4367     0.1538        342        640:  15%|█▍        | 121/809 [02:24<13:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4357     0.1536        174        640:  15%|█▌        | 122/809 [02:25<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4355     0.1539        166        640:  15%|█▌        | 124/809 [02:28<13:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4356     0.1537         62        640:  16%|█▌        | 126/809 [02:30<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.242     0.4349     0.1539        321        640:  16%|█▌        | 127/809 [02:31<13:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239      0.436      0.154        271        640:  16%|█▌        | 129/809 [02:34<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4364     0.1535        220        640:  16%|█▌        | 131/809 [02:36<13:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4359     0.1539        330        640:  16%|█▋        | 132/809 [02:37<13:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233      0.438      0.153        152        640:  17%|█▋        | 134/809 [02:40<13:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4394     0.1524        130        640:  17%|█▋        | 136/809 [02:42<13:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4388     0.1523        299        640:  17%|█▋        | 137/809 [02:43<13:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4389     0.1518        124        640:  17%|█▋        | 139/809 [02:46<13:49,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4377     0.1516        304        640:  17%|█▋        | 141/809 [02:48<13:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4366     0.1519        365        640:  18%|█▊        | 142/809 [02:49<13:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4365     0.1515        180        640:  18%|█▊        | 144/809 [02:52<13:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4365     0.1517        232        640:  18%|█▊        | 146/809 [02:54<13:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4367     0.1515        215        640:  18%|█▊        | 147/809 [02:55<13:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4378     0.1509        196        640:  18%|█▊        | 149/809 [02:58<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4367     0.1532        379        640:  19%|█▊        | 151/809 [03:00<13:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4364     0.1536        535        640:  19%|█▉        | 152/809 [03:01<13:29,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4356     0.1545        394        640:  19%|█▉        | 154/809 [03:04<13:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4362     0.1541        141        640:  19%|█▉        | 156/809 [03:06<13:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4368      0.154        298        640:  19%|█▉        | 157/809 [03:07<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235      0.436     0.1541        184        640:  20%|█▉        | 159/809 [03:10<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4352     0.1552        318        640:  20%|█▉        | 161/809 [03:12<13:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4352     0.1554        251        640:  20%|██        | 162/809 [03:13<13:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4347     0.1557        362        640:  20%|██        | 164/809 [03:16<13:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4352     0.1554        467        640:  21%|██        | 166/809 [03:18<12:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4351     0.1554        113        640:  21%|██        | 167/809 [03:19<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4353     0.1551        132        640:  21%|██        | 169/809 [03:22<12:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238      0.435     0.1547        377        640:  21%|██        | 171/809 [03:24<12:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4349     0.1546        332        640:  21%|██▏       | 172/809 [03:25<12:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4345     0.1554        359        640:  22%|██▏       | 174/809 [03:28<12:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4356     0.1549        349        640:  22%|██▏       | 176/809 [03:30<12:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4363     0.1547         62        640:  22%|██▏       | 177/809 [03:31<12:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4356     0.1552        238        640:  22%|██▏       | 179/809 [03:34<12:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4359     0.1547        240        640:  22%|██▏       | 181/809 [03:36<12:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4366     0.1544        103        640:  22%|██▏       | 182/809 [03:38<12:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4373      0.154        367        640:  23%|██▎       | 184/809 [03:40<13:14,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4375     0.1538        209        640:  23%|██▎       | 185/809 [03:41<12:50,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4379     0.1534        174        640:  23%|██▎       | 187/809 [03:44<12:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4373     0.1533        120        640:  23%|██▎       | 189/809 [03:46<12:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4371     0.1533        197        640:  24%|██▎       | 191/809 [03:48<12:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4369     0.1532        318        640:  24%|██▎       | 192/809 [03:50<12:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4366     0.1527        116        640:  24%|██▍       | 194/809 [03:52<11:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4347     0.1533        148        640:  24%|██▍       | 196/809 [03:54<11:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4341     0.1535        198        640:  24%|██▍       | 198/809 [03:57<12:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241     0.4338     0.1535        178        640:  25%|██▍       | 199/809 [03:58<12:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.244     0.4325     0.1535        231        640:  25%|██▍       | 201/809 [04:00<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.246     0.4321     0.1533        340        640:  25%|██▌       | 203/809 [04:03<11:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.243      0.433     0.1528        147        640:  25%|██▌       | 205/809 [04:05<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.246     0.4322     0.1551        546        640:  25%|██▌       | 206/809 [04:06<12:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.248     0.4312     0.1556        320        640:  26%|██▌       | 208/809 [04:09<12:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249      0.431     0.1555        302        640:  26%|██▌       | 209/809 [04:10<12:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.246     0.4317      0.155        191        640:  26%|██▌       | 211/809 [04:12<12:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.247     0.4315     0.1547        377        640:  26%|██▋       | 213/809 [04:15<12:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.246     0.4316     0.1544        194        640:  26%|██▋       | 214/809 [04:16<11:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.248     0.4309     0.1543        260        640:  27%|██▋       | 216/809 [04:18<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.248     0.4305      0.154        230        640:  27%|██▋       | 218/809 [04:21<11:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249     0.4302     0.1538        323        640:  27%|██▋       | 220/809 [04:23<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249     0.4299     0.1536        228        640:  27%|██▋       | 221/809 [04:24<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.251     0.4292     0.1539         91        640:  28%|██▊       | 223/809 [04:27<11:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.253     0.4287      0.154        363        640:  28%|██▊       | 224/809 [04:28<12:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.251     0.4289     0.1543        448        640:  28%|██▊       | 226/809 [04:30<11:54,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.249     0.4291     0.1539        274        640:  28%|██▊       | 228/809 [04:33<11:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.247     0.4293     0.1536        127        640:  28%|██▊       | 230/809 [04:35<11:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.247     0.4292     0.1535        313        640:  29%|██▊       | 231/809 [04:36<11:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.245     0.4298     0.1529        142        640:  29%|██▉       | 233/809 [04:39<11:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.244     0.4302     0.1527        273        640:  29%|██▉       | 235/809 [04:41<11:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.243     0.4302     0.1526        316        640:  29%|██▉       | 236/809 [04:42<11:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.243     0.4304     0.1526        291        640:  29%|██▉       | 238/809 [04:45<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.242     0.4307     0.1528        258        640:  30%|██▉       | 240/809 [04:47<11:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.242     0.4309     0.1531        345        640:  30%|██▉       | 241/809 [04:48<11:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.243     0.4305     0.1535        585        640:  30%|███       | 243/809 [04:51<11:38,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241      0.431     0.1531        286        640:  30%|███       | 245/809 [04:53<11:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241     0.4309     0.1534        313        640:  30%|███       | 246/809 [04:54<11:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24      0.431     0.1534        363        640:  31%|███       | 248/809 [04:57<11:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.241     0.4307     0.1536        424        640:  31%|███       | 250/809 [04:59<11:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4307     0.1535        225        640:  31%|███       | 251/809 [05:00<11:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4316     0.1533        161        640:  31%|███▏      | 253/809 [05:03<11:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4315     0.1535        281        640:  32%|███▏      | 255/809 [05:05<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4317     0.1533        128        640:  32%|███▏      | 256/809 [05:06<11:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4321     0.1529         98        640:  32%|███▏      | 258/809 [05:09<10:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4324     0.1527        136        640:  32%|███▏      | 260/809 [05:11<10:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4328     0.1524        166        640:  32%|███▏      | 262/809 [05:13<10:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236      0.432      0.153        374        640:  33%|███▎      | 263/809 [05:14<10:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4311     0.1537        312        640:  33%|███▎      | 265/809 [05:17<10:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238      0.431     0.1535        162        640:  33%|███▎      | 267/809 [05:19<10:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4311     0.1535         94        640:  33%|███▎      | 268/809 [05:20<10:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4309     0.1532        209        640:  33%|███▎      | 270/809 [05:23<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4312     0.1529        184        640:  34%|███▎      | 272/809 [05:25<10:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236      0.431     0.1529        258        640:  34%|███▎      | 273/809 [05:26<10:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4303     0.1533        252        640:  34%|███▍      | 275/809 [05:29<10:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.24     0.4298     0.1534        341        640:  34%|███▍      | 277/809 [05:31<10:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4299     0.1532        177        640:  34%|███▍      | 278/809 [05:32<10:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.239     0.4297     0.1532        434        640:  35%|███▍      | 280/809 [05:35<10:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4302     0.1528        194        640:  35%|███▍      | 282/809 [05:37<10:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4305     0.1526        148        640:  35%|███▍      | 283/809 [05:38<10:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4304     0.1529        395        640:  35%|███▌      | 285/809 [05:41<10:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4301     0.1535        287        640:  35%|███▌      | 287/809 [05:43<10:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.238     0.4299     0.1536        333        640:  36%|███▌      | 288/809 [05:44<10:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.237     0.4306     0.1533        159        640:  36%|███▌      | 290/809 [05:47<10:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4309     0.1533        172        640:  36%|███▌      | 292/809 [05:49<10:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4309     0.1531        305        640:  36%|███▌      | 293/809 [05:50<10:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4309     0.1528        273        640:  36%|███▋      | 295/809 [05:53<10:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4307     0.1526        354        640:  37%|███▋      | 297/809 [05:55<10:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4311     0.1524        126        640:  37%|███▋      | 299/809 [05:57<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4308     0.1523        403        640:  37%|███▋      | 300/809 [05:59<10:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4306      0.152        211        640:  37%|███▋      | 302/809 [06:01<09:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.235     0.4305     0.1517        151        640:  38%|███▊      | 304/809 [06:03<10:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.236     0.4306     0.1519        172        640:  38%|███▊      | 305/809 [06:05<10:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4313     0.1516        178        640:  38%|███▊      | 307/809 [06:07<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4318     0.1514        256        640:  38%|███▊      | 309/809 [06:09<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4321     0.1513        182        640:  38%|███▊      | 310/809 [06:10<09:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4324     0.1512         81        640:  39%|███▊      | 312/809 [06:13<09:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4328     0.1509        241        640:  39%|███▉      | 314/809 [06:15<09:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4332     0.1507        119        640:  39%|███▉      | 315/809 [06:16<09:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4342     0.1503        120        640:  39%|███▉      | 317/809 [06:19<09:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4351       0.15        174        640:  39%|███▉      | 319/809 [06:21<09:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4355     0.1497        162        640:  40%|███▉      | 321/809 [06:23<09:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4357     0.1496        211        640:  40%|███▉      | 322/809 [06:24<09:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4357     0.1496        340        640:  40%|████      | 324/809 [06:27<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4363     0.1496        254        640:  40%|████      | 326/809 [06:29<09:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.218     0.4366     0.1495        130        640:  40%|████      | 327/809 [06:30<09:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4362     0.1496        321        640:  41%|████      | 329/809 [06:33<09:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4361     0.1495        154        640:  41%|████      | 331/809 [06:35<09:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4357     0.1497        467        640:  41%|████      | 332/809 [06:37<09:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4351     0.1504        183        640:  41%|████▏     | 334/809 [06:39<09:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4349     0.1504        230        640:  41%|████▏     | 335/809 [06:40<09:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4348     0.1512        283        640:  42%|████▏     | 337/809 [06:43<09:43,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4345     0.1521        248        640:  42%|████▏     | 339/809 [06:45<09:52,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224      0.435     0.1519        158        640:  42%|████▏     | 340/809 [06:46<09:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4348     0.1518        267        640:  42%|████▏     | 342/809 [06:49<09:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4351     0.1517        236        640:  43%|████▎     | 344/809 [06:51<09:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4355     0.1515        120        640:  43%|████▎     | 346/809 [06:53<09:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4357     0.1514        237        640:  43%|████▎     | 347/809 [06:55<09:41,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4359     0.1514        249        640:  43%|████▎     | 349/809 [06:57<09:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4359     0.1515        202        640:  43%|████▎     | 351/809 [07:00<09:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4361     0.1512        281        640:  44%|████▎     | 353/809 [07:02<09:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4363     0.1511        200        640:  44%|████▍     | 354/809 [07:03<09:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4364     0.1508        287        640:  44%|████▍     | 356/809 [07:05<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4367     0.1506        223        640:  44%|████▍     | 358/809 [07:08<08:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4367     0.1505        183        640:  44%|████▍     | 359/809 [07:09<08:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.218     0.4366     0.1505        289        640:  45%|████▍     | 361/809 [07:12<09:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4359     0.1509        281        640:  45%|████▍     | 363/809 [07:14<08:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4361     0.1508        286        640:  45%|████▍     | 364/809 [07:15<08:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4359     0.1517        512        640:  45%|████▌     | 366/809 [07:18<09:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22      0.436     0.1516        250        640:  45%|████▌     | 368/809 [07:20<08:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4361     0.1515        151        640:  46%|████▌     | 369/809 [07:21<08:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4364     0.1513        175        640:  46%|████▌     | 371/809 [07:24<08:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4366     0.1511        165        640:  46%|████▌     | 373/809 [07:26<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4364     0.1515        356        640:  46%|████▌     | 374/809 [07:27<08:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4362     0.1513        344        640:  46%|████▋     | 376/809 [07:29<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4357     0.1516        325        640:  47%|████▋     | 378/809 [07:32<08:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4358     0.1515        189        640:  47%|████▋     | 380/809 [07:34<08:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4355     0.1515        290        640:  47%|████▋     | 381/809 [07:35<08:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4349      0.152        332        640:  47%|████▋     | 383/809 [07:38<08:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4346     0.1521        305        640:  47%|████▋     | 384/809 [07:39<08:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4346     0.1518         84        640:  48%|████▊     | 386/809 [07:41<08:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4351     0.1516        119        640:  48%|████▊     | 388/809 [07:44<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4355     0.1516        193        640:  48%|████▊     | 390/809 [07:46<08:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4357     0.1515        148        640:  48%|████▊     | 391/809 [07:47<08:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4354     0.1516        380        640:  49%|████▊     | 393/809 [07:50<08:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4356     0.1515         97        640:  49%|████▉     | 395/809 [07:52<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4355     0.1513        144        640:  49%|████▉     | 396/809 [07:53<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4355     0.1512        228        640:  49%|████▉     | 398/809 [07:55<08:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4359     0.1509        295        640:  49%|████▉     | 400/809 [07:58<08:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219      0.436     0.1508        102        640:  50%|████▉     | 401/809 [07:59<08:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4354      0.151        391        640:  50%|████▉     | 403/809 [08:02<08:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4346      0.151        185        640:  50%|█████     | 405/809 [08:04<08:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4344      0.151        247        640:  50%|█████     | 406/809 [08:05<08:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224      0.434     0.1511        299        640:  50%|█████     | 408/809 [08:08<08:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224      0.434     0.1509        298        640:  51%|█████     | 410/809 [08:10<07:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4341     0.1508        253        640:  51%|█████     | 411/809 [08:11<07:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4339     0.1506        334        640:  51%|█████     | 413/809 [08:13<07:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4334     0.1509        252        640:  51%|█████▏    | 415/809 [08:16<07:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4332     0.1509        405        640:  51%|█████▏    | 416/809 [08:17<08:04,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4336     0.1506        108        640:  52%|█████▏    | 418/809 [08:20<07:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4334     0.1507        225        640:  52%|█████▏    | 420/809 [08:22<07:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4341     0.1505        121        640:  52%|█████▏    | 422/809 [08:24<07:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4338     0.1509        537        640:  52%|█████▏    | 423/809 [08:26<07:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4335     0.1508        522        640:  53%|█████▎    | 425/809 [08:28<07:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4331     0.1509        300        640:  53%|█████▎    | 427/809 [08:30<07:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4331     0.1508        204        640:  53%|█████▎    | 428/809 [08:31<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4333     0.1507        288        640:  53%|█████▎    | 430/809 [08:34<07:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4333     0.1507        341        640:  53%|█████▎    | 432/809 [08:36<07:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4333     0.1506        152        640:  54%|█████▎    | 433/809 [08:37<07:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4332     0.1504        285        640:  54%|█████▍    | 435/809 [08:40<07:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4329     0.1504        382        640:  54%|█████▍    | 437/809 [08:42<07:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4326     0.1504        321        640:  54%|█████▍    | 438/809 [08:44<07:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4325     0.1504        154        640:  54%|█████▍    | 440/809 [08:46<07:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4325     0.1503        402        640:  55%|█████▍    | 442/809 [08:48<07:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4327     0.1501        138        640:  55%|█████▍    | 443/809 [08:49<07:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4327       0.15        371        640:  55%|█████▌    | 445/809 [08:52<07:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4324       0.15        260        640:  55%|█████▌    | 447/809 [08:54<07:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4325     0.1499        187        640:  55%|█████▌    | 448/809 [08:55<07:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4324     0.1498        286        640:  56%|█████▌    | 450/809 [08:58<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4322     0.1497        191        640:  56%|█████▌    | 452/809 [09:00<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4321     0.1497        285        640:  56%|█████▌    | 453/809 [09:01<06:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4323     0.1496        254        640:  56%|█████▌    | 455/809 [09:04<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4324     0.1494        178        640:  56%|█████▋    | 457/809 [09:06<07:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4326     0.1495        200        640:  57%|█████▋    | 458/809 [09:07<07:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4325     0.1495        211        640:  57%|█████▋    | 460/809 [09:10<06:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225      0.433     0.1494        161        640:  57%|█████▋    | 462/809 [09:12<06:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4335     0.1492        121        640:  57%|█████▋    | 464/809 [09:14<06:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4332     0.1492        405        640:  57%|█████▋    | 465/809 [09:16<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4327     0.1495        351        640:  58%|█████▊    | 467/809 [09:18<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4328     0.1495        342        640:  58%|█████▊    | 469/809 [09:20<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4326     0.1497        288        640:  58%|█████▊    | 470/809 [09:21<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4324     0.1496        248        640:  58%|█████▊    | 472/809 [09:24<06:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4326     0.1494        302        640:  59%|█████▊    | 474/809 [09:26<06:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4329     0.1493        120        640:  59%|█████▊    | 475/809 [09:27<06:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4333     0.1492        180        640:  59%|█████▉    | 477/809 [09:30<06:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4336     0.1489        159        640:  59%|█████▉    | 479/809 [09:32<06:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4337     0.1487        356        640:  59%|█████▉    | 481/809 [09:34<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4336     0.1488        245        640:  60%|█████▉    | 482/809 [09:36<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4333     0.1487        187        640:  60%|█████▉    | 484/809 [09:38<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4329     0.1486        334        640:  60%|██████    | 486/809 [09:40<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4331     0.1485        237        640:  60%|██████    | 487/809 [09:41<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4333     0.1484        238        640:  60%|██████    | 489/809 [09:44<06:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4332     0.1482        278        640:  61%|██████    | 491/809 [09:46<06:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4329     0.1484        392        640:  61%|██████    | 492/809 [09:48<06:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4326     0.1484        340        640:  61%|██████    | 494/809 [09:50<06:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4326     0.1483        223        640:  61%|██████▏   | 496/809 [09:52<06:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4325     0.1482        264        640:  61%|██████▏   | 497/809 [09:53<06:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4327     0.1482        109        640:  62%|██████▏   | 499/809 [09:56<06:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4328      0.148        234        640:  62%|██████▏   | 501/809 [09:58<05:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.219     0.4328     0.1479        273        640:  62%|██████▏   | 503/809 [10:00<06:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.22     0.4326     0.1482        480        640:  62%|██████▏   | 504/809 [10:02<06:19,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4323     0.1484        339        640:  63%|██████▎   | 506/809 [10:04<06:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4321     0.1484        331        640:  63%|██████▎   | 508/809 [10:07<06:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221      0.432     0.1485        211        640:  63%|██████▎   | 509/809 [10:08<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4317     0.1486        380        640:  63%|██████▎   | 511/809 [10:10<05:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4317     0.1492        418        640:  63%|██████▎   | 513/809 [10:13<06:03,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4313     0.1494        334        640:  64%|██████▎   | 515/809 [10:15<05:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4312     0.1494        344        640:  64%|██████▍   | 516/809 [10:16<05:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4313     0.1493        235        640:  64%|██████▍   | 518/809 [10:19<05:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4314     0.1492        193        640:  64%|██████▍   | 520/809 [10:21<05:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4315     0.1491        259        640:  64%|██████▍   | 521/809 [10:22<05:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4317     0.1489        195        640:  65%|██████▍   | 523/809 [10:25<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4318     0.1488        317        640:  65%|██████▍   | 525/809 [10:27<05:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4319     0.1488        176        640:  65%|██████▌   | 526/809 [10:28<05:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4316     0.1489        253        640:  65%|██████▌   | 528/809 [10:31<05:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4316     0.1489        307        640:  66%|██████▌   | 530/809 [10:33<05:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4314     0.1489        216        640:  66%|██████▌   | 531/809 [10:34<05:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4313     0.1489        309        640:  66%|██████▌   | 533/809 [10:36<05:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4311      0.149        276        640:  66%|██████▌   | 535/809 [10:39<05:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4311      0.149        311        640:  66%|██████▋   | 536/809 [10:40<05:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4311     0.1488        124        640:  67%|██████▋   | 538/809 [10:42<05:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4308      0.149        373        640:  67%|██████▋   | 540/809 [10:45<05:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221      0.431     0.1488        191        640:  67%|██████▋   | 542/809 [10:47<05:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221      0.431     0.1488        247        640:  67%|██████▋   | 543/809 [10:48<05:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4307     0.1494        268        640:  67%|██████▋   | 545/809 [10:51<05:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4305     0.1501        388        640:  67%|██████▋   | 546/809 [10:52<05:24,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4304       0.15        119        640:  68%|██████▊   | 548/809 [10:54<05:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4302     0.1499        192        640:  68%|██████▊   | 550/809 [10:57<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4304     0.1498        107        640:  68%|██████▊   | 552/809 [10:59<05:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4308     0.1498         59        640:  68%|██████▊   | 553/809 [11:00<05:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4309     0.1497        114        640:  69%|██████▊   | 555/809 [11:03<04:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222      0.431     0.1497         56        640:  69%|██████▉   | 557/809 [11:05<04:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4312     0.1496        274        640:  69%|██████▉   | 558/809 [11:06<04:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4314     0.1495        265        640:  69%|██████▉   | 560/809 [11:09<04:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4312     0.1494        268        640:  69%|██████▉   | 562/809 [11:11<04:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4312     0.1493        197        640:  70%|██████▉   | 564/809 [11:13<04:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4309     0.1494        274        640:  70%|██████▉   | 565/809 [11:14<04:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4305     0.1494        261        640:  70%|███████   | 567/809 [11:17<04:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4303     0.1494        229        640:  70%|███████   | 568/809 [11:18<04:58,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4302     0.1493        204        640:  70%|███████   | 570/809 [11:21<04:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4301     0.1493        367        640:  71%|███████   | 572/809 [11:23<04:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4302     0.1493        160        640:  71%|███████   | 573/809 [11:24<04:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4302     0.1492        206        640:  71%|███████   | 575/809 [11:27<04:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225       0.43     0.1492        393        640:  71%|███████▏  | 577/809 [11:29<04:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4299     0.1492        263        640:  71%|███████▏  | 578/809 [11:30<04:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4299     0.1492        228        640:  72%|███████▏  | 580/809 [11:33<04:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225       0.43     0.1493        107        640:  72%|███████▏  | 582/809 [11:35<04:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4301     0.1492        229        640:  72%|███████▏  | 583/809 [11:36<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4305      0.149        145        640:  72%|███████▏  | 585/809 [11:38<04:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4304      0.149        303        640:  73%|███████▎  | 587/809 [11:41<04:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4304      0.149        180        640:  73%|███████▎  | 589/809 [11:43<04:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4304     0.1489        305        640:  73%|███████▎  | 590/809 [11:44<04:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4303     0.1496        372        640:  73%|███████▎  | 592/809 [11:47<04:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4306     0.1494        128        640:  73%|███████▎  | 594/809 [11:49<04:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4308     0.1493        206        640:  74%|███████▎  | 595/809 [11:50<04:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4308     0.1492        135        640:  74%|███████▍  | 597/809 [11:53<04:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4309     0.1491        253        640:  74%|███████▍  | 599/809 [11:55<04:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4308     0.1491        261        640:  74%|███████▍  | 600/809 [11:56<04:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4309     0.1491         80        640:  74%|███████▍  | 602/809 [11:59<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4307     0.1491        409        640:  75%|███████▍  | 604/809 [12:01<04:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4309      0.149        224        640:  75%|███████▍  | 605/809 [12:02<03:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4307      0.149        343        640:  75%|███████▌  | 607/809 [12:05<03:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4308     0.1489        233        640:  75%|███████▌  | 609/809 [12:07<03:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4308      0.149        372        640:  76%|███████▌  | 611/809 [12:09<03:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4307     0.1489        196        640:  76%|███████▌  | 612/809 [12:11<03:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.221     0.4307     0.1491         88        640:  76%|███████▌  | 614/809 [12:13<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.222     0.4304     0.1493        197        640:  76%|███████▌  | 616/809 [12:15<03:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4301     0.1494        291        640:  76%|███████▋  | 617/809 [12:17<03:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4299     0.1493        111        640:  77%|███████▋  | 619/809 [12:19<03:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4299     0.1493        208        640:  77%|███████▋  | 621/809 [12:21<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4297     0.1493        312        640:  77%|███████▋  | 622/809 [12:22<03:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4297     0.1493        270        640:  77%|███████▋  | 624/809 [12:25<03:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.223     0.4298     0.1494        190        640:  77%|███████▋  | 626/809 [12:27<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4297     0.1494        372        640:  78%|███████▊  | 627/809 [12:28<03:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4296     0.1498        261        640:  78%|███████▊  | 629/809 [12:31<03:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4297     0.1497        314        640:  78%|███████▊  | 631/809 [12:33<03:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.224     0.4296     0.1497        238        640:  78%|███████▊  | 632/809 [12:34<03:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.225     0.4293       0.15        371        640:  78%|███████▊  | 634/809 [12:37<03:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4292     0.1502        336        640:  79%|███████▊  | 636/809 [12:39<03:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4291     0.1503        285        640:  79%|███████▊  | 637/809 [12:41<03:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4291     0.1503        416        640:  79%|███████▉  | 639/809 [12:43<03:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227      0.429     0.1503        186        640:  79%|███████▉  | 641/809 [12:45<03:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4288     0.1507        455        640:  79%|███████▉  | 642/809 [12:47<03:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227      0.429     0.1505        179        640:  80%|███████▉  | 644/809 [12:49<03:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4288     0.1504        246        640:  80%|███████▉  | 646/809 [12:51<03:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4286     0.1506        379        640:  80%|███████▉  | 647/809 [12:52<03:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4283      0.151        457        640:  80%|████████  | 649/809 [12:55<03:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4281     0.1511        433        640:  80%|████████  | 651/809 [12:57<03:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4282     0.1511        175        640:  81%|████████  | 652/809 [12:59<03:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4282      0.151         88        640:  81%|████████  | 654/809 [13:01<03:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23      0.428      0.151        306        640:  81%|████████  | 656/809 [13:03<03:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4279     0.1511        297        640:  81%|████████  | 657/809 [13:05<03:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4279     0.1511        197        640:  81%|████████▏ | 659/809 [13:07<02:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231      0.428     0.1511        103        640:  82%|████████▏ | 661/809 [13:09<02:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231      0.428      0.151        155        640:  82%|████████▏ | 663/809 [13:12<02:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4279      0.151        294        640:  82%|████████▏ | 664/809 [13:13<02:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4281      0.151         87        640:  82%|████████▏ | 666/809 [13:15<02:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4281      0.151        187        640:  83%|████████▎ | 668/809 [13:17<02:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4284     0.1509        101        640:  83%|████████▎ | 669/809 [13:19<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4285     0.1508        281        640:  83%|████████▎ | 671/809 [13:21<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4282     0.1509        420        640:  83%|████████▎ | 673/809 [13:23<02:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4282     0.1512        259        640:  83%|████████▎ | 674/809 [13:25<02:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4284     0.1512        183        640:  84%|████████▎ | 676/809 [13:27<02:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4286     0.1511        205        640:  84%|████████▍ | 678/809 [13:29<02:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4285     0.1512        292        640:  84%|████████▍ | 680/809 [13:32<02:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4284     0.1511        327        640:  84%|████████▍ | 682/809 [13:34<02:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4285      0.151        187        640:  84%|████████▍ | 683/809 [13:35<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4286     0.1509        194        640:  85%|████████▍ | 685/809 [13:38<02:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4283     0.1511        504        640:  85%|████████▍ | 687/809 [13:40<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4284      0.151        228        640:  85%|████████▌ | 688/809 [13:41<02:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4284     0.1509        352        640:  85%|████████▌ | 690/809 [13:44<02:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4286     0.1508        439        640:  86%|████████▌ | 692/809 [13:46<02:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4287     0.1507        197        640:  86%|████████▌ | 693/809 [13:47<02:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228      0.429     0.1505        134        640:  86%|████████▌ | 695/809 [13:50<02:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4291     0.1504        395        640:  86%|████████▌ | 697/809 [13:52<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4291     0.1504        356        640:  86%|████████▋ | 698/809 [13:53<02:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228      0.429     0.1503        150        640:  87%|████████▋ | 700/809 [13:56<02:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4289     0.1504        363        640:  87%|████████▋ | 702/809 [13:58<02:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4288     0.1506        249        640:  87%|████████▋ | 703/809 [13:59<02:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4289     0.1506        119        640:  87%|████████▋ | 705/809 [14:02<02:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4288     0.1505        259        640:  87%|████████▋ | 707/809 [14:04<02:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4288     0.1504        169        640:  88%|████████▊ | 709/809 [14:06<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4288     0.1504        217        640:  88%|████████▊ | 710/809 [14:07<01:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4288     0.1504        391        640:  88%|████████▊ | 712/809 [14:10<02:04,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4288     0.1504        212        640:  88%|████████▊ | 713/809 [14:11<01:58,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4288     0.1505        182        640:  88%|████████▊ | 715/809 [14:14<01:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4287     0.1506        150        640:  89%|████████▊ | 717/809 [14:16<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4285     0.1509        430        640:  89%|████████▉ | 718/809 [14:17<01:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4286     0.1507        199        640:  89%|████████▉ | 720/809 [14:20<01:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4287     0.1506        252        640:  89%|████████▉ | 722/809 [14:22<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4291     0.1505        175        640:  89%|████████▉ | 724/809 [14:24<01:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4291     0.1504        315        640:  90%|████████▉ | 725/809 [14:25<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4292     0.1504        198        640:  90%|████████▉ | 727/809 [14:28<01:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4294     0.1503         91        640:  90%|█████████ | 729/809 [14:30<01:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4294     0.1502        186        640:  90%|█████████ | 730/809 [14:31<01:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4291     0.1503        314        640:  90%|█████████ | 732/809 [14:34<01:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4291     0.1503        281        640:  91%|█████████ | 734/809 [14:36<01:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227      0.429     0.1503        230        640:  91%|█████████ | 735/809 [14:37<01:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228      0.429     0.1509        395        640:  91%|█████████ | 737/809 [14:40<01:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4293     0.1508        308        640:  91%|█████████▏| 739/809 [14:42<01:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4294     0.1507        153        640:  91%|█████████▏| 740/809 [14:43<01:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4297     0.1505        102        640:  92%|█████████▏| 742/809 [14:46<01:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4298     0.1507        149        640:  92%|█████████▏| 744/809 [14:48<01:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4296     0.1511        478        640:  92%|█████████▏| 745/809 [14:49<01:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4297     0.1511        222        640:  92%|█████████▏| 747/809 [14:52<01:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4297      0.151         56        640:  93%|█████████▎| 749/809 [14:54<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4298     0.1509        228        640:  93%|█████████▎| 751/809 [14:56<01:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4297     0.1509        186        640:  93%|█████████▎| 752/809 [14:57<01:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.226     0.4296      0.151        202        640:  93%|█████████▎| 754/809 [15:00<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4294     0.1517        454        640:  93%|█████████▎| 756/809 [15:02<01:05,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4293     0.1517        311        640:  94%|█████████▎| 757/809 [15:04<01:04,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4297     0.1517         93        640:  94%|█████████▍| 759/809 [15:06<01:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4296     0.1516        257        640:  94%|█████████▍| 761/809 [15:08<00:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.227     0.4294     0.1519        399        640:  94%|█████████▍| 762/809 [15:10<00:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.228     0.4293     0.1518        319        640:  94%|█████████▍| 764/809 [15:12<00:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4291     0.1519        178        640:  95%|█████████▍| 766/809 [15:14<00:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229      0.429     0.1519        274        640:  95%|█████████▍| 767/809 [15:16<00:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.229     0.4287      0.152        354        640:  95%|█████████▌| 769/809 [15:18<00:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G       1.23     0.4285     0.1524        461        640:  95%|█████████▌| 770/809 [15:19<00:48,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4284     0.1524        312        640:  95%|█████████▌| 772/809 [15:22<00:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4283     0.1524        256        640:  96%|█████████▌| 774/809 [15:24<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.231     0.4281     0.1528        586        640:  96%|█████████▌| 775/809 [15:25<00:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4279     0.1528        243        640:  96%|█████████▌| 777/809 [15:28<00:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232      0.428     0.1528        225        640:  96%|█████████▋| 779/809 [15:30<00:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4279     0.1528        221        640:  97%|█████████▋| 781/809 [15:32<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.232     0.4278     0.1529        322        640:  97%|█████████▋| 782/809 [15:34<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4276     0.1529        185        640:  97%|█████████▋| 784/809 [15:36<00:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4274     0.1528        217        640:  97%|█████████▋| 786/809 [15:38<00:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4273     0.1531        507        640:  97%|█████████▋| 787/809 [15:40<00:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4273     0.1532        327        640:  98%|█████████▊| 789/809 [15:42<00:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4273     0.1532        229        640:  98%|█████████▊| 790/809 [15:44<00:23,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4273     0.1532        382        640:  98%|█████████▊| 792/809 [15:46<00:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4272     0.1531        249        640:  98%|█████████▊| 794/809 [15:48<00:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4271     0.1532        264        640:  98%|█████████▊| 795/809 [15:50<00:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4273     0.1531        177        640:  99%|█████████▊| 797/809 [15:52<00:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4272     0.1532        369        640:  99%|█████████▉| 799/809 [15:54<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4274     0.1531        103        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.233     0.4272     0.1532        382        640:  99%|█████████▉| 802/809 [15:58<00:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234      0.427     0.1532        285        640:  99%|█████████▉| 804/809 [16:00<00:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4268     0.1532        311        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     26/300      13.5G      1.234     0.4269     0.1531        125        640: 100%|█████████▉| 807/809 [16:04<00:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   1%|▏         | 1/69 [00:00<00:19,  3.51it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  13%|█▎        | 9/69 [00:02<00:15,  3.96it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  25%|██▍       | 17/69 [00:04<00:13,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  36%|███▌      | 25/69 [00:06<00:11,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:09,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:07,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:05,  3.57it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.486     0.3847     0.1952        289        640:   0%|          | 1/809 [00:01<16:32,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.099     0.4885     0.1383        177        640:   0%|          | 3/809 [00:03<15:47,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.302     0.4263     0.1946        353        640:   1%|          | 5/809 [00:06<16:22,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.207     0.4532     0.1742        102        640:   1%|          | 6/809 [00:07<16:05,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.265     0.4312     0.1773        217        640:   1%|          | 8/809 [00:09<16:08,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4252     0.1756        167        640:   1%|          | 10/809 [00:12<15:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4328     0.1695        209        640:   1%|▏         | 11/809 [00:13<15:58,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4237      0.169        255        640:   2%|▏         | 13/809 [00:15<15:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.24     0.4281     0.1658        240        640:   2%|▏         | 15/809 [00:18<16:09,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4301     0.1612         94        640:   2%|▏         | 16/809 [00:19<15:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.211     0.4339     0.1543        106        640:   2%|▏         | 18/809 [00:21<15:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.235     0.4279     0.1569        440        640:   2%|▏         | 20/809 [00:24<15:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4245     0.1559        312        640:   3%|▎         | 22/809 [00:26<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4336     0.1513        197        640:   3%|▎         | 24/809 [00:28<15:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4317     0.1506        251        640:   3%|▎         | 25/809 [00:29<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.208     0.4344     0.1481        179        640:   3%|▎         | 27/809 [00:32<15:36,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.197      0.437     0.1449        150        640:   4%|▎         | 29/809 [00:34<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.208     0.4348     0.1475        281        640:   4%|▎         | 30/809 [00:35<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.201     0.4349     0.1461        186        640:   4%|▍         | 32/809 [00:38<15:29,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4287      0.146        263        640:   4%|▍         | 34/809 [00:40<15:35,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226      0.426     0.1464        187        640:   4%|▍         | 35/809 [00:41<15:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4242     0.1458        216        640:   5%|▍         | 37/809 [00:44<15:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4205     0.1485        189        640:   5%|▍         | 39/809 [00:46<15:30,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4211     0.1479        144        640:   5%|▍         | 40/809 [00:47<15:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.235     0.4245     0.1473        356        640:   5%|▌         | 42/809 [00:50<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4253     0.1449        233        640:   5%|▌         | 44/809 [00:52<14:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227      0.427      0.146        240        640:   6%|▌         | 46/809 [00:54<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4262     0.1459        298        640:   6%|▌         | 47/809 [00:56<15:07,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4299     0.1441        187        640:   6%|▌         | 49/809 [00:58<14:43,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4325     0.1431        252        640:   6%|▋         | 51/809 [01:00<14:40,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4304     0.1438        380        640:   6%|▋         | 52/809 [01:01<14:48,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4283     0.1445        431        640:   7%|▋         | 54/809 [01:04<14:52,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225       0.43     0.1445        130        640:   7%|▋         | 56/809 [01:06<14:51,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22      0.432     0.1438        274        640:   7%|▋         | 58/809 [01:08<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229       0.43     0.1482        570        640:   7%|▋         | 59/809 [01:10<15:03,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4303      0.148        316        640:   8%|▊         | 61/809 [01:12<14:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4299     0.1482        288        640:   8%|▊         | 62/809 [01:14<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4301     0.1478        156        640:   8%|▊         | 64/809 [01:16<14:40,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4314     0.1471        241        640:   8%|▊         | 66/809 [01:18<14:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4289     0.1481        247        640:   8%|▊         | 68/809 [01:20<14:34,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.236     0.4287     0.1475        238        640:   9%|▊         | 69/809 [01:21<14:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4273     0.1475        242        640:   9%|▉         | 71/809 [01:24<14:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4239      0.148        332        640:   9%|▉         | 73/809 [01:26<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4238     0.1476        149        640:   9%|▉         | 74/809 [01:28<15:11,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.254     0.4221     0.1479        106        640:   9%|▉         | 76/809 [01:30<14:41,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.261     0.4202     0.1486        309        640:  10%|▉         | 78/809 [01:32<14:58,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.259     0.4204      0.148        192        640:  10%|▉         | 79/809 [01:34<14:56,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.264     0.4191     0.1491        382        640:  10%|█         | 81/809 [01:36<14:42,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.26     0.4206     0.1483        113        640:  10%|█         | 82/809 [01:38<15:57,  1.32s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.26     0.4215     0.1481        208        640:  10%|█         | 84/809 [01:40<14:58,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.255     0.4227     0.1476        174        640:  11%|█         | 86/809 [01:42<14:26,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.254     0.4233     0.1471        221        640:  11%|█         | 87/809 [01:43<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.261     0.4207     0.1482        344        640:  11%|█         | 89/809 [01:46<14:25,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.265     0.4194     0.1477        212        640:  11%|█         | 91/809 [01:48<14:26,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.27     0.4173     0.1489        324        640:  11%|█▏        | 92/809 [01:49<14:27,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.271     0.4161     0.1487        181        640:  12%|█▏        | 94/809 [01:52<14:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.271     0.4152     0.1482        259        640:  12%|█▏        | 96/809 [01:54<14:05,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.271     0.4153     0.1478        142        640:  12%|█▏        | 97/809 [01:55<14:15,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.267      0.416     0.1471        296        640:  12%|█▏        | 99/809 [01:58<14:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.265     0.4168     0.1466        228        640:  12%|█▏        | 101/809 [02:00<14:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.265     0.4168     0.1464        300        640:  13%|█▎        | 102/809 [02:01<13:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.265     0.4163     0.1463        330        640:  13%|█▎        | 104/809 [02:04<14:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.26     0.4173     0.1455        391        640:  13%|█▎        | 106/809 [02:06<14:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.256      0.419     0.1459        269        640:  13%|█▎        | 108/809 [02:09<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.254     0.4197     0.1455        198        640:  13%|█▎        | 109/809 [02:10<13:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251       0.42     0.1452        308        640:  14%|█▎        | 111/809 [02:12<14:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.245     0.4217     0.1443        153        640:  14%|█▍        | 113/809 [02:15<13:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4207     0.1447        404        640:  14%|█▍        | 114/809 [02:16<13:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4203     0.1457        463        640:  14%|█▍        | 116/809 [02:18<14:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4201     0.1462        484        640:  14%|█▍        | 117/809 [02:19<14:14,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4197     0.1459        288        640:  15%|█▍        | 119/809 [02:22<14:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4212     0.1452         96        640:  15%|█▍        | 121/809 [02:24<13:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4209     0.1449        258        640:  15%|█▌        | 123/809 [02:27<13:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4203     0.1449        263        640:  15%|█▌        | 124/809 [02:28<13:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4198     0.1454        131        640:  16%|█▌        | 126/809 [02:30<13:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247      0.421     0.1447        214        640:  16%|█▌        | 128/809 [02:33<13:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4213     0.1446        425        640:  16%|█▌        | 129/809 [02:34<13:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.244     0.4222     0.1454        228        640:  16%|█▌        | 131/809 [02:36<13:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4226     0.1455        153        640:  16%|█▋        | 133/809 [02:38<13:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4238      0.145        213        640:  17%|█▋        | 134/809 [02:40<13:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4235     0.1449        108        640:  17%|█▋        | 136/809 [02:42<13:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4235     0.1453        200        640:  17%|█▋        | 138/809 [02:44<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4239     0.1454        177        640:  17%|█▋        | 140/809 [02:47<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4247     0.1455        167        640:  17%|█▋        | 141/809 [02:48<13:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.236     0.4256     0.1449        258        640:  18%|█▊        | 143/809 [02:50<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4256     0.1453        315        640:  18%|█▊        | 145/809 [02:53<13:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4252     0.1453        346        640:  18%|█▊        | 146/809 [02:54<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4249     0.1456         95        640:  18%|█▊        | 148/809 [02:56<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.238     0.4252     0.1454        257        640:  19%|█▊        | 150/809 [02:59<13:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.238     0.4252     0.1452        254        640:  19%|█▊        | 151/809 [03:00<13:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4254     0.1451        227        640:  19%|█▉        | 153/809 [03:02<13:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.236     0.4261     0.1448        260        640:  19%|█▉        | 155/809 [03:05<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4257     0.1461        338        640:  19%|█▉        | 156/809 [03:06<13:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.236     0.4261      0.146        282        640:  20%|█▉        | 158/809 [03:08<12:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.235     0.4264     0.1457        297        640:  20%|█▉        | 160/809 [03:11<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.236     0.4257     0.1458        368        640:  20%|█▉        | 161/809 [03:12<12:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.233     0.4265     0.1451        246        640:  20%|██        | 163/809 [03:14<12:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4278     0.1444        165        640:  20%|██        | 165/809 [03:16<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4282     0.1441        162        640:  21%|██        | 166/809 [03:18<12:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4287     0.1443        246        640:  21%|██        | 168/809 [03:20<12:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4293     0.1443        116        640:  21%|██        | 170/809 [03:22<12:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4295     0.1441        218        640:  21%|██        | 171/809 [03:24<12:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4292     0.1448        430        640:  21%|██▏       | 173/809 [03:26<12:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4292     0.1455        577        640:  22%|██▏       | 175/809 [03:29<13:13,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4295     0.1453        221        640:  22%|██▏       | 176/809 [03:30<13:15,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4295     0.1453        259        640:  22%|██▏       | 178/809 [03:32<12:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4297     0.1452        297        640:  22%|██▏       | 180/809 [03:35<12:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4294     0.1459        295        640:  22%|██▏       | 181/809 [03:36<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4295     0.1455        218        640:  23%|██▎       | 183/809 [03:38<12:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4292     0.1462        413        640:  23%|██▎       | 185/809 [03:40<12:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4279     0.1465        194        640:  23%|██▎       | 187/809 [03:43<12:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4275     0.1465        333        640:  23%|██▎       | 188/809 [03:44<12:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4271     0.1473        163        640:  23%|██▎       | 190/809 [03:46<12:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4278     0.1472        327        640:  24%|██▎       | 192/809 [03:49<12:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4287     0.1472        341        640:  24%|██▍       | 194/809 [03:51<12:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4287     0.1472        320        640:  24%|██▍       | 195/809 [03:52<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219      0.428     0.1476        298        640:  24%|██▍       | 197/809 [03:55<12:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4288     0.1471        267        640:  25%|██▍       | 199/809 [03:57<12:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22      0.428     0.1487        490        640:  25%|██▍       | 200/809 [03:59<12:33,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4273     0.1492        405        640:  25%|██▍       | 202/809 [04:01<12:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4278     0.1491        116        640:  25%|██▌       | 204/809 [04:03<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4278     0.1489        290        640:  25%|██▌       | 205/809 [04:04<11:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4282     0.1485        182        640:  26%|██▌       | 207/809 [04:07<12:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.215     0.4285     0.1482        266        640:  26%|██▌       | 209/809 [04:09<12:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.215     0.4287     0.1481        219        640:  26%|██▌       | 210/809 [04:11<12:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4287     0.1479        142        640:  26%|██▌       | 212/809 [04:13<11:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4289     0.1478        106        640:  26%|██▋       | 214/809 [04:15<11:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4281      0.149        513        640:  27%|██▋       | 215/809 [04:17<12:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.216     0.4284     0.1506        478        640:  27%|██▋       | 217/809 [04:19<12:12,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.215      0.429     0.1503        198        640:  27%|██▋       | 218/809 [04:20<11:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.212       0.43     0.1499        171        640:  27%|██▋       | 220/809 [04:23<11:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.21     0.4304     0.1496        107        640:  27%|██▋       | 222/809 [04:25<11:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.208     0.4314     0.1492        161        640:  28%|██▊       | 224/809 [04:27<11:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.208     0.4316     0.1493        263        640:  28%|██▊       | 225/809 [04:29<11:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.21     0.4314     0.1498        429        640:  28%|██▊       | 227/809 [04:31<11:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4304     0.1511        505        640:  28%|██▊       | 229/809 [04:33<11:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.216       0.43     0.1515        318        640:  28%|██▊       | 230/809 [04:35<11:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4312     0.1511        271        640:  29%|██▊       | 232/809 [04:37<11:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4312     0.1511        295        640:  29%|██▉       | 234/809 [04:39<11:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.216     0.4309     0.1514        359        640:  29%|██▉       | 235/809 [04:41<11:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.215     0.4309      0.151        135        640:  29%|██▉       | 237/809 [04:43<11:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.214     0.4311     0.1506        334        640:  30%|██▉       | 239/809 [04:45<11:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.216     0.4306     0.1509        304        640:  30%|██▉       | 240/809 [04:47<11:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4301     0.1509        278        640:  30%|██▉       | 242/809 [04:49<11:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4298     0.1507        171        640:  30%|███       | 244/809 [04:51<11:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4298     0.1507        240        640:  30%|███       | 245/809 [04:53<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4299     0.1505        189        640:  31%|███       | 247/809 [04:55<11:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4296     0.1513        523        640:  31%|███       | 249/809 [04:57<11:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4297     0.1514        281        640:  31%|███       | 250/809 [04:59<11:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4295      0.151        145        640:  31%|███       | 252/809 [05:01<11:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4289     0.1508        223        640:  31%|███▏      | 254/809 [05:03<10:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4289     0.1509        248        640:  32%|███▏      | 255/809 [05:05<11:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226      0.428      0.151        481        640:  32%|███▏      | 257/809 [05:07<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4273     0.1514        428        640:  32%|███▏      | 259/809 [05:09<11:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4272     0.1513        131        640:  32%|███▏      | 260/809 [05:10<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4262     0.1512        201        640:  32%|███▏      | 262/809 [05:13<10:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4265     0.1508        144        640:  33%|███▎      | 264/809 [05:15<10:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23      0.427     0.1505        292        640:  33%|███▎      | 266/809 [05:18<10:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4266     0.1506        584        640:  33%|███▎      | 267/809 [05:19<10:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4276     0.1502        261        640:  33%|███▎      | 269/809 [05:21<10:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4277     0.1501        283        640:  33%|███▎      | 271/809 [05:23<10:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4278     0.1501        338        640:  34%|███▎      | 272/809 [05:25<10:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4287     0.1499        101        640:  34%|███▍      | 274/809 [05:27<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4287     0.1497        406        640:  34%|███▍      | 276/809 [05:29<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4297     0.1495         57        640:  34%|███▍      | 277/809 [05:30<10:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4307     0.1492        244        640:  34%|███▍      | 279/809 [05:33<10:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4302     0.1498        323        640:  35%|███▍      | 281/809 [05:35<10:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4302     0.1497        138        640:  35%|███▍      | 283/809 [05:38<10:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227       0.43     0.1496        249        640:  35%|███▌      | 284/809 [05:39<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4301     0.1495        177        640:  35%|███▌      | 286/809 [05:41<10:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4298     0.1495        334        640:  36%|███▌      | 288/809 [05:44<10:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227       0.43     0.1493        340        640:  36%|███▌      | 289/809 [05:45<10:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4299     0.1492        324        640:  36%|███▌      | 291/809 [05:47<10:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4299     0.1493        268        640:  36%|███▌      | 293/809 [05:50<10:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4298     0.1501        202        640:  36%|███▋      | 294/809 [05:51<10:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4302     0.1498        190        640:  37%|███▋      | 296/809 [05:53<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225      0.431     0.1494        174        640:  37%|███▋      | 298/809 [05:56<10:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4312     0.1492        277        640:  37%|███▋      | 299/809 [05:57<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4305     0.1519        293        640:  37%|███▋      | 301/809 [05:59<10:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4306     0.1517        241        640:  37%|███▋      | 303/809 [06:02<10:28,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4303     0.1519        340        640:  38%|███▊      | 304/809 [06:03<10:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4303     0.1519        147        640:  38%|███▊      | 306/809 [06:05<09:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4295      0.152        411        640:  38%|███▊      | 308/809 [06:08<10:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4294     0.1519        292        640:  38%|███▊      | 309/809 [06:09<09:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.234     0.4284     0.1519        362        640:  38%|███▊      | 311/809 [06:11<10:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.234     0.4284     0.1517        310        640:  39%|███▊      | 313/809 [06:14<09:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.234     0.4285     0.1516        246        640:  39%|███▉      | 314/809 [06:15<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.233     0.4288     0.1513        176        640:  39%|███▉      | 316/809 [06:17<09:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4293     0.1509        326        640:  39%|███▉      | 318/809 [06:19<09:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4296     0.1506        101        640:  39%|███▉      | 319/809 [06:21<09:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4299     0.1505        271        640:  40%|███▉      | 321/809 [06:23<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4301     0.1502        438        640:  40%|███▉      | 323/809 [06:25<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4309     0.1499        135        640:  40%|████      | 325/809 [06:28<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4311     0.1497        137        640:  40%|████      | 326/809 [06:29<09:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4312     0.1497        176        640:  41%|████      | 328/809 [06:31<09:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226      0.431     0.1496        234        640:  41%|████      | 329/809 [06:33<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4317     0.1495        309        640:  41%|████      | 331/809 [06:35<09:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4323     0.1492         43        640:  41%|████      | 333/809 [06:37<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4327      0.149        344        640:  41%|████▏     | 335/809 [06:40<09:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4319     0.1492        423        640:  42%|████▏     | 337/809 [06:42<09:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4315     0.1491        166        640:  42%|████▏     | 338/809 [06:43<09:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225     0.4312     0.1492        229        640:  42%|████▏     | 340/809 [06:46<09:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4307     0.1494        150        640:  42%|████▏     | 342/809 [06:48<09:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4302     0.1493        267        640:  43%|████▎     | 344/809 [06:50<09:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228       0.43     0.1494        410        640:  43%|████▎     | 345/809 [06:52<09:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4304     0.1494        361        640:  43%|████▎     | 347/809 [06:54<09:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4312     0.1491        104        640:  43%|████▎     | 349/809 [06:56<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4312      0.149        237        640:  43%|████▎     | 350/809 [06:57<08:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4312     0.1491        324        640:  44%|████▎     | 352/809 [07:00<09:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226     0.4315     0.1489        250        640:  44%|████▍     | 354/809 [07:02<08:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4324     0.1487        132        640:  44%|████▍     | 355/809 [07:03<09:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4328     0.1485        113        640:  44%|████▍     | 357/809 [07:06<08:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4331     0.1484        363        640:  44%|████▍     | 359/809 [07:08<09:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4333     0.1483        162        640:  44%|████▍     | 360/809 [07:09<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4333     0.1484        250        640:  45%|████▍     | 362/809 [07:12<08:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4331     0.1485        169        640:  45%|████▍     | 364/809 [07:14<08:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225     0.4326     0.1489        342        640:  45%|████▌     | 365/809 [07:15<08:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4319     0.1493        349        640:  45%|████▌     | 367/809 [07:18<09:03,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4314     0.1497        566        640:  46%|████▌     | 369/809 [07:20<08:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4312     0.1502        457        640:  46%|████▌     | 370/809 [07:21<09:02,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4317       0.15        171        640:  46%|████▌     | 372/809 [07:24<08:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23      0.431     0.1506        226        640:  46%|████▌     | 374/809 [07:26<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4312     0.1505        107        640:  46%|████▋     | 375/809 [07:28<08:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4315     0.1504        140        640:  47%|████▋     | 377/809 [07:30<08:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4314     0.1505        237        640:  47%|████▋     | 379/809 [07:32<08:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4314     0.1504        170        640:  47%|████▋     | 380/809 [07:33<08:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4314     0.1503        189        640:  47%|████▋     | 382/809 [07:36<08:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229      0.431     0.1505        368        640:  47%|████▋     | 384/809 [07:38<08:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4312     0.1504        194        640:  48%|████▊     | 385/809 [07:40<08:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4307     0.1503        193        640:  48%|████▊     | 387/809 [07:42<08:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4308     0.1501        149        640:  48%|████▊     | 389/809 [07:44<08:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4307     0.1506        247        640:  48%|████▊     | 390/809 [07:45<08:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4304     0.1509        349        640:  48%|████▊     | 392/809 [07:48<08:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4311     0.1506        173        640:  49%|████▊     | 394/809 [07:50<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.227     0.4314     0.1505        230        640:  49%|████▉     | 395/809 [07:51<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226      0.432     0.1503        223        640:  49%|████▉     | 397/809 [07:54<08:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4325       0.15        197        640:  49%|████▉     | 399/809 [07:56<08:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4326     0.1499        142        640:  50%|████▉     | 401/809 [07:58<07:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4325     0.1498        266        640:  50%|████▉     | 402/809 [08:00<07:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4323     0.1497        311        640:  50%|████▉     | 404/809 [08:02<08:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4322     0.1495        165        640:  50%|█████     | 406/809 [08:04<08:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4321     0.1495        314        640:  50%|█████     | 407/809 [08:06<08:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4322     0.1496        320        640:  51%|█████     | 409/809 [08:08<07:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224      0.432     0.1497        262        640:  51%|█████     | 411/809 [08:10<07:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4325     0.1496        122        640:  51%|█████     | 412/809 [08:12<07:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4326     0.1495        194        640:  51%|█████     | 414/809 [08:14<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4327     0.1494        170        640:  51%|█████▏    | 416/809 [08:16<07:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4325     0.1495        406        640:  52%|█████▏    | 417/809 [08:17<07:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4324     0.1494        240        640:  52%|█████▏    | 419/809 [08:20<07:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4324     0.1493        219        640:  52%|█████▏    | 421/809 [08:22<07:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.224     0.4324      0.149        220        640:  52%|█████▏    | 423/809 [08:25<07:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4326     0.1489        202        640:  52%|█████▏    | 424/809 [08:26<07:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222      0.433     0.1487        225        640:  53%|█████▎    | 426/809 [08:28<07:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4334     0.1485        152        640:  53%|█████▎    | 428/809 [08:30<07:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4338     0.1483        153        640:  53%|█████▎    | 429/809 [08:32<07:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4341     0.1485        174        640:  53%|█████▎    | 431/809 [08:34<07:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4343     0.1482        169        640:  54%|█████▎    | 433/809 [08:36<07:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4347      0.148        113        640:  54%|█████▍    | 435/809 [08:39<07:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4349     0.1479        148        640:  54%|█████▍    | 436/809 [08:40<07:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218      0.435     0.1482        440        640:  54%|█████▍    | 438/809 [08:42<07:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218     0.4349     0.1481        155        640:  54%|█████▍    | 440/809 [08:45<07:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218      0.435      0.148        224        640:  55%|█████▍    | 441/809 [08:46<07:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218      0.435     0.1479        214        640:  55%|█████▍    | 443/809 [08:48<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4353     0.1477        143        640:  55%|█████▌    | 445/809 [08:50<07:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4352     0.1478        219        640:  55%|█████▌    | 446/809 [08:52<07:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.217     0.4352     0.1476        305        640:  55%|█████▌    | 448/809 [08:54<07:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.218      0.435     0.1479        393        640:  56%|█████▌    | 450/809 [08:56<07:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4345     0.1487        530        640:  56%|█████▌    | 451/809 [08:58<07:31,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4349     0.1485        175        640:  56%|█████▌    | 453/809 [09:00<07:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4349     0.1483        197        640:  56%|█████▌    | 455/809 [09:03<07:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4347     0.1485        365        640:  56%|█████▋    | 456/809 [09:04<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4345     0.1484        285        640:  57%|█████▋    | 458/809 [09:06<07:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4344     0.1484        173        640:  57%|█████▋    | 460/809 [09:08<06:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4342     0.1484        221        640:  57%|█████▋    | 461/809 [09:10<06:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4339     0.1488        486        640:  57%|█████▋    | 463/809 [09:12<07:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221      0.434     0.1488        188        640:  57%|█████▋    | 465/809 [09:15<06:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4341     0.1486        204        640:  58%|█████▊    | 466/809 [09:16<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221      0.434     0.1488        589        640:  58%|█████▊    | 468/809 [09:18<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4345     0.1487        295        640:  58%|█████▊    | 470/809 [09:20<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1487        257        640:  58%|█████▊    | 471/809 [09:22<06:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1486        278        640:  58%|█████▊    | 473/809 [09:24<06:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4344     0.1484         81        640:  59%|█████▊    | 475/809 [09:26<06:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4345     0.1483        110        640:  59%|█████▉    | 477/809 [09:29<06:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4346     0.1482        240        640:  59%|█████▉    | 478/809 [09:30<06:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4347     0.1481        292        640:  59%|█████▉    | 480/809 [09:32<06:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4349     0.1479        176        640:  60%|█████▉    | 482/809 [09:35<06:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4348     0.1479        187        640:  60%|█████▉    | 483/809 [09:36<06:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.219     0.4348     0.1478        235        640:  60%|█████▉    | 485/809 [09:38<06:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.22     0.4345     0.1479        266        640:  60%|██████    | 487/809 [09:41<06:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1483        262        640:  60%|██████    | 489/809 [09:43<06:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1483        214        640:  61%|██████    | 490/809 [09:44<06:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4344     0.1483         94        640:  61%|██████    | 492/809 [09:47<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1482        251        640:  61%|██████    | 494/809 [09:49<06:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.221     0.4343     0.1483        293        640:  61%|██████    | 495/809 [09:50<06:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4342     0.1484        267        640:  61%|██████▏   | 497/809 [09:53<06:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.222     0.4342     0.1484        479        640:  62%|██████▏   | 499/809 [09:55<06:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.223     0.4339     0.1487        289        640:  62%|██████▏   | 500/809 [09:56<06:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225     0.4334     0.1492        277        640:  62%|██████▏   | 502/809 [09:59<06:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225     0.4334      0.149        255        640:  62%|��█████▏   | 504/809 [10:01<06:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.225     0.4333     0.1491        289        640:  62%|██████▏   | 505/809 [10:02<06:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.226      0.433      0.149        259        640:  63%|██████▎   | 507/809 [10:05<05:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4327     0.1491        229        640:  63%|██████▎   | 509/809 [10:07<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.228     0.4325      0.149        190        640:  63%|██████▎   | 511/809 [10:09<05:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4323      0.149        224        640:  63%|██████▎   | 512/809 [10:11<05:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229      0.432     0.1494        322        640:  64%|██████▎   | 514/809 [10:13<05:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4315     0.1494        338        640:  64%|██████▍   | 516/809 [10:15<05:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4312     0.1495        284        640:  64%|██████▍   | 517/809 [10:17<05:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4313     0.1495        117        640:  64%|██████▍   | 519/809 [10:19<05:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4316     0.1494        189        640:  64%|██████▍   | 521/809 [10:21<05:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229      0.432     0.1493         66        640:  65%|██████▍   | 522/809 [10:22<05:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23      0.432     0.1495        189        640:  65%|██████▍   | 524/809 [10:25<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4321     0.1493        275        640:  65%|██████▌   | 526/809 [10:27<05:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4319     0.1494        406        640:  65%|██████▌   | 527/809 [10:28<05:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4319     0.1495        232        640:  65%|██████▌   | 529/809 [10:31<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23      0.432     0.1494        358        640:  66%|██████▌   | 531/809 [10:33<05:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4323     0.1493        141        640:  66%|██████▌   | 533/809 [10:35<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4324     0.1492        282        640:  66%|██████▌   | 534/809 [10:37<05:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4325     0.1493         78        640:  66%|██████▋   | 536/809 [10:39<05:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4327     0.1493        281        640:  67%|██████▋   | 538/809 [10:41<05:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4326     0.1492        241        640:  67%|██████▋   | 539/809 [10:43<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4324     0.1494        213        640:  67%|██████▋   | 541/809 [10:45<05:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4323     0.1493        198        640:  67%|██████▋   | 543/809 [10:47<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4322     0.1495        385        640:  67%|██████▋   | 544/809 [10:48<05:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.23     0.4323     0.1495        276        640:  67%|██████▋   | 546/809 [10:51<05:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4323     0.1495        302        640:  68%|██████▊   | 548/809 [10:53<05:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4325     0.1494        221        640:  68%|██████▊   | 549/809 [10:54<05:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4322     0.1495        128        640:  68%|██████▊   | 551/809 [10:57<05:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.229     0.4321     0.1495        176        640:  68%|██████▊   | 553/809 [10:59<05:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.231     0.4316     0.1502        361        640:  68%|██████▊   | 554/809 [11:00<05:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4313     0.1507        534        640:  69%|██████▊   | 556/809 [11:03<05:13,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4312     0.1507        407        640:  69%|██████▉   | 558/809 [11:05<05:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.232     0.4311     0.1508        353        640:  69%|██████▉   | 559/809 [11:07<05:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.234     0.4306      0.151        224        640:  69%|██████▉   | 561/809 [11:09<05:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.235     0.4301     0.1511        210        640:  70%|██████▉   | 563/809 [11:11<04:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.234     0.4302     0.1511        152        640:  70%|██████▉   | 564/809 [11:12<04:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4293     0.1517        414        640:  70%|██████▉   | 566/809 [11:15<04:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4291     0.1518        104        640:  70%|███████   | 568/809 [11:17<04:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.237     0.4289     0.1518        179        640:  70%|███████   | 569/809 [11:19<04:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.238     0.4285     0.1519        259        640:  71%|███████   | 571/809 [11:21<04:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4284     0.1518        232        640:  71%|███████   | 573/809 [11:23<04:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4282      0.152        442        640:  71%|███████   | 574/809 [11:25<04:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.24      0.428     0.1522        177        640:  71%|███████   | 576/809 [11:27<04:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.24     0.4278     0.1522        182        640:  71%|███████▏  | 578/809 [11:29<04:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4276     0.1522        336        640:  72%|███████▏  | 579/809 [11:31<04:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4275     0.1522        226        640:  72%|███████▏  | 581/809 [11:33<04:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4274     0.1521        126        640:  72%|███████▏  | 583/809 [11:35<04:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4274     0.1521        169        640:  72%|███████▏  | 584/809 [11:36<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4269     0.1523        275        640:  72%|███████▏  | 586/809 [11:39<04:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242      0.427     0.1523        142        640:  73%|███████▎  | 588/809 [11:41<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4269     0.1524        351        640:  73%|███████▎  | 590/809 [11:44<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4268     0.1524        191        640:  73%|███████▎  | 591/809 [11:45<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4265     0.1525        470        640:  73%|███████▎  | 593/809 [11:47<04:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.244     0.4264     0.1526        302        640:  74%|███████▎  | 595/809 [11:50<04:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4264     0.1525        184        640:  74%|███████▎  | 596/809 [11:51<04:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4264     0.1524        249        640:  74%|███████▍  | 598/809 [11:53<04:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4266     0.1524        309        640:  74%|███████▍  | 600/809 [11:56<04:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4265     0.1524        234        640:  74%|███████▍  | 601/809 [11:57<04:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4266     0.1523        232        640:  75%|███████▍  | 603/809 [11:59<04:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4264     0.1525        355        640:  75%|███████▍  | 605/809 [12:01<04:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4264     0.1525        293        640:  75%|███████▍  | 606/809 [12:03<04:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4264     0.1525        242        640:  75%|███████▌  | 608/809 [12:05<04:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4263     0.1525        319        640:  75%|███████▌  | 610/809 [12:07<03:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4263     0.1525        325        640:  76%|███████▌  | 611/809 [12:09<03:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4262     0.1525        201        640:  76%|███████▌  | 613/809 [12:11<03:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4263     0.1524        194        640:  76%|███████▌  | 615/809 [12:13<03:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4266     0.1523        129        640:  76%|███████▋  | 617/809 [12:16<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4267     0.1523        297        640:  76%|███████▋  | 618/809 [12:17<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4265     0.1527        418        640:  77%|███████▋  | 620/809 [12:19<03:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4267     0.1526        251        640:  77%|███████▋  | 622/809 [12:22<03:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4265     0.1526        354        640:  77%|█���█████▋  | 623/809 [12:23<03:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4267     0.1526        358        640:  77%|███████▋  | 625/809 [12:25<03:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4265     0.1526        335        640:  78%|███████▊  | 627/809 [12:28<03:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4266     0.1524        141        640:  78%|███████▊  | 629/809 [12:30<03:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4266     0.1524        308        640:  78%|███████▊  | 630/809 [12:31<03:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4266     0.1524        146        640:  78%|███████▊  | 632/809 [12:34<03:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4267     0.1525        256        640:  78%|███████▊  | 634/809 [12:36<03:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4268     0.1524        186        640:  78%|███████▊  | 635/809 [12:37<03:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.239     0.4273     0.1522         96        640:  79%|███████▊  | 637/809 [12:40<03:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.24     0.4272     0.1524        183        640:  79%|███████▉  | 639/809 [12:42<03:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.24     0.4271     0.1524        217        640:  79%|███████▉  | 640/809 [12:43<03:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241      0.427     0.1525        297        640:  79%|███████▉  | 642/809 [12:46<03:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241     0.4271     0.1526        288        640:  80%|███████▉  | 644/809 [12:48<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.241      0.427     0.1526        261        640:  80%|███████▉  | 645/809 [12:49<03:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.242     0.4267     0.1531        223        640:  80%|███████▉  | 647/809 [12:52<03:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.243     0.4263     0.1538        314        640:  80%|████████  | 649/809 [12:54<03:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.245     0.4259     0.1544        613        640:  80%|████████  | 650/809 [12:56<03:25,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.245     0.4258     0.1546        509        640:  81%|████████  | 652/809 [12:58<03:19,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.245     0.4257     0.1546        199        640:  81%|████████  | 653/809 [12:59<03:13,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.245     0.4258     0.1549        180        640:  81%|████████  | 655/809 [13:02<03:12,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4257     0.1552        325        640:  81%|████████  | 657/809 [13:04<03:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4255     0.1552        276        640:  81%|████████▏ | 658/809 [13:05<03:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4254     0.1552        406        640:  82%|████████▏ | 660/809 [13:08<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4254     0.1553        113        640:  82%|████████▏ | 662/809 [13:10<02:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4253     0.1552        202        640:  82%|████████▏ | 663/809 [13:11<02:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4251     0.1555        219        640:  82%|████████▏ | 665/809 [13:14<02:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248      0.425     0.1556        242        640:  82%|████████▏ | 667/809 [13:16<02:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4248     0.1556        306        640:  83%|████████▎ | 668/809 [13:17<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4244     0.1558        623        640:  83%|████████▎ | 670/809 [13:20<02:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4241     0.1559        375        640:  83%|████████▎ | 672/809 [13:22<02:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252      0.424     0.1559        159        640:  83%|████████▎ | 673/809 [13:23<02:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4239     0.1559         74        640:  83%|████████▎ | 675/809 [13:26<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4239     0.1559        180        640:  84%|████████▎ | 677/809 [13:28<02:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4239      0.156        268        640:  84%|████████▍ | 678/809 [13:30<02:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4238      0.156        261        640:  84%|████████▍ | 680/809 [13:32<02:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4237      0.156        353        640:  84%|████████▍ | 682/809 [13:34<02:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4238     0.1559        223        640:  84%|████████▍ | 683/809 [13:35<02:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4237     0.1558        357        640:  85%|████████▍ | 685/809 [13:38<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4237     0.1557        256        640:  85%|████████▍ | 687/809 [13:40<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4237     0.1556        318        640:  85%|████████▌ | 688/809 [13:41<02:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4239     0.1555         86        640:  85%|████████▌ | 690/809 [13:44<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251      0.424     0.1554         90        640:  86%|████████▌ | 692/809 [13:46<02:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4238     0.1554        278        640:  86%|████████▌ | 694/809 [13:48<02:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4238     0.1554        232        640:  86%|████████▌ | 695/809 [13:50<02:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.252     0.4238     0.1553        212        640:  86%|████████▌ | 697/809 [13:52<02:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4239     0.1551        228        640:  86%|████████▋ | 699/809 [13:54<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251      0.424     0.1551        281        640:  87%|████████▋ | 700/809 [13:56<02:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251      0.424     0.1551        174        640:  87%|████████▋ | 702/809 [13:58<02:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4239     0.1552        250        640:  87%|████████▋ | 704/809 [14:00<02:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4238     0.1552        180        640:  87%|████████▋ | 705/809 [14:02<02:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4241      0.155        127        640:  87%|████████▋ | 707/809 [14:04<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4241     0.1548        293        640:  88%|████████▊ | 709/809 [14:06<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4241     0.1548        199        640:  88%|████████▊ | 710/809 [14:07<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25      0.424     0.1548        229        640:  88%|████████▊ | 712/809 [14:10<01:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4241     0.1546        175        640:  88%|████████▊ | 714/809 [14:12<01:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4242     0.1545        319        640:  89%|████████▊ | 716/809 [14:15<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4244     0.1544        164        640:  89%|████████▊ | 717/809 [14:16<01:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25      0.424     0.1547        474        640:  89%|████████▉ | 719/809 [14:18<01:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4241     0.1546         75        640:  89%|████████▉ | 721/809 [14:21<01:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4239     0.1547        616        640:  89%|████████▉ | 722/809 [14:22<01:47,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4238     0.1546        162        640:  89%|████████▉ | 724/809 [14:24<01:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4239     0.1546        156        640:  90%|████████▉ | 726/809 [14:27<01:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4238     0.1546        252        640:  90%|████████▉ | 727/809 [14:28<01:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4235     0.1547        411        640:  90%|█████████ | 729/809 [14:30<01:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4234     0.1548        402        640:  90%|█████████ | 730/809 [14:31<01:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.251     0.4233     0.1548        147        640:  90%|█████████ | 732/809 [14:34<01:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4233     0.1547        231        640:  91%|█████████ | 734/809 [14:36<01:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4233     0.1547        381        640:  91%|█████████ | 736/809 [14:39<01:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4233     0.1547        183        640:  91%|█████████ | 737/809 [14:40<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4232     0.1548        228        640:  91%|█████████▏| 739/809 [14:42<01:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4232     0.1548        173        640:  92%|█████████▏| 741/809 [14:45<01:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4234     0.1547         93        640:  92%|█████████▏| 742/809 [14:46<01:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249     0.4236     0.1546        262        640:  92%|█████████▏| 744/809 [14:48<01:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4238     0.1546        297        640:  92%|█████████▏| 746/809 [14:51<01:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4238     0.1546        209        640:  92%|█████████▏| 747/809 [14:52<01:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4238     0.1546        398        640:  93%|█████████▎| 749/809 [14:54<01:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4241     0.1546        160        640:  93%|█████████▎| 751/809 [14:56<01:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4241     0.1546        155        640:  93%|█████████▎| 752/809 [14:58<01:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4241     0.1545        187        640:  93%|█████████▎| 754/809 [15:00<01:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4244     0.1544        162        640:  93%|█████████▎| 756/809 [15:02<01:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4243     0.1546        133        640:  94%|█████████▎| 758/809 [15:05<00:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4242     0.1546        424        640:  94%|█████████▍| 759/809 [15:06<00:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4241     0.1546        235        640:  94%|█████████▍| 761/809 [15:08<00:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4243     0.1545        211        640:  94%|█████████▍| 763/809 [15:11<00:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4242     0.1546        271        640:  94%|█████████▍| 764/809 [15:12<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4242     0.1545        250        640:  95%|█████████▍| 766/809 [15:14<00:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4242     0.1544        175        640:  95%|█████████▍| 768/809 [15:17<00:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4242     0.1545        341        640:  95%|█████████▌| 769/809 [15:18<00:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246      0.424     0.1545        400        640:  95%|█████████▌| 771/809 [15:20<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4241     0.1545        168        640:  96%|█████████▌| 773/809 [15:22<00:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246     0.4241     0.1545        218        640:  96%|█████████▌| 774/809 [15:24<00:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.246      0.424     0.1545        268        640:  96%|█████████▌| 776/809 [15:26<00:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4238     0.1546        355        640:  96%|█████████▌| 778/809 [15:28<00:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4238     0.1545        407        640:  96%|█████████▋| 780/809 [15:31<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4237     0.1545        143        640:  97%|█████████▋| 781/809 [15:32<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4236     0.1545        396        640:  97%|█████████▋| 783/809 [15:34<00:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.247     0.4235     0.1545        236        640:  97%|█████████▋| 784/809 [15:36<00:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4234     0.1545        325        640:  97%|█████████▋| 786/809 [15:38<00:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.248     0.4234     0.1544        196        640:  97%|█████████▋| 788/809 [15:40<00:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249      0.423     0.1549        379        640:  98%|█████████▊| 790/809 [15:43<00:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4229     0.1553        561        640:  98%|█████████▊| 791/809 [15:44<00:22,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G      1.249      0.423     0.1552        212        640:  98%|█████████▊| 793/809 [15:47<00:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4228     0.1553        438        640:  98%|█████████▊| 795/809 [15:49<00:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227     0.1553        208        640:  99%|█████████▊| 797/809 [15:51<00:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227     0.1553        333        640:  99%|█████████▊| 798/809 [15:53<00:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227     0.1552        287        640:  99%|█████████▉| 800/809 [15:55<00:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227     0.1551        205        640:  99%|█████████▉| 802/809 [15:57<00:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4226     0.1551        302        640:  99%|█████████▉| 803/809 [15:59<00:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4224      0.155        333        640: 100%|█████████▉| 805/809 [16:01<00:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227      0.155        421        640: 100%|█████████▉| 807/809 [16:03<00:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     27/300      14.2G       1.25     0.4227     0.1549        190        640: 100%|█████████▉| 808/809 [16:04<00:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   9%|▊         | 6/69 [00:01<00:15,  3.99it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  20%|██        | 14/69 [00:03<00:14,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  32%|███▏      | 22/69 [00:05<00:11,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  43%|████▎     | 30/69 [00:07<00:09,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  55%|█████▌    | 38/69 [00:09<00:07,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.69it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  77%|███████▋  | 53/69 [00:13<00:04,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  88%|████████▊ | 61/69 [00:15<00:02,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.6G     0.7778     0.5505    0.09189        132        640:   0%|          | 1/809 [00:01<15:47,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.6G     0.7778     0.5505    0.09189        132        640:   0%|          | 1/809 [00:01<15:47,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.6G      1.027     0.4824     0.1162        396        640:   0%|          | 3/809 [00:03<16:00,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.6G      1.078     0.4695     0.1172        354        640:   0%|          | 4/809 [00:04<15:58,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.078     0.4664     0.1184        239        640:   1%|          | 6/809 [00:07<16:06,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.052     0.4754     0.1153        148        640:   1%|          | 7/809 [00:08<16:28,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G     0.9857     0.4904     0.1148        183        640:   1%|          | 9/809 [00:10<16:09,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G     0.9664     0.5076     0.1131        109        640:   1%|▏         | 11/809 [00:13<15:49,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G     0.9718     0.5069     0.1162         73        640:   2%|▏         | 13/809 [00:15<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G     0.9658      0.511     0.1157         79        640:   2%|▏         | 14/809 [00:16<15:49,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.019     0.4966     0.1224        310        640:   2%|▏         | 16/809 [00:19<15:59,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.06     0.4842       0.13        484        640:   2%|▏         | 17/809 [00:20<16:07,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.115     0.4715     0.1474        492        640:   2%|▏         | 19/809 [00:22<16:16,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.135     0.4676     0.1479        245        640:   3%|▎         | 21/809 [00:25<15:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.135     0.4681     0.1473        195        640:   3%|▎         | 22/809 [00:26<15:56,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.119     0.4731      0.142        187        640:   3%|▎         | 24/809 [00:28<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.125     0.4715     0.1436        228        640:   3%|▎         | 26/809 [00:31<15:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.146     0.4635     0.1453        355        640:   3%|▎         | 27/809 [00:33<15:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.146     0.4628     0.1457        230        640:   4%|▎         | 29/809 [00:34<15:39,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.142     0.4635     0.1431        307        640:   4%|▍         | 31/809 [00:37<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.147     0.4598      0.144        310        640:   4%|▍         | 33/809 [00:39<15:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.152     0.4586     0.1476        259        640:   4%|▍         | 34/809 [00:40<15:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.161      0.453     0.1477        180        640:   4%|▍         | 36/809 [00:43<15:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.172     0.4485     0.1486        495        640:   5%|▍         | 38/809 [00:45<15:50,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.177     0.4459     0.1484        415        640:   5%|▍         | 39/809 [00:46<15:38,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.192      0.441     0.1518        492        640:   5%|▌         | 41/809 [00:49<15:41,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.192     0.4396     0.1507        246        640:   5%|▌         | 43/809 [00:51<15:15,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.185     0.4419     0.1496        136        640:   5%|▌         | 44/809 [00:52<15:07,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.187     0.4407     0.1504        378        640:   6%|▌         | 46/809 [00:55<15:17,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.194     0.4391     0.1512        313        640:   6%|▌         | 48/809 [00:57<15:02,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.194     0.4384     0.1511        281        640:   6%|▌         | 49/809 [00:58<15:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.195     0.4392     0.1525        150        640:   6%|▋         | 51/809 [01:01<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.192     0.4402     0.1521        172        640:   7%|▋         | 53/809 [01:03<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.189     0.4414     0.1508        169        640:   7%|▋         | 54/809 [01:04<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.205     0.4379     0.1548        413        640:   7%|▋         | 56/809 [01:07<14:49,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.204      0.439     0.1535        249        640:   7%|▋         | 58/809 [01:09<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.204     0.4393     0.1525        322        640:   7%|▋         | 60/809 [01:11<14:45,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.202     0.4408     0.1521        103        640:   8%|▊         | 61/809 [01:12<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.198     0.4426     0.1514        259        640:   8%|▊         | 63/809 [01:15<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.196     0.4432     0.1507        260        640:   8%|▊         | 65/809 [01:17<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.194     0.4431       0.15        200        640:   8%|▊         | 66/809 [01:18<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.205     0.4397     0.1529        361        640:   8%|▊         | 68/809 [01:21<14:46,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.204     0.4412     0.1519        124        640:   9%|▊         | 70/809 [01:23<14:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.205     0.4408     0.1514        215        640:   9%|▉         | 71/809 [01:24<14:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.205     0.4413     0.1508        462        640:   9%|▉         | 73/809 [01:27<14:45,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4387     0.1521        475        640:   9%|▉         | 75/809 [01:29<14:46,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.216     0.4377     0.1523        267        640:   9%|▉         | 76/809 [01:30<14:43,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4377     0.1523        315        640:  10%|▉         | 78/809 [01:33<14:54,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4374     0.1512        151        640:  10%|▉         | 80/809 [01:35<14:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4383     0.1503         95        640:  10%|█         | 81/809 [01:36<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222      0.438     0.1508        279        640:  10%|█         | 83/809 [01:39<14:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.218     0.4389     0.1497        192        640:  11%|█         | 85/809 [01:41<14:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4393      0.149        237        640:  11%|█         | 87/809 [01:43<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4391     0.1486        158        640:  11%|█         | 88/809 [01:45<14:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.216     0.4392     0.1486        298        640:  11%|█         | 90/809 [01:47<14:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4399     0.1486        509        640:  11%|█▏        | 92/809 [01:49<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4401     0.1483        291        640:  11%|█▏        | 93/809 [01:51<14:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.211     0.4406     0.1472        195        640:  12%|█▏        | 95/809 [01:53<14:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.217     0.4388     0.1491        463        640:  12%|█▏        | 97/809 [01:55<14:05,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.217     0.4391     0.1487        154        640:  12%|█▏        | 98/809 [01:56<13:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.216     0.4399     0.1487        132        640:  12%|█▏        | 100/809 [01:59<13:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4403     0.1485         96        640:  13%|█▎        | 102/809 [02:01<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22      0.439     0.1492        205        640:  13%|█▎        | 103/809 [02:02<14:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4401     0.1487        166        640:  13%|█▎        | 105/809 [02:05<13:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4389     0.1498        384        640:  13%|█▎        | 107/809 [02:07<14:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4378     0.1504        291        640:  13%|█▎        | 108/809 [02:08<14:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4363     0.1505        292        640:  14%|█▎        | 110/809 [02:11<14:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4366     0.1501         99        640:  14%|█▍        | 112/809 [02:13<13:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4368     0.1497        239        640:  14%|█▍        | 114/809 [02:15<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231      0.436     0.1498        309        640:  14%|█▍        | 115/809 [02:17<13:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4367     0.1495        131        640:  14%|█▍        | 117/809 [02:19<13:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4366     0.1491        243        640:  15%|█▍        | 118/809 [02:20<14:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4366     0.1492        585        640:  15%|█▍        | 120/809 [02:23<13:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4352     0.1503        456        640:  15%|█▌        | 122/809 [02:25<13:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.243     0.4336     0.1536        522        640:  15%|█▌        | 123/809 [02:27<14:25,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.246     0.4331      0.154        242        640:  15%|█▌        | 125/809 [02:29<14:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.245     0.4332     0.1531        215        640:  16%|█▌        | 127/809 [02:32<13:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.249     0.4321     0.1541        442        640:  16%|█▌        | 128/809 [02:33<14:08,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.246     0.4334     0.1541        165        640:  16%|█▌        | 130/809 [02:35<13:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.246     0.4337     0.1535        290        640:  16%|█▋        | 132/809 [02:38<13:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.242     0.4348     0.1533        133        640:  16%|█▋        | 133/809 [02:39<13:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4352      0.153        282        640:  17%|█▋        | 135/809 [02:41<13:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4354     0.1526        157        640:  17%|█▋        | 137/809 [02:44<13:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4364     0.1522        222        640:  17%|█▋        | 138/809 [02:45<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4382     0.1515         91        640:  17%|█▋        | 140/809 [02:47<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4377     0.1515        346        640:  18%|█▊        | 142/809 [02:49<13:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4373     0.1521        288        640:  18%|█▊        | 144/809 [02:52<13:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4373     0.1517        126        640:  18%|█▊        | 145/809 [02:53<13:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4361     0.1525        186        640:  18%|█▊        | 147/809 [02:55<13:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4358     0.1521        219        640:  18%|█▊        | 149/809 [02:58<13:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4359     0.1521        228        640:  19%|█▊        | 150/809 [02:59<13:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4366     0.1515        209        640:  19%|█▉        | 152/809 [03:01<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4366     0.1511        239        640:  19%|█▉        | 154/809 [03:04<13:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4358      0.151        405        640:  19%|█▉        | 156/809 [03:06<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4357     0.1513        480        640:  19%|█▉        | 157/809 [03:07<13:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4369     0.1509        134        640:  20%|█▉        | 159/809 [03:10<12:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4369     0.1508        284        640:  20%|█▉        | 160/809 [03:11<12:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4361     0.1516        181        640:  20%|██        | 162/809 [03:13<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4357      0.152        398        640:  20%|██        | 164/809 [03:16<12:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4364     0.1512        137        640:  21%|██        | 166/809 [03:18<12:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4367      0.151        230        640:  21%|██        | 167/809 [03:19<12:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.436     0.1512        264        640:  21%|██        | 169/809 [03:22<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4359     0.1514        223        640:  21%|██        | 171/809 [03:24<12:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4362     0.1512        227        640:  21%|██▏       | 172/809 [03:25<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.242     0.4356     0.1516        339        640:  22%|██▏       | 174/809 [03:28<12:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4369      0.151        140        640:  22%|██▏       | 176/809 [03:30<12:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4371     0.1506        179        640:  22%|██▏       | 177/809 [03:31<12:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4373     0.1508        349        640:  22%|██▏       | 179/809 [03:34<12:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4376     0.1508        386        640:  22%|██▏       | 181/809 [03:36<13:04,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4378     0.1508        278        640:  22%|██▏       | 182/809 [03:37<13:03,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4374     0.1504        183        640:  23%|██▎       | 184/809 [03:40<12:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4372     0.1501        164        640:  23%|██▎       | 186/809 [03:42<12:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4368     0.1499        216        640:  23%|██▎       | 187/809 [03:43<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4354     0.1508        431        640:  23%|██▎       | 189/809 [03:46<12:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4357     0.1505        255        640:  24%|██▎       | 191/809 [03:48<12:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237      0.436     0.1504        196        640:  24%|██▎       | 192/809 [03:49<12:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4362       0.15        264        640:  24%|██▍       | 194/809 [03:52<12:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4368     0.1497        331        640:  24%|██▍       | 196/809 [03:54<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4372     0.1495        338        640:  24%|██▍       | 197/809 [03:55<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4368     0.1499        248        640:  25%|██▍       | 199/809 [03:58<12:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4369     0.1496        143        640:  25%|██▍       | 201/809 [04:00<12:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4374     0.1494        158        640:  25%|██▍       | 202/809 [04:01<12:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4384     0.1491        100        640:  25%|██▌       | 204/809 [04:04<11:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4382     0.1488        223        640:  25%|██▌       | 206/809 [04:06<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4379     0.1486        153        640:  26%|██▌       | 208/809 [04:08<11:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4379     0.1486        280        640:  26%|██▌       | 209/809 [04:10<11:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4392     0.1483        122        640:  26%|██▌       | 211/809 [04:12<11:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222       0.44     0.1479        177        640:  26%|██▋       | 213/809 [04:14<11:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4395      0.148        284        640:  26%|██▋       | 214/809 [04:15<11:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4401     0.1476        264        640:  27%|██▋       | 216/809 [04:18<11:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4405     0.1475        119        640:  27%|██▋       | 218/809 [04:20<11:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4408     0.1473         56        640:  27%|██▋       | 219/809 [04:21<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4412     0.1472        180        640:  27%|██▋       | 221/809 [04:24<11:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4411     0.1473        176        640:  28%|██▊       | 223/809 [04:26<11:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.223     0.4404     0.1478        550        640:  28%|██▊       | 224/809 [04:27<11:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4404     0.1474        130        640:  28%|██▊       | 226/809 [04:30<11:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4398     0.1473        374        640:  28%|██▊       | 228/809 [04:32<11:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4396     0.1472        182        640:  28%|██▊       | 229/809 [04:33<11:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4389     0.1471        360        640:  29%|██▊       | 231/809 [04:36<11:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4391     0.1469        275        640:  29%|██▉       | 233/809 [04:38<11:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225      0.439     0.1468        187        640:  29%|██▉       | 234/809 [04:39<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4383     0.1468        437        640:  29%|██▉       | 236/809 [04:42<11:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4383     0.1466        121        640:  29%|██▉       | 238/809 [04:44<11:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4385     0.1464        116        640:  30%|██▉       | 239/809 [04:45<11:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4386     0.1462        112        640:  30%|██▉       | 241/809 [04:48<11:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4384      0.146        392        640:  30%|███       | 243/809 [04:50<11:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4378     0.1461        450        640:  30%|███       | 245/809 [04:52<11:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229      0.437     0.1466        415        640:  30%|███       | 246/809 [04:54<11:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4374     0.1466        364        640:  31%|███       | 248/809 [04:56<11:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4371      0.147        440        640:  31%|███       | 250/809 [04:58<11:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4371      0.147        327        640:  31%|███       | 251/809 [05:00<11:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4377     0.1468        213        640:  31%|███▏      | 253/809 [05:02<10:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4376     0.1465        204        640:  32%|███▏      | 255/809 [05:04<10:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4374     0.1466        240        640:  32%|███▏      | 256/809 [05:05<10:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227      0.437     0.1467        272        640:  32%|███▏      | 258/809 [05:08<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4367     0.1467        382        640:  32%|███▏      | 260/809 [05:10<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4361      0.147        265        640:  32%|███▏      | 261/809 [05:11<10:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4358      0.147        296        640:  33%|███▎      | 263/809 [05:14<10:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233      0.435     0.1471        316        640:  33%|███▎      | 265/809 [05:16<10:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4351     0.1471        126        640:  33%|███▎      | 266/809 [05:17<10:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4348      0.147        172        640:  33%|███▎      | 268/809 [05:20<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4341     0.1479        141        640:  33%|███▎      | 270/809 [05:22<11:08,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4342     0.1477        190        640:  33%|███▎      | 271/809 [05:24<10:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4342     0.1474        167        640:  34%|███▎      | 273/809 [05:26<10:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4346     0.1472        237        640:  34%|███▍      | 275/809 [05:28<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4345     0.1479        429        640:  34%|███▍      | 276/809 [05:29<10:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4348     0.1476        143        640:  34%|███▍      | 278/809 [05:32<10:52,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4351     0.1473        104        640:  35%|███▍      | 280/809 [05:34<10:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4351     0.1472        354        640:  35%|███▍      | 281/809 [05:35<10:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4359     0.1469        149        640:  35%|███▍      | 283/809 [05:38<10:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4358      0.147        234        640:  35%|███▌      | 285/809 [05:40<10:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4359     0.1474        340        640:  35%|███▌      | 287/809 [05:43<10:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4358     0.1475        398        640:  36%|███▌      | 288/809 [05:44<10:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4366     0.1472        111        640:  36%|███▌      | 290/809 [05:46<10:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4373     0.1468        408        640:  36%|███▌      | 292/809 [05:48<10:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225      0.437     0.1467        340        640:  36%|███▌      | 293/809 [05:50<10:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4383     0.1466        125        640:  36%|███▋      | 295/809 [05:52<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4391     0.1462         75        640:  37%|███▋      | 297/809 [05:54<09:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4402     0.1458        132        640:  37%|███▋      | 299/809 [05:57<09:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4396      0.147        241        640:  37%|███▋      | 301/809 [05:59<10:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4402     0.1469        112        640:  37%|███▋      | 302/809 [06:00<10:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4401     0.1467        289        640:  38%|███▊      | 304/809 [06:03<10:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4397     0.1465        152        640:  38%|███▊      | 306/809 [06:05<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.218     0.4399     0.1464        209        640:  38%|███▊      | 307/809 [06:06<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.217     0.4402     0.1461        265        640:  38%|███▊      | 309/809 [06:09<09:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.216     0.4405     0.1459        224        640:  38%|███▊      | 311/809 [06:11<09:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4409     0.1457        120        640:  39%|███▊      | 312/809 [06:12<09:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4408     0.1456        246        640:  39%|███▉      | 314/809 [06:14<09:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4411     0.1454        138        640:  39%|███▉      | 316/809 [06:17<09:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.213     0.4415     0.1453        139        640:  39%|███▉      | 317/809 [06:18<09:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4412     0.1451        220        640:  39%|███▉      | 319/809 [06:21<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4406     0.1453        189        640:  40%|███▉      | 321/809 [06:23<09:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.214     0.4412     0.1451         82        640:  40%|███▉      | 322/809 [06:24<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.215     0.4407     0.1457        335        640:  40%|████      | 324/809 [06:26<09:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.216     0.4402     0.1458        322        640:  40%|████      | 326/809 [06:29<09:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.217     0.4399     0.1459        281        640:  40%|████      | 327/809 [06:30<09:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.217     0.4396     0.1457        216        640:  41%|████      | 329/809 [06:32<09:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4392      0.146        228        640:  41%|████      | 331/809 [06:35<09:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4387     0.1464        458        640:  41%|████      | 332/809 [06:36<09:51,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4382     0.1473        160        640:  41%|████▏     | 334/809 [06:39<09:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4384     0.1475        539        640:  42%|████▏     | 336/809 [06:41<09:45,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4384     0.1474        210        640:  42%|████▏     | 337/809 [06:42<09:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4389     0.1473        198        640:  42%|████▏     | 339/809 [06:45<09:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.219     0.4395     0.1471        192        640:  42%|████▏     | 341/809 [06:47<09:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.218     0.4397      0.147        204        640:  42%|████▏     | 342/809 [06:48<09:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.218     0.4398      0.147        228        640:  43%|████▎     | 344/809 [06:51<09:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.218     0.4396     0.1469         93        640:  43%|████▎     | 346/809 [06:53<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4388     0.1475        536        640:  43%|████▎     | 348/809 [06:55<09:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4384     0.1479        478        640:  43%|████▎     | 349/809 [06:57<09:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.22     0.4386     0.1476        152        640:  43%|████▎     | 351/809 [06:59<09:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4385     0.1476        274        640:  44%|████▎     | 352/809 [07:00<09:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4383     0.1476        380        640:  44%|████▍     | 354/809 [07:03<09:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4381     0.1475        309        640:  44%|████▍     | 356/809 [07:05<09:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4383     0.1473        276        640:  44%|████▍     | 357/809 [07:06<08:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.221     0.4385      0.147        172        640:  44%|████▍     | 359/809 [07:09<08:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.223      0.438     0.1469        350        640:  45%|████▍     | 361/809 [07:11<08:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.223     0.4379     0.1468        184        640:  45%|████▍     | 363/809 [07:13<08:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.222     0.4382     0.1467        233        640:  45%|████▍     | 364/809 [07:14<08:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.224     0.4377     0.1468        268        640:  45%|████▌     | 366/809 [07:17<08:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4373      0.147        253        640:  45%|████▌     | 368/809 [07:19<08:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4375     0.1469        176        640:  46%|████▌     | 369/809 [07:20<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4376     0.1468         87        640:  46%|████▌     | 371/809 [07:23<08:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.226     0.4373     0.1467        353        640:  46%|████▌     | 373/809 [07:25<08:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.225     0.4375     0.1466        237        640:  46%|████▌     | 374/809 [07:26<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4371     0.1466        222        640:  46%|████▋     | 376/809 [07:29<08:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4371     0.1465        346        640:  47%|████▋     | 378/809 [07:31<08:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4375     0.1464         74        640:  47%|████▋     | 379/809 [07:32<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4372     0.1465        297        640:  47%|████▋     | 381/809 [07:35<08:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229      0.437     0.1464        255        640:  47%|████▋     | 383/809 [07:37<08:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4369     0.1465        284        640:  47%|████▋     | 384/809 [07:38<08:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4369     0.1465        262        640:  48%|████▊     | 386/809 [07:41<08:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4369     0.1463        303        640:  48%|████▊     | 388/809 [07:43<08:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4369     0.1462        217        640:  48%|████▊     | 389/809 [07:44<08:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4368     0.1462        316        640:  48%|████▊     | 391/809 [07:47<08:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4368      0.146        141        640:  49%|████▊     | 393/809 [07:49<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4364     0.1462        392        640:  49%|████▊     | 394/809 [07:50<08:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4365     0.1461        265        640:  49%|████▉     | 396/809 [07:53<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4365     0.1461        426        640:  49%|████▉     | 398/809 [07:55<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4361     0.1463        304        640:  49%|████▉     | 399/809 [07:56<08:29,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4361     0.1463         47        640:  50%|████▉     | 401/809 [07:59<08:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4358     0.1464        282        640:  50%|████▉     | 403/809 [08:01<08:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4359     0.1463        137        640:  50%|████▉     | 404/809 [08:02<08:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4354     0.1467        297        640:  50%|█████     | 406/809 [08:05<08:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4353     0.1466        233        640:  50%|█████     | 408/809 [08:07<07:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4354     0.1466        168        640:  51%|█████     | 409/809 [08:08<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4351     0.1468        374        640:  51%|█████     | 411/809 [08:11<07:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4356     0.1468        179        640:  51%|█████     | 413/809 [08:13<07:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4352     0.1473        288        640:  51%|█████     | 414/809 [08:14<08:04,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4348     0.1475        178        640:  51%|█████▏    | 416/809 [08:17<07:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4352     0.1473        225        640:  52%|█████▏    | 418/809 [08:19<07:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235      0.435     0.1473        378        640:  52%|█████▏    | 420/809 [08:22<07:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4352     0.1473        254        640:  52%|█████▏    | 421/809 [08:23<07:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4353     0.1472        262        640:  52%|█████▏    | 423/809 [08:25<07:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4351     0.1473        167        640:  53%|█████▎    | 425/809 [08:27<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4356     0.1472         74        640:  53%|█████▎    | 426/809 [08:29<07:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4359     0.1471        274        640:  53%|█████▎    | 428/809 [08:31<07:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4359      0.147        335        640:  53%|█████▎    | 430/809 [08:33<07:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4357      0.147        340        640:  53%|█████▎    | 431/809 [08:35<07:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4356     0.1469        238        640:  54%|█████▎    | 433/809 [08:37<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4357     0.1468        238        640:  54%|█████▍    | 435/809 [08:39<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4357     0.1468        234        640:  54%|█████▍    | 436/809 [08:41<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.232     0.4357     0.1468        288        640:  54%|█████▍    | 438/809 [08:43<07:41,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4361     0.1466        198        640:  54%|█████▍    | 440/809 [08:45<07:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4362     0.1466        152        640:  55%|█████▍    | 441/809 [08:47<07:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4365     0.1465        185        640:  55%|█████▍    | 443/809 [08:49<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4368     0.1463         96        640:  55%|█████▌    | 445/809 [08:51<07:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4368     0.1462        159        640:  55%|█████▌    | 446/809 [08:53<07:35,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227     0.4369     0.1461        199        640:  55%|█████▌    | 448/809 [08:55<07:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.227      0.437     0.1459        157        640:  56%|█████▌    | 450/809 [08:57<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4368     0.1464        379        640:  56%|█████▌    | 451/809 [08:59<07:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4367     0.1464        271        640:  56%|█████▌    | 453/809 [09:01<07:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.228     0.4366     0.1463        151        640:  56%|█████▌    | 455/809 [09:03<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4363     0.1466        189        640:  56%|█████▋    | 457/809 [09:06<06:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.229     0.4361     0.1465        287        640:  57%|█████▋    | 458/809 [09:07<06:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.23     0.4357     0.1475        391        640:  57%|█████▋    | 460/809 [09:09<06:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.231     0.4354     0.1478        253        640:  57%|█████▋    | 461/809 [09:10<06:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4349     0.1489        247        640:  57%|█████▋    | 463/809 [09:13<07:15,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4344     0.1506        298        640:  57%|█████▋    | 465/809 [09:16<07:13,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4344     0.1504        205        640:  58%|█████▊    | 467/809 [09:18<06:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4345     0.1504        143        640:  58%|█████▊    | 468/809 [09:19<06:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4344     0.1504        187        640:  58%|█████▊    | 470/809 [09:22<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4344     0.1502        287        640:  58%|█████▊    | 472/809 [09:24<06:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4341     0.1502        411        640:  58%|█████▊    | 473/809 [09:25<06:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4339     0.1501        295        640:  59%|█████▊    | 475/809 [09:28<06:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.233     0.4339       0.15        154        640:  59%|█████▉    | 477/809 [09:30<06:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4337       0.15        298        640:  59%|█████▉    | 478/809 [09:31<06:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4336     0.1507        199        640:  59%|█████▉    | 480/809 [09:33<06:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4331      0.151        206        640:  60%|█████▉    | 482/809 [09:36<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4333     0.1508        212        640:  60%|█████▉    | 484/809 [09:38<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4332      0.151        450        640:  60%|█████▉    | 485/809 [09:39<06:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4335     0.1508         87        640:  60%|██████    | 487/809 [09:42<06:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4336     0.1507        123        640:  60%|██████    | 488/809 [09:43<06:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4333     0.1525        177        640:  61%|██████    | 490/809 [09:45<06:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4334     0.1526         62        640:  61%|██████    | 492/809 [09:48<06:39,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4333     0.1526        284        640:  61%|██████    | 493/809 [09:49<06:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4333     0.1527        251        640:  61%|██████    | 495/809 [09:52<06:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4332     0.1529        387        640:  61%|██████▏   | 497/809 [09:54<06:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4332     0.1528        219        640:  62%|██████▏   | 498/809 [09:55<06:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4331     0.1528        216        640:  62%|██████▏   | 500/809 [09:58<06:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4325     0.1536        397        640:  62%|██████▏   | 502/809 [10:00<06:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4321     0.1546        268        640:  62%|██████▏   | 503/809 [10:01<06:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4321     0.1545        176        640:  62%|██████▏   | 505/809 [10:04<06:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4316     0.1548        383        640:  63%|██████▎   | 507/809 [10:06<06:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4318     0.1547         97        640:  63%|██████▎   | 508/809 [10:07<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4317     0.1548        164        640:  63%|██████▎   | 510/809 [10:10<05:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.242     0.4316     0.1548        462        640:  63%|██████▎   | 512/809 [10:12<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4317     0.1547        175        640:  63%|██████▎   | 513/809 [10:13<06:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4315     0.1547        225        640:  64%|██████▎   | 515/809 [10:16<06:21,  1.30s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4317     0.1545         95        640:  64%|██████▍   | 517/809 [10:18<05:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4317     0.1544        210        640:  64%|██████▍   | 518/809 [10:20<05:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4318     0.1542        183        640:  64%|██████▍   | 520/809 [10:22<05:56,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4318     0.1541        235        640:  64%|██████▍   | 521/809 [10:23<05:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4323      0.154        253        640:  65%|██████▍   | 523/809 [10:26<05:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4323     0.1539        306        640:  65%|██████▍   | 525/809 [10:28<05:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4323     0.1539        371        640:  65%|██████▌   | 527/809 [10:30<05:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4323     0.1539        193        640:  65%|██████▌   | 528/809 [10:31<05:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4324     0.1537        161        640:  66%|██████▌   | 530/809 [10:34<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4325     0.1537        285        640:  66%|██████▌   | 532/809 [10:36<05:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4322     0.1537        241        640:  66%|██████▌   | 533/809 [10:37<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4321     0.1537        337        640:  66%|██████▌   | 535/809 [10:40<05:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4321     0.1536        190        640:  66%|██████▋   | 537/809 [10:42<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4322     0.1535        240        640:  67%|██████▋   | 538/809 [10:43<05:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4323     0.1535         74        640:  67%|██████▋   | 540/809 [10:46<05:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4321     0.1535        234        640:  67%|██████▋   | 542/809 [10:48<05:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4319     0.1535        378        640:  67%|██████▋   | 543/809 [10:49<05:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4322     0.1535        215        640:  67%|██████▋   | 545/809 [10:52<05:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4325     0.1534        154        640:  68%|██████▊   | 547/809 [10:54<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4324     0.1533        294        640:  68%|██████▊   | 548/809 [10:55<05:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4324     0.1533        191        640:  68%|██████▊   | 550/809 [10:58<05:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4328     0.1531         60        640:  68%|██████▊   | 552/809 [11:00<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237      0.433      0.153        210        640:  68%|██████▊   | 554/809 [11:02<04:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4334     0.1528        134        640:  69%|██████▊   | 555/809 [11:04<04:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4333     0.1529        217        640:  69%|██████▉   | 557/809 [11:06<04:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4332     0.1533        321        640:  69%|██████▉   | 559/809 [11:08<05:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4329     0.1536        372        640:  69%|██████▉   | 560/809 [11:10<04:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4332     0.1535        218        640:  69%|██████▉   | 562/809 [11:12<04:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4328     0.1541        530        640:  70%|██████▉   | 564/809 [11:14<04:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.433     0.1541        209        640:  70%|██████▉   | 565/809 [11:16<04:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4331     0.1539        160        640:  70%|███████   | 567/809 [11:18<04:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4331     0.1539        272        640:  70%|███████   | 569/809 [11:20<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4329     0.1539        297        640:  70%|███████   | 570/809 [11:22<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4329     0.1539        442        640:  71%|███████   | 572/809 [11:24<04:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4329     0.1538        260        640:  71%|███████   | 574/809 [11:26<04:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4329     0.1538        206        640:  71%|███████   | 575/809 [11:28<04:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4329     0.1538        256        640:  71%|███████▏  | 577/809 [11:30<04:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333     0.1538         46        640:  72%|███████▏  | 579/809 [11:32<04:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333     0.1538        302        640:  72%|███████▏  | 580/809 [11:33<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333     0.1538        222        640:  72%|███████▏  | 582/809 [11:36<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4339     0.1537        230        640:  72%|███████▏  | 584/809 [11:38<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4341     0.1536        143        640:  72%|███████▏  | 586/809 [11:41<04:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4343     0.1536        207        640:  73%|███████▎  | 587/809 [11:42<04:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4342     0.1538        347        640:  73%|███████▎  | 589/809 [11:44<04:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4339     0.1541        182        640:  73%|███████▎  | 591/809 [11:47<04:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.434      0.154        272        640:  73%|███████▎  | 592/809 [11:48<04:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4339      0.154        437        640:  73%|███████▎  | 594/809 [11:50<04:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4339     0.1539        220        640:  74%|███████▎  | 596/809 [11:52<04:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4337      0.154        430        640:  74%|███████▍  | 597/809 [11:54<04:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4341     0.1538        162        640:  74%|███████▍  | 599/809 [11:56<04:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4342     0.1538         96        640:  74%|███████▍  | 601/809 [11:58<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4341     0.1539        415        640:  74%|███████▍  | 602/809 [12:00<04:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.434     0.1539        360        640:  75%|███████▍  | 604/809 [12:02<04:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4342     0.1539        338        640:  75%|███████▍  | 606/809 [12:04<04:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4342     0.1539        266        640:  75%|███████▌  | 607/809 [12:06<03:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.434     0.1539        387        640:  75%|███████▌  | 609/809 [12:08<03:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4343     0.1537        147        640:  76%|███████▌  | 611/809 [12:10<03:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4343     0.1536        282        640:  76%|███████▌  | 613/809 [12:13<03:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4344     0.1535        273        640:  76%|███████▌  | 614/809 [12:14<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4344     0.1538        190        640:  76%|███████▌  | 616/809 [12:16<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4344     0.1538        264        640:  76%|███████▋  | 618/809 [12:19<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4343      0.154        353        640:  77%|███████▋  | 620/809 [12:21<03:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4344      0.154        242        640:  77%|███████▋  | 621/809 [12:22<03:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4343     0.1541        250        640:  77%|███████▋  | 623/809 [12:25<03:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4344     0.1541        204        640:  77%|███████▋  | 625/809 [12:27<03:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4343     0.1543        331        640:  77%|███████▋  | 626/809 [12:28<03:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4345     0.1542        311        640:  78%|███████▊  | 628/809 [12:30<03:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4345     0.1543        299        640:  78%|███████▊  | 630/809 [12:33<03:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4347     0.1542        255        640:  78%|███████▊  | 632/809 [12:35<03:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4347     0.1541        174        640:  78%|███████▊  | 633/809 [12:36<03:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4347     0.1543         95        640:  78%|███████▊  | 635/809 [12:39<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4347     0.1544        131        640:  79%|███████▊  | 637/809 [12:41<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4346     0.1544        283        640:  79%|███████▉  | 638/809 [12:43<03:33,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341      0.155        569        640:  79%|███████▉  | 640/809 [12:45<03:33,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24      0.434     0.1549        201        640:  79%|███████▉  | 641/809 [12:46<03:28,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4339     0.1551        366        640:  79%|███████▉  | 643/809 [12:49<03:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24      0.434      0.155        256        640:  80%|███████▉  | 645/809 [12:51<03:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341      0.155        234        640:  80%|███████▉  | 646/809 [12:52<03:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341      0.155         78        640:  80%|████████  | 648/809 [12:55<03:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4339     0.1551        399        640:  80%|████████  | 650/809 [12:57<03:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4339     0.1551        280        640:  80%|████████  | 651/809 [12:58<03:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4339      0.155        185        640:  81%|████████  | 653/809 [13:01<03:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4336     0.1553        460        640:  81%|█████���██  | 655/809 [13:03<03:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4337     0.1552        227        640:  81%|████████  | 656/809 [13:04<03:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4336     0.1552        416        640:  81%|████████▏ | 658/809 [13:07<03:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4336     0.1551        225        640:  82%|████████▏ | 660/809 [13:09<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4337      0.155        200        640:  82%|████████▏ | 662/809 [13:11<02:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4337      0.155        360        640:  82%|████████▏ | 663/809 [13:13<02:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4338     0.1548        159        640:  82%|████████▏ | 665/809 [13:15<02:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4336     0.1549        378        640:  82%|████████▏ | 667/809 [13:17<02:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4336     0.1548        177        640:  83%|████████▎ | 668/809 [13:18<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4335     0.1547        204        640:  83%|████████▎ | 670/809 [13:21<02:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4338     0.1546        204        640:  83%|████████▎ | 672/809 [13:23<02:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4339     0.1546        200        640:  83%|████████▎ | 673/809 [13:24<02:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.434     0.1547        260        640:  83%|████████▎ | 675/809 [13:27<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239      0.434     0.1546        179        640:  84%|████████▎ | 677/809 [13:29<02:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4341     0.1547        174        640:  84%|████████▍ | 678/809 [13:30<02:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4341     0.1547        420        640:  84%|████████▍ | 680/809 [13:33<02:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4339     0.1548        410        640:  84%|████████▍ | 682/809 [13:35<02:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4338     0.1549        281        640:  84%|████████▍ | 683/809 [13:36<02:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4338     0.1549        297        640:  85%|████████▍ | 685/809 [13:39<02:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4337      0.155        407        640:  85%|████████▍ | 687/809 [13:41<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4337      0.155        240        640:  85%|████████▌ | 688/809 [13:42<02:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241      0.434     0.1548        318        640:  85%|████████▌ | 690/809 [13:45<02:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341     0.1547        190        640:  86%|████████▌ | 692/809 [13:47<02:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341     0.1547        379        640:  86%|████████▌ | 693/809 [13:48<02:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4341     0.1546        272        640:  86%|████████▌ | 695/809 [13:51<02:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4342     0.1545        261        640:  86%|████████▌ | 697/809 [13:53<02:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4342     0.1545        207        640:  86%|████████▋ | 698/809 [13:55<02:17,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24      0.434     0.1545        184        640:  87%|████████▋ | 700/809 [13:57<02:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4339     0.1544        188        640:  87%|████████▋ | 702/809 [13:59<02:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4336     0.1546        316        640:  87%|████████▋ | 703/809 [14:01<02:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.241     0.4335     0.1546        234        640:  87%|████████▋ | 705/809 [14:03<02:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4335     0.1544        118        640:  87%|████████▋ | 707/809 [14:05<01:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4335     0.1544        112        640:  88%|████████▊ | 708/809 [14:06<01:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4337     0.1543        185        640:  88%|████████▊ | 710/809 [14:09<01:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4336     0.1542        239        640:  88%|████████▊ | 712/809 [14:11<01:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4336     0.1542        210        640:  88%|████████▊ | 713/809 [14:13<01:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4337     0.1542        261        640:  88%|████████▊ | 715/809 [14:15<01:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G       1.24     0.4334     0.1545        321        640:  89%|████████▊ | 717/809 [14:17<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4335     0.1544        110        640:  89%|████████▉ | 718/809 [14:19<01:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4336     0.1543        182        640:  89%|████████▉ | 720/809 [14:21<01:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4336     0.1543        225        640:  89%|████████▉ | 722/809 [14:23<01:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4336     0.1542        230        640:  89%|████████▉ | 724/809 [14:26<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4336     0.1541        253        640:  90%|████████▉ | 725/809 [14:27<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4335     0.1541        272        640:  90%|████████▉ | 727/809 [14:29<01:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4335     0.1541        369        640:  90%|█████████ | 729/809 [14:32<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4335     0.1541        248        640:  90%|█████████ | 730/809 [14:33<01:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4335      0.154        274        640:  90%|█████████ | 732/809 [14:35<01:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4335     0.1539        422        640:  91%|█████████ | 733/809 [14:36<01:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333      0.154        346        640:  91%|█████████ | 735/809 [14:39<01:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333     0.1539        148        640:  91%|█████████ | 737/809 [14:41<01:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333      0.154        278        640:  91%|█████████▏| 739/809 [14:44<01:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4332     0.1539        197        640:  91%|█████████▏| 740/809 [14:45<01:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4331     0.1539        142        640:  92%|█████████▏| 742/809 [14:47<01:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4332     0.1538        200        640:  92%|█████████▏| 744/809 [14:49<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4334     0.1537        222        640:  92%|█████████▏| 745/809 [14:51<01:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4336     0.1536        292        640:  92%|█████████▏| 747/809 [14:53<01:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4336     0.1537        205        640:  93%|█████████▎| 749/809 [14:55<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4335     0.1537        364        640:  93%|█████████▎| 750/809 [14:57<01:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4331     0.1541        404        640:  93%|█████████▎| 752/809 [14:59<01:10,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4331      0.154        239        640:  93%|█████████▎| 754/809 [15:01<01:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4332      0.154        230        640:  93%|█████████▎| 755/809 [15:03<01:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4333     0.1539        229        640:  94%|█████████▎| 757/809 [15:05<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.239     0.4332     0.1542        193        640:  94%|█████████▍| 759/809 [15:07<01:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4332     0.1542        237        640:  94%|█████████▍| 761/809 [15:10<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4333     0.1542        192        640:  94%|█████████▍| 762/809 [15:11<00:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4336     0.1541         51        640:  94%|█████████▍| 764/809 [15:13<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4338      0.154        188        640:  95%|█████████▍| 766/809 [15:16<00:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4339     0.1539         96        640:  95%|█████████▍| 767/809 [15:17<00:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.236     0.4343     0.1538        198        640:  95%|█████████▌| 769/809 [15:19<00:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4344     0.1538        313        640:  95%|█████████▌| 771/809 [15:21<00:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4344     0.1537        299        640:  96%|█████████▌| 773/809 [15:24<00:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4346     0.1537        143        640:  96%|█████████▌| 774/809 [15:25<00:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4349     0.1536         99        640:  96%|█████████▌| 776/809 [15:27<00:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4348     0.1535        261        640:  96%|█████████▌| 778/809 [15:30<00:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4346     0.1538        410        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.234     0.4346     0.1537        311        640:  97%|█████████▋| 781/809 [15:33<00:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4345     0.1539        352        640:  97%|█████████▋| 783/809 [15:36<00:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4342     0.1538        139        640:  97%|█████████▋| 785/809 [15:38<00:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.235     0.4341     0.1538        176        640:  97%|█████████▋| 786/809 [15:39<00:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4337     0.1541        384        640:  97%|█████████▋| 788/809 [15:42<00:25,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4334     0.1541        383        640:  98%|█████████▊| 790/809 [15:44<00:23,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4333     0.1541        187        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4333     0.1541         84        640:  98%|█████████▊| 793/809 [15:48<00:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4332      0.154        209        640:  98%|█████████▊| 795/809 [15:50<00:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4331     0.1543        387        640:  98%|█████████▊| 796/809 [15:51<00:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238      0.433     0.1545        425        640:  99%|█████████▊| 798/809 [15:54<00:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.238     0.4331     0.1545        292        640:  99%|█████████▉| 800/809 [15:56<00:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4332     0.1543         97        640:  99%|█████████▉| 802/809 [15:58<00:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4333     0.1543        114        640:  99%|█████████▉| 803/809 [16:00<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4332     0.1543        226        640: 100%|█████████▉| 805/809 [16:02<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4331     0.1543        480        640: 100%|█████████▉| 807/809 [16:04<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     28/300      13.7G      1.237     0.4331     0.1543        237        640: 100%|█████████▉| 808/809 [16:06<00:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   9%|▊         | 6/69 [00:01<00:15,  4.03it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  20%|██        | 14/69 [00:03<00:14,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  32%|███▏      | 22/69 [00:05<00:11,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  43%|████▎     | 30/69 [00:07<00:09,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  55%|█████▌    | 38/69 [00:09<00:07,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  65%|██████▌   | 45/69 [00:11<00:06,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  77%|███████▋  | 53/69 [00:13<00:04,  3.77it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  88%|████████▊ | 61/69 [00:15<00:02,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.424      0.282      0.284      0.148
-     29/300      13.5G      1.167     0.4589     0.1409        217        640:   0%|          | 1/809 [00:01<15:26,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.336     0.4044     0.1431        202        640:   0%|          | 2/809 [00:02<15:38,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.291     0.4073     0.1386        194        640:   0%|          | 4/809 [00:04<15:41,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.245     0.4177     0.1363        260        640:   1%|          | 6/809 [00:07<15:45,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.201     0.4365     0.1327        298        640:   1%|          | 8/809 [00:09<15:35,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4233     0.1344        294        640:   1%|          | 9/809 [00:10<15:39,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.194     0.4303     0.1293        208        640:   1%|▏         | 11/809 [00:12<15:42,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.183     0.4334      0.129        238        640:   2%|▏         | 13/809 [00:15<16:14,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.183      0.433     0.1288        274        640:   2%|▏         | 14/809 [00:16<16:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.174      0.438     0.1279        297        640:   2%|▏         | 16/809 [00:18<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.17     0.4448     0.1279        287        640:   2%|▏         | 18/809 [00:21<15:47,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.178     0.4423     0.1289        206        640:   2%|▏         | 19/809 [00:22<15:44,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.204     0.4386     0.1437        190        640:   3%|▎         | 21/809 [00:25<16:00,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.203     0.4402     0.1472        294        640:   3%|▎         | 23/809 [00:27<15:50,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.197     0.4431     0.1451        133        640:   3%|▎         | 24/809 [00:28<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4367     0.1494        197        640:   3%|▎         | 26/809 [00:30<15:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4326     0.1495        246        640:   3%|▎         | 28/809 [00:33<15:42,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.243     0.4296     0.1497        435        640:   4%|▎         | 29/809 [00:34<15:55,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.241     0.4292     0.1494        310        640:   4%|▍         | 31/809 [00:37<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.248     0.4264     0.1511        363        640:   4%|▍         | 33/809 [00:39<15:19,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.245     0.4281     0.1512        258        640:   4%|▍         | 34/809 [00:40<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4293     0.1495        411        640:   4%|▍         | 36/809 [00:42<15:04,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4324     0.1496        262        640:   5%|▍         | 38/809 [00:45<15:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4353     0.1498        126        640:   5%|▍         | 39/809 [00:46<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217      0.437     0.1491        340        640:   5%|▌         | 41/809 [00:48<15:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4381     0.1481        248        640:   5%|▌         | 43/809 [00:51<14:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222      0.436     0.1507        231        640:   6%|▌         | 45/809 [00:53<15:21,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4344     0.1511        487        640:   6%|▌         | 46/809 [00:54<15:19,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214      0.438      0.149        218        640:   6%|▌         | 48/809 [00:57<15:27,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4386     0.1482        288        640:   6%|▌         | 50/809 [00:59<14:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4413     0.1474        180        640:   6%|▋         | 51/809 [01:00<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4391     0.1489        294        640:   7%|▋         | 53/809 [01:03<15:14,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4408     0.1476        120        640:   7%|▋         | 55/809 [01:05<15:07,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4389     0.1506        468        640:   7%|▋         | 56/809 [01:06<15:07,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4388     0.1508        441        640:   7%|▋         | 58/809 [01:09<14:51,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4387     0.1514        369        640:   7%|▋         | 60/809 [01:11<14:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4387     0.1506        152        640:   8%|▊         | 61/809 [01:12<14:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4419     0.1491        115        640:   8%|▊         | 63/809 [01:14<14:36,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G        1.2     0.4437     0.1481        143        640:   8%|▊         | 65/809 [01:17<14:29,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.198     0.4447     0.1474        146        640:   8%|▊         | 66/809 [01:18<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.201     0.4437     0.1469        285        640:   8%|▊         | 68/809 [01:20<14:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.201     0.4424      0.146        262        640:   9%|▊         | 70/809 [01:23<14:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.201     0.4424     0.1462        375        640:   9%|▉         | 72/809 [01:25<14:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.197     0.4437     0.1454        182        640:   9%|▉         | 73/809 [01:26<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.194     0.4463     0.1447        306        640:   9%|▉         | 75/809 [01:29<14:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.19     0.4479     0.1436        239        640:  10%|▉         | 77/809 [01:31<14:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.187     0.4494     0.1433        121        640:  10%|▉         | 78/809 [01:32<14:18,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.191     0.4483     0.1446        121        640:  10%|▉         | 80/809 [01:35<14:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.195     0.4472     0.1455        394        640:  10%|█         | 82/809 [01:37<14:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.197     0.4467     0.1456        161        640:  10%|█         | 83/809 [01:38<14:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.195     0.4464     0.1446        122        640:  11%|█         | 85/809 [01:40<14:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.195     0.4466     0.1451        440        640:  11%|█         | 87/809 [01:43<14:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.187     0.4499     0.1436        163        640:  11%|█         | 89/809 [01:45<14:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.184     0.4503     0.1432        159        640:  11%|█         | 90/809 [01:46<13:55,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.178     0.4518     0.1425         98        640:  11%|█▏        | 92/809 [01:49<14:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.174     0.4538     0.1418        189        640:  12%|█▏        | 94/809 [01:51<14:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.173      0.454     0.1411        116        640:  12%|█▏        | 95/809 [01:52<14:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.171     0.4538     0.1408        139        640:  12%|█▏        | 97/809 [01:55<13:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.169     0.4534     0.1402        270        640:  12%|█▏        | 99/809 [01:57<14:17,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.169     0.4528     0.1403        207        640:  12%|█▏        | 100/809 [01:58<14:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.17     0.4521     0.1404        244        640:  13%|█▎        | 102/809 [02:01<14:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.167     0.4533     0.1394        210        640:  13%|█▎        | 104/809 [02:03<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.159     0.4554     0.1384         90        640:  13%|█▎        | 106/809 [02:05<13:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.157     0.4563      0.138        211        640:  13%|█▎        | 107/809 [02:06<13:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.16     0.4554     0.1392        183        640:  13%|█▎        | 109/809 [02:09<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.159     0.4559      0.139        177        640:  14%|█▎        | 111/809 [02:11<13:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.161     0.4552     0.1392        329        640:  14%|█▍        | 112/809 [02:12<13:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.16     0.4562     0.1392        476        640:  14%|█▍        | 114/809 [02:15<13:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.161     0.4566     0.1396        404        640:  14%|█▍        | 116/809 [02:17<13:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.165     0.4558     0.1404        287        640:  14%|█▍        | 117/809 [02:18<13:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.166     0.4555     0.1403        187        640:  15%|█▍        | 119/809 [02:21<14:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.168     0.4547     0.1409        288        640:  15%|█▍        | 121/809 [02:23<14:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.169     0.4544     0.1406        226        640:  15%|█▌        | 122/809 [02:24<13:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.176      0.452     0.1412        439        640:  15%|█▌        | 124/809 [02:27<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.174     0.4517     0.1407        201        640:  16%|█▌        | 126/809 [02:29<13:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.171     0.4526     0.1404        119        640:  16%|█▌        | 127/809 [02:30<13:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.176     0.4511     0.1409        179        640:  16%|█▌        | 129/809 [02:33<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.179     0.4492     0.1408        378        640:  16%|█▌        | 131/809 [02:35<13:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.185     0.4474     0.1421        630        640:  16%|█▋        | 133/809 [02:38<14:06,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.186     0.4468      0.142        218        640:  17%|█▋        | 134/809 [02:39<13:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.19     0.4452     0.1418        224        640:  17%|█▋        | 136/809 [02:41<13:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.196     0.4431     0.1421        379        640:  17%|█▋        | 138/809 [02:43<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.204     0.4407     0.1428        381        640:  17%|█▋        | 140/809 [02:46<13:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209      0.439     0.1435        426        640:  17%|█▋        | 141/809 [02:47<13:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208     0.4392     0.1432        410        640:  18%|█▊        | 143/809 [02:50<13:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.21     0.4386     0.1434        428        640:  18%|█▊        | 144/809 [02:51<13:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4387      0.143        294        640:  18%|█▊        | 146/809 [02:53<13:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208      0.439     0.1424        321        640:  18%|█▊        | 148/809 [02:56<13:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4382     0.1426        382        640:  18%|█▊        | 149/809 [02:57<13:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215      0.437     0.1443        278        640:  19%|█▊        | 151/809 [02:59<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4371      0.144        169        640:  19%|█▉        | 153/809 [03:02<12:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4368     0.1441        186        640:  19%|█▉        | 155/809 [03:04<12:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4364     0.1444        539        640:  19%|█▉        | 156/809 [03:05<13:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4355     0.1443        366        640:  20%|█▉        | 158/809 [03:08<13:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4346     0.1462        511        640:  20%|█▉        | 159/809 [03:09<13:31,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4343     0.1464        252        640:  20%|█▉        | 161/809 [03:11<13:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4343     0.1461        254        640:  20%|██        | 162/809 [03:13<13:22,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4328     0.1483        197        640:  20%|██        | 164/809 [03:15<13:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4322     0.1487        374        640:  21%|██        | 166/809 [03:18<13:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4327     0.1484        100        640:  21%|██        | 168/809 [03:20<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4325     0.1484        179        640:  21%|██        | 169/809 [03:21<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4318     0.1484        353        640:  21%|██        | 171/809 [03:23<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4312     0.1491        415        640:  21%|██▏       | 173/809 [03:26<13:03,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4306     0.1494        379        640:  22%|██▏       | 174/809 [03:27<12:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236       0.43     0.1501        197        640:  22%|██▏       | 176/809 [03:30<12:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4299     0.1498        208        640:  22%|██▏       | 178/809 [03:32<12:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4299     0.1497        327        640:  22%|██▏       | 179/809 [03:33<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4297     0.1494        150        640:  22%|██▏       | 181/809 [03:35<12:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4293     0.1492        268        640:  23%|██▎       | 183/809 [03:38<12:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4297     0.1489        161        640:  23%|██▎       | 184/809 [03:39<12:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4296     0.1487        192        640:  23%|██▎       | 186/809 [03:41<12:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4292     0.1486        115        640:  23%|██▎       | 188/809 [03:44<12:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4294     0.1482        205        640:  23%|██▎       | 190/809 [03:46<12:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4299      0.148        170        640:  24%|██▎       | 191/809 [03:47<12:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4304     0.1476        224        640:  24%|██▍       | 193/809 [03:50<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4309     0.1473        260        640:  24%|██▍       | 195/809 [03:52<12:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229      0.431      0.147        179        640:  24%|██▍       | 196/809 [03:53<12:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4303     0.1469        154        640:  24%|██▍       | 198/809 [03:56<12:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4298     0.1467        160        640:  25%|██▍       | 200/809 [03:58<12:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4298     0.1466        227        640:  25%|██▍       | 201/809 [03:59<12:26,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232      0.429     0.1467        328        640:  25%|██▌       | 203/809 [04:02<12:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4287     0.1467        393        640:  25%|██▌       | 204/809 [04:03<12:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4283     0.1469        472        640:  25%|██▌       | 206/809 [04:05<12:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4284     0.1472        320        640:  26%|██▌       | 208/809 [04:08<12:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4287     0.1472        300        640:  26%|██▌       | 209/809 [04:09<12:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4292     0.1469        321        640:  26%|██▌       | 211/809 [04:11<11:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4286     0.1474        426        640:  26%|██▋       | 213/809 [04:14<12:16,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4285     0.1472        181        640:  26%|██▋       | 214/809 [04:15<12:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4293     0.1472        161        640:  27%|██▋       | 216/809 [04:17<11:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4289     0.1473        217        640:  27%|██▋       | 218/809 [04:20<12:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4288     0.1472        310        640:  27%|██▋       | 219/809 [04:21<11:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4287      0.147        325        640:  27%|██▋       | 221/809 [04:24<11:59,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4274     0.1475        301        640:  28%|██▊       | 223/809 [04:26<11:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4276     0.1474        191        640:  28%|██▊       | 224/809 [04:27<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.24      0.427     0.1479        316        640:  28%|██▊       | 226/809 [04:30<11:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.239     0.4275     0.1479        105        640:  28%|██▊       | 228/809 [04:32<11:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.239     0.4276     0.1478        211        640:  28%|██▊       | 229/809 [04:33<11:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4281     0.1474        253        640:  29%|██▊       | 231/809 [04:35<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4279     0.1475        398        640:  29%|██▉       | 233/809 [04:38<11:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4283     0.1474        254        640:  29%|██▉       | 235/809 [04:40<11:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4287     0.1472        113        640:  29%|██▉       | 236/809 [04:41<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4287     0.1469        317        640:  29%|██▉       | 238/809 [04:44<11:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4284     0.1474        560        640:  30%|██▉       | 240/809 [04:46<11:38,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4281     0.1474        404        640:  30%|██▉       | 241/809 [04:47<11:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.24     0.4274     0.1481        407        640:  30%|███       | 243/809 [04:50<11:42,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.241     0.4272     0.1482        371        640:  30%|███       | 244/809 [04:51<11:40,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.238     0.4282     0.1477        233        640:  30%|███       | 246/809 [04:53<11:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4285     0.1476        314        640:  31%|███       | 248/809 [04:56<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237     0.4284     0.1477        337        640:  31%|███       | 250/809 [04:58<11:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4286     0.1474        125        640:  31%|███       | 251/809 [04:59<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4285     0.1476        440        640:  31%|███▏      | 253/809 [05:02<11:28,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234      0.429     0.1472        242        640:  32%|███▏      | 255/809 [05:04<11:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4289     0.1473        335        640:  32%|███▏      | 256/809 [05:05<11:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4289     0.1471        334        640:  32%|███▏      | 258/809 [05:08<10:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.236     0.4282     0.1473        270        640:  32%|███▏      | 260/809 [05:10<10:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.237      0.428     0.1474        342        640:  32%|███▏      | 261/809 [05:11<10:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4289     0.1471        218        640:  33%|███▎      | 263/809 [05:14<10:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4292     0.1467        219        640:  33%|███▎      | 265/809 [05:16<10:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4288     0.1472        548        640:  33%|███▎      | 266/809 [05:17<10:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4287      0.147        411        640:  33%|███▎      | 268/809 [05:20<10:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4294     0.1467        201        640:  33%|███▎      | 270/809 [05:22<10:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231      0.429     0.1469        239        640:  33%|███▎      | 271/809 [05:23<10:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4284     0.1468        239        640:  34%|███▎      | 273/809 [05:26<10:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4285     0.1466        111        640:  34%|███▍      | 275/809 [05:28<10:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4285     0.1465        231        640:  34%|███▍      | 276/809 [05:29<10:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4275     0.1476        476        640:  34%|███▍      | 278/809 [05:32<10:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4272      0.148        355        640:  35%|███▍      | 280/809 [05:34<10:49,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4274     0.1479        200        640:  35%|███▍      | 282/809 [05:36<10:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4274     0.1477        146        640:  35%|███▌      | 284/809 [05:39<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4277     0.1476        134        640:  35%|███▌      | 285/809 [05:40<10:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4277     0.1473        228        640:  35%|███▌      | 287/809 [05:42<10:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4273     0.1472        310        640:  36%|███▌      | 289/809 [05:45<10:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235      0.427     0.1472        161        640:  36%|███▌      | 290/809 [05:46<10:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4267     0.1472        244        640:  36%|███▌      | 292/809 [05:48<10:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4269     0.1469        282        640:  36%|███▋      | 294/809 [05:51<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4275     0.1465        175        640:  37%|███▋      | 296/809 [05:53<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4272     0.1465        205        640:  37%|███▋      | 297/809 [05:54<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4272     0.1464        114        640:  37%|███▋      | 299/809 [05:56<09:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4268     0.1473        161        640:  37%|███▋      | 301/809 [05:59<10:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4269     0.1472        237        640:  37%|███▋      | 302/809 [06:00<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4283     0.1469         59        640:  38%|███▊      | 304/809 [06:02<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4288      0.147        267        640:  38%|███▊      | 306/809 [06:05<09:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4297     0.1468        109        640:  38%|███▊      | 307/809 [06:06<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4302     0.1467        271        640:  38%|███▊      | 309/809 [06:08<10:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227       0.43     0.1466        178        640:  38%|███▊      | 311/809 [06:11<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4296     0.1466        258        640:  39%|███▊      | 312/809 [06:12<09:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4297     0.1464        168        640:  39%|███▉      | 314/809 [06:14<09:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4302     0.1461        124        640:  39%|███▉      | 316/809 [06:17<09:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228       0.43     0.1463        355        640:  39%|███▉      | 318/809 [06:19<09:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4299     0.1462        306        640:  39%|███▉      | 319/809 [06:20<09:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231      0.429     0.1471        276        640:  40%|███▉      | 321/809 [06:23<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4285     0.1471        380        640:  40%|███▉      | 323/809 [06:25<09:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4288     0.1469        205        640:  40%|████      | 324/809 [06:26<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23      0.429     0.1466        143        640:  40%|████      | 326/809 [06:28<09:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4291     0.1465        258        640:  41%|████      | 328/809 [06:31<09:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4289     0.1466        269        640:  41%|████      | 329/809 [06:32<09:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4291     0.1464        208        640:  41%|████      | 331/809 [06:34<09:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4292     0.1462        183        640:  41%|████      | 333/809 [06:37<09:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4292     0.1461        165        640:  41%|████▏     | 334/809 [06:38<09:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4287     0.1467        657        640:  42%|████▏     | 336/809 [06:40<09:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4289     0.1467        161        640:  42%|████▏     | 338/809 [06:43<09:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4284     0.1472        510        640:  42%|████▏     | 339/809 [06:44<09:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4281     0.1475        426        640:  42%|████▏     | 341/809 [06:47<09:39,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4288     0.1472        163        640:  42%|████▏     | 343/809 [06:49<09:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4286     0.1472        321        640:  43%|████▎     | 344/809 [06:50<09:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4287     0.1474        445        640:  43%|████▎     | 346/809 [06:52<09:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4283     0.1476        239        640:  43%|████▎     | 348/809 [06:55<09:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4285     0.1475        219        640:  43%|████▎     | 349/809 [06:56<09:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4286     0.1474        248        640:  43%|████▎     | 351/809 [06:58<09:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4281     0.1476        404        640:  44%|████▎     | 353/809 [07:01<09:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4284     0.1474        154        640:  44%|████▍     | 354/809 [07:02<09:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4281     0.1473        322        640:  44%|████▍     | 356/809 [07:04<09:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4279     0.1475        478        640:  44%|████▍     | 358/809 [07:07<09:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4278     0.1474        207        640:  44%|████▍     | 359/809 [07:08<09:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4278     0.1474        373        640:  45%|████▍     | 361/809 [07:10<09:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4278     0.1484        414        640:  45%|████▍     | 363/809 [07:13<09:10,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235      0.428     0.1482        246        640:  45%|████▍     | 364/809 [07:14<08:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4281     0.1483        264        640:  45%|████▌     | 366/809 [07:17<08:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.235     0.4281     0.1486        290        640:  45%|████▌     | 368/809 [07:19<08:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.234     0.4284     0.1485        229        640:  46%|████▌     | 369/809 [07:20<08:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4285     0.1483        179        640:  46%|████▌     | 371/809 [07:22<08:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4284     0.1483        608        640:  46%|████▌     | 373/809 [07:25<08:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.233     0.4286     0.1482        191        640:  46%|████▌     | 374/809 [07:26<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4289     0.1479        192        640:  46%|████▋     | 376/809 [07:28<08:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4295     0.1478         70        640:  47%|████▋     | 378/809 [07:31<08:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4296     0.1476        165        640:  47%|████▋     | 380/809 [07:33<08:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4297     0.1474        179        640:  47%|████▋     | 381/809 [07:34<08:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4298     0.1473        330        640:  47%|████▋     | 383/809 [07:37<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4296     0.1474        452        640:  48%|████▊     | 385/809 [07:39<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4295     0.1474        131        640:  48%|████▊     | 386/809 [07:40<08:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4296     0.1474         73        640:  48%|████▊     | 388/809 [07:43<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231     0.4298     0.1474        162        640:  48%|████▊     | 390/809 [07:45<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.232     0.4298     0.1475        229        640:  48%|████▊     | 391/809 [07:46<08:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.231       0.43     0.1474        241        640:  49%|████▊     | 393/809 [07:49<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4305     0.1473        380        640:  49%|████▉     | 395/809 [07:51<08:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4305     0.1472        349        640:  49%|████▉     | 396/809 [07:52<08:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4309      0.147        151        640:  49%|████▉     | 398/809 [07:54<08:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4309     0.1468        250        640:  49%|████▉     | 400/809 [07:57<08:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23     0.4308     0.1469        373        640:  50%|████▉     | 402/809 [07:59<08:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.23      0.431     0.1468        204        640:  50%|████▉     | 403/809 [08:00<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4312     0.1466        243        640:  50%|█████     | 405/809 [08:03<07:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227      0.432     0.1463        182        640:  50%|█████     | 407/809 [08:05<07:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227      0.432     0.1464        311        640:  50%|█████     | 408/809 [08:06<07:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4322     0.1463        210        640:  51%|█████     | 410/809 [08:08<07:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4328     0.1461        134        640:  51%|█████     | 412/809 [08:11<07:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4329     0.1467        219        640:  51%|█████     | 414/809 [08:13<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.225     0.4332     0.1466        151        640:  51%|█████▏    | 415/809 [08:14<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.225     0.4334     0.1466        219        640:  52%|█████▏    | 417/809 [08:17<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4337     0.1464        131        640:  52%|█████▏    | 419/809 [08:19<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4338     0.1464        320        640:  52%|█████▏    | 420/809 [08:20<07:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4338     0.1462        178        640:  52%|█████▏    | 422/809 [08:23<07:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4344      0.146        210        640:  52%|█████▏    | 424/809 [08:25<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4344     0.1459        306        640:  53%|█████▎    | 425/809 [08:26<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4345     0.1458        118        640:  53%|█████▎    | 427/809 [08:29<07:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4344     0.1458        121        640:  53%|█████▎    | 429/809 [08:31<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4346     0.1457        131        640:  53%|█████▎    | 430/809 [08:32<07:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4345     0.1461        297        640:  53%|█████▎    | 432/809 [08:35<07:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4346     0.1462        158        640:  54%|█████▎    | 434/809 [08:37<07:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4348     0.1462        111        640:  54%|█████▍    | 436/809 [08:39<07:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4348     0.1464        144        640:  54%|█████▍    | 438/809 [08:42<07:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4344     0.1468        475        640:  54%|█████▍    | 439/809 [08:43<07:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4339     0.1468        459        640:  55%|█████▍    | 441/809 [08:45<07:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4338     0.1467        229        640:  55%|█████▍    | 443/809 [08:48<07:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4336     0.1468        238        640:  55%|█████▍    | 444/809 [08:49<07:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4336     0.1467        235        640:  55%|█████▌    | 446/809 [08:51<07:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4335     0.1466        212        640:  55%|█████▌    | 448/809 [08:54<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4331     0.1469        284        640:  56%|█████▌    | 449/809 [08:55<07:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4328     0.1468        117        640:  56%|█████▌    | 451/809 [08:57<06:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4321      0.147        293        640:  56%|█████▌    | 453/809 [09:00<07:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4316     0.1471        307        640:  56%|█████▌    | 455/809 [09:02<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4314     0.1471        251        640:  56%|█████▋    | 456/809 [09:03<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4314      0.147        180        640:  57%|█████▋    | 458/809 [09:05<06:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.229     0.4311      0.147        224        640:  57%|█████▋    | 460/809 [09:08<06:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.228     0.4313      0.147        153        640:  57%|█████▋    | 461/809 [09:09<06:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4318     0.1467        151        640:  57%|█████▋    | 463/809 [09:11<06:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4318     0.1467        197        640:  57%|█████▋    | 465/809 [09:14<06:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4319     0.1467        277        640:  58%|█████▊    | 466/809 [09:15<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.227     0.4319     0.1468        395        640:  58%|█████▊    | 468/809 [09:17<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226      0.432     0.1467        220        640:  58%|█████▊    | 470/809 [09:20<06:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4321     0.1467        377        640:  58%|█████▊    | 472/809 [09:22<06:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226      0.432     0.1467        316        640:  58%|█████▊    | 473/809 [09:23<06:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.225      0.432     0.1467        311        640:  59%|█████▊    | 475/809 [09:26<06:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.225      0.432     0.1466        165        640:  59%|█████▉    | 476/809 [09:27<06:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.226     0.4318     0.1472        115        640:  59%|█████▉    | 478/809 [09:29<06:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4321     0.1471        205        640:  59%|█████▉    | 480/809 [09:32<06:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4321     0.1471        198        640:  60%|█████▉    | 482/809 [09:34<06:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224     0.4321      0.147        334        640:  60%|█████▉    | 483/809 [09:35<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.224      0.432     0.1473        392        640:  60%|█████▉    | 485/809 [09:38<06:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4324     0.1472        187        640:  60%|██████    | 487/809 [09:40<06:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.223     0.4324     0.1472        261        640:  60%|██████    | 488/809 [09:41<06:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4326     0.1473         90        640:  61%|██████    | 490/809 [09:43<06:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4324     0.1472        158        640:  61%|██████    | 492/809 [09:46<06:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4325     0.1472        291        640:  61%|██████    | 493/809 [09:47<06:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4323     0.1473        199        640:  61%|██████    | 495/809 [09:49<06:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4323     0.1473        326        640:  61%|██████▏   | 497/809 [09:52<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4322     0.1472        243        640:  62%|██████▏   | 499/809 [09:54<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.221     0.4322     0.1473        331        640:  62%|██████▏   | 500/809 [09:55<06:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.222     0.4321     0.1475        137        640:  62%|██████▏   | 502/809 [09:58<06:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.22     0.4324     0.1473        231        640:  62%|██████▏   | 504/809 [10:00<06:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4326     0.1473        203        640:  62%|██████▏   | 505/809 [10:01<06:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.22     0.4324     0.1474        377        640:  63%|██████▎   | 507/809 [10:04<06:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4325     0.1473        349        640:  63%|██████▎   | 509/809 [10:06<06:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4326     0.1472        270        640:  63%|██████▎   | 510/809 [10:07<05:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4325     0.1471        293        640:  63%|██████▎   | 512/809 [10:10<05:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218     0.4327     0.1469        148        640:  64%|██████▎   | 514/809 [10:12<05:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218     0.4327     0.1469        141        640:  64%|██████▎   | 515/809 [10:13<05:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4329     0.1467        101        640:  64%|██████▍   | 517/809 [10:15<05:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4326     0.1466        270        640:  64%|██████▍   | 519/809 [10:18<05:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4329     0.1465         55        640:  64%|██████▍   | 521/809 [10:20<05:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4328     0.1464        138        640:  65%|██████▍   | 522/809 [10:21<05:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4326     0.1464        278        640:  65%|██████▍   | 524/809 [10:24<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4324     0.1464        131        640:  65%|██████▌   | 526/809 [10:26<05:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4324     0.1463        146        640:  65%|██████▌   | 527/809 [10:27<05:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4324     0.1462        337        640:  65%|██████▌   | 529/809 [10:30<05:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4329      0.146        142        640:  66%|██████▌   | 531/809 [10:32<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215      0.433     0.1461        316        640:  66%|██████▌   | 532/809 [10:33<05:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4331     0.1461         53        640:  66%|██████▌   | 534/809 [10:35<05:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4332     0.1459        145        640:  66%|██████▋   | 536/809 [10:38<05:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4333     0.1458        155        640:  67%|██████▋   | 538/809 [10:40<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4333     0.1457        233        640:  67%|██████▋   | 539/809 [10:41<05:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4334     0.1457        154        640:  67%|██████▋   | 541/809 [10:44<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4337     0.1455        211        640:  67%|██████▋   | 543/809 [10:46<05:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4336     0.1456        131        640:  67%|██████▋   | 544/809 [10:47<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4334     0.1456        274        640:  67%|██████▋   | 546/809 [10:50<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4335     0.1454        224        640:  68%|██████▊   | 548/809 [10:52<05:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4336     0.1454        293        640:  68%|██████▊   | 549/809 [10:53<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4334     0.1454        287        640:  68%|██████▊   | 551/809 [10:55<05:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4334     0.1453        256        640:  68%|██████▊   | 553/809 [10:58<05:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4333     0.1452        304        640:  69%|██████▊   | 555/809 [11:00<04:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4332     0.1452        314        640:  69%|██████▊   | 556/809 [11:01<04:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4333     0.1451        160        640:  69%|██████▉   | 558/809 [11:04<04:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4331     0.1451        222        640:  69%|██████▉   | 560/809 [11:06<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4332      0.145        252        640:  69%|██████▉   | 561/809 [11:07<05:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214      0.433      0.145        262        640:  70%|██████▉   | 563/809 [11:10<04:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214      0.433      0.145        295        640:  70%|██████▉   | 565/809 [11:12<04:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4329     0.1451        410        640:  70%|██████▉   | 566/809 [11:13<04:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4324     0.1452        376        640:  70%|███████   | 568/809 [11:16<04:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4324     0.1452        129        640:  70%|███████   | 570/809 [11:18<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4324     0.1452        192        640:  71%|███████   | 571/809 [11:19<04:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4321     0.1452        347        640:  71%|███████   | 573/809 [11:22<04:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216      0.432     0.1451        264        640:  71%|███████   | 575/809 [11:24<04:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4323      0.145        187        640:  71%|███████   | 576/809 [11:25<04:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4323      0.145        185        640:  71%|███████▏  | 578/809 [11:28<04:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4325     0.1449        177        640:  72%|███████▏  | 580/809 [11:30<04:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4325     0.1449        290        640:  72%|███████▏  | 581/809 [11:31<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4327     0.1447        169        640:  72%|███████▏  | 583/809 [11:34<04:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4328     0.1447        176        640:  72%|███████▏  | 585/809 [11:36<04:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214      0.433     0.1446        352        640:  73%|███████▎  | 587/809 [11:38<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4331     0.1446        184        640:  73%|███████▎  | 588/809 [11:39<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213      0.433     0.1447        325        640:  73%|███████▎  | 590/809 [11:42<04:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4333     0.1447        163        640:  73%|███████▎  | 592/809 [11:44<04:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4333     0.1446        284        640:  73%|███████▎  | 594/809 [11:47<04:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4333     0.1446        186        640:  74%|███████▎  | 595/809 [11:48<04:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4332     0.1448        317        640:  74%|███████▍  | 597/809 [11:50<04:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4332     0.1447        146        640:  74%|███████▍  | 599/809 [11:53<04:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214      0.433     0.1449        291        640:  74%|███████▍  | 600/809 [11:54<04:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4329     0.1449        203        640:  74%|███████▍  | 602/809 [11:56<04:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4327     0.1449        162        640:  75%|███████▍  | 604/809 [11:59<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4326     0.1449        198        640:  75%|███████▍  | 605/809 [12:00<04:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4326     0.1451        338        640:  75%|███████▌  | 607/809 [12:02<04:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4323     0.1454        454        640:  75%|███████▌  | 609/809 [12:05<04:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4324     0.1454        151        640:  75%|███████▌  | 610/809 [12:06<04:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4325     0.1452        156        640:  76%|███████▌  | 612/809 [12:08<03:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4323     0.1453        451        640:  76%|███████▌  | 614/809 [12:11<03:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4323     0.1453        193        640:  76%|███████▌  | 616/809 [12:13<03:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4322     0.1452        255        640:  76%|███████▋  | 617/809 [12:14<03:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218     0.4319     0.1454        214        640:  77%|███████▋  | 619/809 [12:16<03:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4316     0.1455        251        640:  77%|███████▋  | 621/809 [12:19<03:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4317     0.1454        153        640:  77%|███████▋  | 622/809 [12:20<03:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4315     0.1454        474        640:  77%|███████▋  | 624/809 [12:22<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4315     0.1454        186        640:  77%|███████▋  | 626/809 [12:25<03:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4317     0.1453        160        640:  78%|███████▊  | 627/809 [12:26<03:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4319     0.1453        221        640:  78%|███████▊  | 629/809 [12:28<03:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4318     0.1453        301        640:  78%|███████▊  | 631/809 [12:31<03:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218     0.4321     0.1452        185        640:  78%|███████▊  | 632/809 [12:32<03:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218     0.4322     0.1451        309        640:  78%|███████▊  | 634/809 [12:34<03:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218      0.432     0.1451        223        640:  79%|███████▊  | 636/809 [12:37<03:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.219     0.4318     0.1452        462        640:  79%|███████▊  | 637/809 [12:38<03:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.218      0.432     0.1451        241        640:  79%|███████▉  | 639/809 [12:40<03:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.217     0.4323     0.1449        121        640:  79%|███████▉  | 641/809 [12:43<03:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4325     0.1449        202        640:  79%|███████▉  | 643/809 [12:45<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4326     0.1448        235        640:  80%|███████▉  | 644/809 [12:46<03:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4327     0.1447        200        640:  80%|███████▉  | 646/809 [12:49<03:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4325     0.1447        292        640:  80%|████████  | 648/809 [12:51<03:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4325     0.1447        171        640:  80%|████████  | 649/809 [12:52<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4326     0.1446        264        640:  80%|████████  | 651/809 [12:54<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.216     0.4326     0.1445        226        640:  81%|████████  | 653/809 [12:57<03:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.215     0.4327     0.1444        187        640:  81%|████████  | 654/809 [12:58<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4331     0.1443        211        640:  81%|████████  | 656/809 [13:00<03:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.214     0.4332     0.1442        280        640:  81%|████████▏ | 658/809 [13:03<02:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4335     0.1441        291        640:  81%|████████▏ | 659/809 [13:05<02:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4337     0.1441         89        640:  82%|████████▏ | 661/809 [13:06<02:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4338     0.1441        351        640:  82%|████████▏ | 663/809 [13:09<02:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211      0.434     0.1439        175        640:  82%|████████▏ | 665/809 [13:11<02:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212      0.434     0.1442        356        640:  82%|████████▏ | 666/809 [13:12<02:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4339     0.1442        380        640:  83%|████████▎ | 668/809 [13:14<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4341     0.1442        349        640:  83%|████████▎ | 670/809 [13:17<02:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4341     0.1441        228        640:  83%|████████▎ | 671/809 [13:18<02:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.21     0.4343     0.1439        136        640:  83%|████████▎ | 673/809 [13:20<02:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4346     0.1438        258        640:  83%|████████▎ | 675/809 [13:23<02:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4346     0.1439        409        640:  84%|████████▎ | 676/809 [13:24<02:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4345      0.144        280        640:  84%|████████▍ | 678/809 [13:26<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4344      0.144        256        640:  84%|████████▍ | 680/809 [13:29<02:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.21      0.434      0.144        271        640:  84%|████████▍ | 682/809 [13:31<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.21      0.434      0.144        173        640:  84%|████████▍ | 683/809 [13:32<02:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1443        278        640:  85%|████████▍ | 685/809 [13:35<02:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4336     0.1443        309        640:  85%|████████▍ | 687/809 [13:37<02:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1443        202        640:  85%|████████▌ | 688/809 [13:38<02:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1443        160        640:  85%|████████▌ | 690/809 [13:41<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4337     0.1444        427        640:  86%|████████▌ | 692/809 [13:43<02:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4338     0.1443        179        640:  86%|████████▌ | 693/809 [13:44<02:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1443        162        640:  86%|████████▌ | 695/809 [13:47<02:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4337     0.1443        253        640:  86%|████████▌ | 697/809 [13:49<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4337     0.1443        170        640:  86%|████████▋ | 698/809 [13:50<02:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4334     0.1446        422        640:  87%|████████▋ | 700/809 [13:53<02:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4334     0.1446        248        640:  87%|████████▋ | 702/809 [13:55<02:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1446        279        640:  87%|████████▋ | 703/809 [13:56<02:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4334     0.1449        279        640:  87%|████████▋ | 705/809 [13:59<02:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4333     0.1449        298        640:  87%|████████▋ | 707/809 [14:01<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4331     0.1449        279        640:  88%|████████▊ | 708/809 [14:02<01:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4332     0.1449        170        640:  88%|████████▊ | 710/809 [14:05<01:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213      0.433     0.1451        340        640:  88%|████████▊ | 712/809 [14:07<01:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4332      0.145        164        640:  88%|████████▊ | 714/809 [14:09<01:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4331     0.1452        379        640:  88%|████████▊ | 715/809 [14:10<01:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4333     0.1456        475        640:  89%|████████▊ | 717/809 [14:13<01:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.213     0.4332     0.1456        281        640:  89%|████████▉ | 718/809 [14:14<01:51,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4333     0.1456        182        640:  89%|████████▉ | 720/809 [14:17<01:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4332     0.1456        252        640:  89%|████████▉ | 722/809 [14:19<01:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.212     0.4333     0.1455        188        640:  89%|████████▉ | 724/809 [14:21<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4334     0.1454        135        640:  90%|████████▉ | 725/809 [14:22<01:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1454        177        640:  90%|████████▉ | 727/809 [14:25<01:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1454        162        640:  90%|█████████ | 729/809 [14:27<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.211     0.4336     0.1454        118        640:  90%|█████████ | 730/809 [14:28<01:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G       1.21     0.4337     0.1454        159        640:  90%|█████████ | 732/809 [14:31<01:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209      0.434     0.1452        131        640:  91%|█████████ | 734/809 [14:33<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209      0.434     0.1452        255        640:  91%|█████████ | 735/809 [14:34<01:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4344     0.1452        220        640:  91%|█████████ | 737/809 [14:37<01:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208     0.4346     0.1451         72        640:  91%|█████████▏| 739/809 [14:39<01:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208     0.4348     0.1452         45        640:  92%|█████████▏| 741/809 [14:41<01:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208     0.4349     0.1451        211        640:  92%|█████████▏| 742/809 [14:43<01:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207     0.4352      0.145        167        640:  92%|█████████▏| 744/809 [14:45<01:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207     0.4352      0.145        216        640:  92%|█████████▏| 746/809 [14:47<01:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207     0.4353     0.1449        171        640:  92%|█████████▏| 747/809 [14:48<01:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4355     0.1448        125        640:  93%|█████████▎| 749/809 [14:51<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4355     0.1448        335        640:  93%|█████████▎| 751/809 [14:53<01:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4356     0.1447        239        640:  93%|█████████▎| 753/809 [14:55<01:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4359     0.1446        102        640:  93%|█████████▎| 754/809 [14:57<01:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4358      0.145        250        640:  93%|█████████▎| 756/809 [14:59<01:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4357     0.1452        430        640:  94%|█████████▎| 758/809 [15:02<01:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4359     0.1451        133        640:  94%|█████████▍| 759/809 [15:03<01:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4358     0.1452        369        640:  94%|█████████▍| 761/809 [15:05<00:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4359      0.145        276        640:  94%|█████████▍| 763/809 [15:08<00:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206      0.436      0.145        119        640:  94%|█████████▍| 764/809 [15:09<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206      0.436     0.1449        175        640:  95%|█████████▍| 766/809 [15:11<00:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206      0.436     0.1449        236        640:  95%|█████████▍| 768/809 [15:14<00:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4362     0.1447        170        640:  95%|█████████▌| 770/809 [15:16<00:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.204     0.4364     0.1447         94        640:  95%|█████████▌| 771/809 [15:17<00:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.204     0.4363     0.1446        146        640:  96%|█████████▌| 773/809 [15:19<00:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4362     0.1446        284        640:  96%|█████████▌| 775/809 [15:22<00:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4361     0.1446        185        640:  96%|█████████▌| 776/809 [15:23<00:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4361     0.1445        294        640:  96%|█████████▌| 778/809 [15:26<00:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4361     0.1445        231        640:  96%|█████████▋| 780/809 [15:28<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4362     0.1444        173        640:  97%|█████████▋| 781/809 [15:29<00:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.204     0.4364     0.1443        214        640:  97%|█████████▋| 783/809 [15:31<00:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4363     0.1444        431        640:  97%|█████████▋| 785/809 [15:34<00:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4364     0.1443        170        640:  97%|█████████▋| 786/809 [15:35<00:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4363     0.1445        298        640:  97%|█████████▋| 788/809 [15:37<00:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.205     0.4364     0.1446        144        640:  98%|█████████▊| 790/809 [15:40<00:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4363     0.1445        159        640:  98%|█████████▊| 791/809 [15:41<00:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4363     0.1445        377        640:  98%|█████████▊| 793/809 [15:43<00:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207      0.436     0.1447        335        640:  98%|█████████▊| 795/809 [15:46<00:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.206     0.4361     0.1446        199        640:  98%|█████████▊| 796/809 [15:47<00:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207      0.436     0.1453        545        640:  99%|█████████▊| 798/809 [15:50<00:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207     0.4359     0.1452        324        640:  99%|█████████▉| 800/809 [15:52<00:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207      0.436     0.1452        362        640:  99%|█████████▉| 801/809 [15:53<00:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.207      0.436     0.1454        268        640:  99%|█████████▉| 803/809 [15:56<00:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.208     0.4357     0.1462        717        640:  99%|█████████▉| 804/809 [15:57<00:06,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4355     0.1468        534        640: 100%|█████████▉| 806/809 [16:00<00:03,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     29/300      13.5G      1.209     0.4357     0.1466        339        640: 100%|█████████▉| 808/809 [16:02<00:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:00<00:15,  4.09it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:03<00:14,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:05<00:12,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:11,  3.69it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  3.84it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:11<00:06,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:12<00:05,  3.71it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  84%|████████▍ | 58/69 [00:14<00:02,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300       7.6G      1.871     0.2705     0.3954        905        640:   0%|          | 1/809 [00:01<22:23,  1.66s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300      8.49G      1.479     0.3768     0.2412        298        640:   0%|          | 3/809 [00:03<17:11,  1.28s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.343     0.4278     0.1962        474        640:   1%|          | 5/809 [00:06<16:42,  1.25s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.303     0.4318     0.1844        194        640:   1%|          | 6/809 [00:07<16:18,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255       0.44     0.1695        255        640:   1%|          | 8/809 [00:09<15:58,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4454     0.1669        117        640:   1%|          | 10/809 [00:12<15:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253      0.437     0.1652        284        640:   1%|▏         | 11/809 [00:13<15:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4366     0.1721        182        640:   2%|▏         | 13/809 [00:16<16:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4371     0.1687        411        640:   2%|▏         | 15/809 [00:18<15:53,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4333     0.1684        342        640:   2%|▏         | 16/809 [00:19<16:07,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4358     0.1614        212        640:   2%|▏         | 18/809 [00:21<15:39,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4319     0.1596        315        640:   2%|▏         | 20/809 [00:24<16:04,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4326     0.1578        210        640:   3%|▎         | 21/809 [00:25<15:44,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.274     0.4208     0.1659        253        640:   3%|▎         | 23/809 [00:28<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.276     0.4179     0.1624        304        640:   3%|▎         | 25/809 [00:30<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.285     0.4139     0.1611        335        640:   3%|▎         | 27/809 [00:32<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.289     0.4118     0.1599        117        640:   3%|▎         | 28/809 [00:34<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.269     0.4178      0.155        111        640:   4%|▎         | 30/809 [00:36<15:13,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.271     0.4163     0.1537        256        640:   4%|▍         | 32/809 [00:38<15:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.278     0.4162     0.1556        398        640:   4%|▍         | 33/809 [00:39<15:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.286     0.4159     0.1613        361        640:   4%|▍         | 35/809 [00:42<15:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.274     0.4202     0.1586        258        640:   5%|▍         | 37/809 [00:44<15:27,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.264     0.4234     0.1569        101        640:   5%|▍         | 38/809 [00:45<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4303     0.1537         96        640:   5%|▍         | 40/809 [00:48<14:54,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.25     0.4305     0.1579        260        640:   5%|▌         | 42/809 [00:50<15:21,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.25     0.4298      0.158        287        640:   5%|▌         | 43/809 [00:51<15:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4321     0.1572        229        640:   6%|▌         | 45/809 [00:54<15:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4311     0.1558        209        640:   6%|▌         | 47/809 [00:56<14:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4322     0.1552        288        640:   6%|▌         | 49/809 [00:58<14:40,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4324     0.1541        359        640:   6%|▌         | 50/809 [00:59<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4287     0.1538        224        640:   6%|▋         | 52/809 [01:02<15:09,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.233     0.4305     0.1521        225        640:   7%|▋         | 54/809 [01:04<14:44,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.235     0.4291     0.1519        269        640:   7%|▋         | 55/809 [01:05<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4286     0.1521        168        640:   7%|▋         | 57/809 [01:08<14:42,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4263     0.1526        343        640:   7%|▋         | 59/809 [01:10<14:46,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4272     0.1522        254        640:   7%|▋         | 60/809 [01:11<14:56,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.231     0.4276     0.1534        216        640:   8%|▊         | 62/809 [01:14<15:03,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.219     0.4312     0.1515        207        640:   8%|▊         | 64/809 [01:16<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.218      0.431     0.1507        312        640:   8%|▊         | 65/809 [01:17<14:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.22     0.4296       0.15        466        640:   8%|▊         | 67/809 [01:20<14:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.215     0.4308     0.1486        234        640:   9%|▊         | 69/809 [01:22<14:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.216     0.4303     0.1491        609        640:   9%|▊         | 70/809 [01:23<14:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.213     0.4318     0.1485        134        640:   9%|▉         | 72/809 [01:26<14:37,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.207      0.434     0.1475        125        640:   9%|▉         | 74/809 [01:28<14:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.207     0.4339     0.1471        246        640:   9%|▉         | 75/809 [01:29<14:18,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.209     0.4329     0.1467        224        640:  10%|▉         | 77/809 [01:32<14:29,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.21     0.4318     0.1461        237        640:  10%|▉         | 79/809 [01:34<14:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.205     0.4336     0.1453        232        640:  10%|█         | 81/809 [01:36<14:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.202     0.4348     0.1449        130        640:  10%|█         | 82/809 [01:37<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.199     0.4352     0.1443         82        640:  10%|█         | 84/809 [01:40<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.195     0.4374     0.1436        117        640:  11%|█         | 86/809 [01:42<14:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.192      0.438     0.1427        253        640:  11%|█         | 88/809 [01:44<13:55,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.189     0.4386     0.1421        226        640:  11%|█         | 90/809 [01:47<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.185     0.4398     0.1415        117        640:  11%|█         | 91/809 [01:48<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.184       0.44     0.1413        185        640:  11%|█▏        | 93/809 [01:50<13:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.189     0.4385     0.1418        276        640:  12%|█▏        | 95/809 [01:53<13:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.191     0.4376     0.1417        294        640:  12%|█▏        | 96/809 [01:54<14:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.192     0.4372     0.1412        214        640:  12%|█▏        | 98/809 [01:56<14:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.194     0.4363     0.1409        192        640:  12%|█▏        | 100/809 [01:59<14:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.194     0.4361     0.1408        267        640:  12%|█▏        | 101/809 [02:00<14:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.193     0.4361     0.1412        184        640:  13%|█▎        | 103/809 [02:02<13:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.192     0.4361     0.1411        294        640:  13%|█▎        | 105/809 [02:05<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.194     0.4356     0.1407        250        640:  13%|█▎        | 107/809 [02:07<13:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.191      0.436     0.1403        141        640:  13%|█▎        | 108/809 [02:08<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.192     0.4359     0.1401         93        640:  14%|█▎        | 110/809 [02:10<13:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.19      0.437       0.14         92        640:  14%|█▍        | 112/809 [02:13<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.193     0.4367     0.1405        227        640:  14%|█▍        | 113/809 [02:14<13:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.196     0.4353     0.1412        437        640:  14%|█▍        | 115/809 [02:16<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.195     0.4349      0.141        321        640:  14%|█▍        | 117/809 [02:19<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.202      0.433     0.1429        241        640:  15%|█▍        | 119/809 [02:21<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.199     0.4342     0.1424        113        640:  15%|█▍        | 120/809 [02:22<13:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.196     0.4354     0.1423         29        640:  15%|█▌        | 122/809 [02:24<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.197      0.435     0.1424        243        640:  15%|█▌        | 124/809 [02:27<13:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G        1.2     0.4341     0.1429        290        640:  15%|█▌        | 125/809 [02:28<13:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.198     0.4341     0.1424        185        640:  16%|█▌        | 127/809 [02:30<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.197     0.4344     0.1422        194        640:  16%|█▌        | 129/809 [02:33<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.197     0.4344     0.1419        212        640:  16%|█▌        | 130/809 [02:34<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G        1.2     0.4334     0.1422        353        640:  16%|█▋        | 132/809 [02:36<13:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.198     0.4337     0.1424        110        640:  17%|█▋        | 134/809 [02:39<13:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.199     0.4334     0.1429        397        640:  17%|█▋        | 135/809 [02:41<13:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.197      0.434     0.1428        229        640:  17%|█▋        | 137/809 [02:42<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G        1.2      0.433     0.1432        351        640:  17%|█▋        | 139/809 [02:45<13:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.202     0.4323     0.1436        279        640:  17%|█▋        | 140/809 [02:46<13:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.202     0.4329     0.1433        230        640:  18%|█▊        | 142/809 [02:48<13:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.201     0.4335     0.1433        127        640:  18%|█▊        | 144/809 [02:51<13:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.202     0.4327     0.1431        217        640:  18%|█▊        | 146/809 [02:53<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.203     0.4325     0.1431        178        640:  18%|█▊        | 147/809 [02:54<12:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.201     0.4332     0.1427        238        640:  18%|█▊        | 149/809 [02:57<12:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.205     0.4323     0.1436        398        640:  19%|█▊        | 150/809 [02:58<13:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.209     0.4314     0.1458        458        640:  19%|█▉        | 152/809 [03:00<13:29,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.212      0.431     0.1463        332        640:  19%|█▉        | 154/809 [03:03<13:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.211     0.4314     0.1466        143        640:  19%|█▉        | 155/809 [03:04<13:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.213     0.4309     0.1467        267        640:  19%|█▉        | 157/809 [03:06<13:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.215     0.4305      0.147        372        640:  20%|█▉        | 159/809 [03:09<13:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.215     0.4296     0.1467        258        640:  20%|█▉        | 161/809 [03:11<12:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.213     0.4307     0.1463        228        640:  20%|██        | 162/809 [03:12<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.213     0.4308     0.1459        181        640:  20%|██        | 164/809 [03:15<12:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.214     0.4305     0.1459        439        640:  21%|██        | 166/809 [03:17<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.216     0.4297     0.1462        392        640:  21%|██        | 167/809 [03:18<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.217     0.4291     0.1468        156        640:  21%|██        | 169/809 [03:21<12:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.217     0.4288     0.1465        260        640:  21%|██        | 171/809 [03:23<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.219     0.4284     0.1465        404        640:  21%|██▏       | 172/809 [03:24<12:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.22     0.4282     0.1466        256        640:  22%|██▏       | 174/809 [03:27<12:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.221     0.4281     0.1467        256        640:  22%|██▏       | 176/809 [03:29<12:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.223     0.4273     0.1466        287        640:  22%|██▏       | 177/809 [03:30<12:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.224     0.4266     0.1463        338        640:  22%|██▏       | 179/809 [03:33<12:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.23     0.4248     0.1471        483        640:  22%|██▏       | 181/809 [03:35<12:56,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.231     0.4245     0.1469        262        640:  22%|██▏       | 182/809 [03:36<12:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.235     0.4233     0.1468        160        640:  23%|██▎       | 184/809 [03:39<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.235     0.4234     0.1464        121        640:  23%|██▎       | 186/809 [03:41<12:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4231     0.1462        157        640:  23%|██▎       | 187/809 [03:42<12:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4222     0.1464        375        640:  23%|██▎       | 189/809 [03:45<12:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4214      0.147        266        640:  24%|██▎       | 191/809 [03:47<12:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4215     0.1467        147        640:  24%|██▎       | 192/809 [03:48<12:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4215     0.1465        264        640:  24%|██▍       | 194/809 [03:51<12:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4218      0.146        164        640:  24%|██▍       | 196/809 [03:53<12:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.422     0.1458        210        640:  24%|██▍       | 197/809 [03:54<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4214     0.1457        152        640:  25%|██▍       | 199/809 [03:57<12:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4217     0.1452        232        640:  25%|██▍       | 201/809 [03:59<12:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4211     0.1452        186        640:  25%|██▍       | 202/809 [04:00<12:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4202      0.145        257        640:  25%|██▌       | 204/809 [04:03<12:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4205     0.1447        155        640:  25%|██▌       | 206/809 [04:05<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4208     0.1445        203        640:  26%|██▌       | 207/809 [04:06<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4208      0.145        253        640:  26%|██▌       | 209/809 [04:09<11:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4211     0.1447        344        640:  26%|██▌       | 211/809 [04:11<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4213     0.1446        200        640:  26%|██▋       | 213/809 [04:13<11:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4213     0.1443        222        640:  26%|██▋       | 214/809 [04:14<11:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4216     0.1439        181        640:  27%|██▋       | 216/809 [04:17<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4213     0.1439        114        640:  27%|██▋       | 218/809 [04:19<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238      0.421     0.1439        418        640:  27%|██▋       | 219/809 [04:20<11:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4208     0.1439        272        640:  27%|██▋       | 221/809 [04:23<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4213     0.1436        312        640:  28%|██▊       | 223/809 [04:25<11:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236      0.422     0.1431        128        640:  28%|██▊       | 225/809 [04:27<11:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.235     0.4221      0.143        149        640:  28%|██▊       | 226/809 [04:29<11:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4218     0.1429        251        640:  28%|██▊       | 228/809 [04:31<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.235     0.4223     0.1427        253        640:  28%|██▊       | 230/809 [04:33<11:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4226     0.1426        156        640:  29%|██▊       | 231/809 [04:34<11:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.233     0.4229     0.1423        189        640:  29%|██▉       | 233/809 [04:37<11:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.232     0.4236      0.142         89        640:  29%|██▉       | 235/809 [04:39<11:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.23     0.4243     0.1418        158        640:  29%|██▉       | 236/809 [04:40<11:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.232     0.4233     0.1422        147        640:  29%|██▉       | 238/809 [04:43<11:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.231     0.4236     0.1419        170        640:  30%|██▉       | 240/809 [04:45<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4229     0.1424         89        640:  30%|██▉       | 242/809 [04:47<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4218     0.1429        276        640:  30%|███       | 244/809 [04:50<11:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4223     0.1427        221        640:  30%|███       | 245/809 [04:51<11:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4231     0.1426        218        640:  31%|███       | 247/809 [04:53<11:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234      0.423     0.1425        203        640:  31%|███       | 249/809 [04:56<10:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4231     0.1422        142        640:  31%|███       | 251/809 [04:58<10:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4229     0.1421        206        640:  31%|███       | 252/809 [04:59<11:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.234     0.4229     0.1421        381        640:  31%|███▏      | 254/809 [05:02<10:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4222     0.1423        577        640:  32%|███▏      | 256/809 [05:04<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4218     0.1425        322        640:  32%|███▏      | 257/809 [05:05<11:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4219     0.1428         86        640:  32%|███▏      | 259/809 [05:08<10:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4217     0.1429        442        640:  32%|███▏      | 260/809 [05:09<11:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4209      0.143        257        640:  32%|███▏      | 262/809 [05:11<11:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4203     0.1434        122        640:  33%|███▎      | 264/809 [05:14<10:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4198     0.1435        163        640:  33%|███▎      | 266/809 [05:16<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4198     0.1436        217        640:  33%|███▎      | 267/809 [05:17<10:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4191     0.1435        394        640:  33%|███▎      | 269/809 [05:20<10:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4189     0.1433        219        640:  33%|███▎      | 271/809 [05:22<10:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4183     0.1439        526        640:  34%|███▎      | 272/809 [05:23<10:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251      0.418     0.1439        377        640:  34%|███▍      | 274/809 [05:26<10:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251      0.418     0.1438        188        640:  34%|███▍      | 276/809 [05:28<10:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4181      0.144        207        640:  34%|███���      | 277/809 [05:29<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.254     0.4175     0.1447        314        640:  34%|███▍      | 279/809 [05:32<10:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255     0.4171     0.1448        236        640:  35%|███▍      | 281/809 [05:34<10:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4167     0.1453        440        640:  35%|███▍      | 282/809 [05:35<10:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4163     0.1461        465        640:  35%|███▌      | 284/809 [05:38<11:10,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26      0.416     0.1462        440        640:  35%|███▌      | 285/809 [05:39<11:05,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4164      0.146        213        640:  35%|███▌      | 287/809 [05:41<10:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26     0.4164     0.1465        241        640:  36%|███▌      | 289/809 [05:44<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26     0.4164     0.1465        155        640:  36%|███▌      | 291/809 [05:46<10:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26     0.4162     0.1465        246        640:  36%|███▌      | 292/809 [05:47<10:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.262     0.4156     0.1468        292        640:  36%|███▋      | 294/809 [05:50<10:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.262     0.4154     0.1467        210        640:  36%|███▋      | 295/809 [05:51<10:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.263     0.4151     0.1478        586        640:  37%|███▋      | 297/809 [05:54<11:10,  1.31s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.261     0.4158     0.1477         86        640:  37%|███▋      | 298/809 [05:56<10:46,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26     0.4165     0.1476         50        640:  37%|███▋      | 300/809 [05:57<10:30,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.262     0.4157     0.1481        376        640:  37%|███▋      | 302/809 [06:00<10:35,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.261     0.4161      0.148        131        640:  37%|███▋      | 303/809 [06:01<10:35,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.261     0.4165     0.1481        132        640:  38%|███▊      | 305/809 [06:04<10:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.261     0.4166      0.148        272        640:  38%|███▊      | 307/809 [06:06<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259      0.417     0.1479        102        640:  38%|███▊      | 308/809 [06:07<10:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4173     0.1478        185        640:  38%|███▊      | 310/809 [06:09<09:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258     0.4174     0.1477        162        640:  39%|███▊      | 312/809 [06:12<09:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4171     0.1478        347        640:  39%|███▊      | 313/809 [06:13<09:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.26     0.4168     0.1482        140        640:  39%|███▉      | 315/809 [06:15<09:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4171     0.1481        333        640:  39%|███▉      | 317/809 [06:18<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258     0.4174     0.1479        231        640:  39%|███▉      | 319/809 [06:20<09:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258     0.4172     0.1478        208        640:  40%|███▉      | 320/809 [06:21<09:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258     0.4168     0.1478        272        640:  40%|███▉      | 322/809 [06:24<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4183     0.1477        183        640:  40%|████      | 324/809 [06:26<09:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4184     0.1476        211        640:  40%|████      | 325/809 [06:27<09:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4182     0.1478        314        640:  40%|████      | 327/809 [06:30<09:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4183     0.1476        280        640:  41%|████      | 329/809 [06:32<09:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4182     0.1476        350        640:  41%|████      | 330/809 [06:33<09:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4179     0.1476        481        640:  41%|████      | 332/809 [06:36<09:46,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4175     0.1476        192        640:  41%|████▏     | 334/809 [06:38<09:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4177     0.1475        256        640:  41%|████▏     | 335/809 [06:39<09:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4178     0.1473        209        640:  42%|████▏     | 337/809 [06:42<09:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.254     0.4184      0.147        308        640:  42%|████▏     | 339/809 [06:44<09:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255     0.4182     0.1472        250        640:  42%|████▏     | 340/809 [06:45<09:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4176     0.1477        187        640:  42%|████▏     | 342/809 [06:48<09:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4173     0.1479        309        640:  43%|████▎     | 344/809 [06:50<09:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259     0.4171     0.1483        213        640:  43%|████▎     | 345/809 [06:51<09:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.259      0.417     0.1481        207        640:  43%|████▎     | 347/809 [06:54<09:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258     0.4173     0.1479        140        640:  43%|████▎     | 349/809 [06:56<09:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.258      0.417     0.1478        191        640:  43%|████▎     | 351/809 [06:58<08:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.257     0.4172     0.1476        210        640:  44%|████▎     | 352/809 [06:59<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4175     0.1475        205        640:  44%|████▍     | 354/809 [07:02<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4175     0.1477        289        640:  44%|████▍     | 356/809 [07:04<09:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255     0.4178     0.1477        112        640:  44%|████▍     | 357/809 [07:05<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4177     0.1475        340        640:  44%|████▍     | 359/809 [07:08<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4176     0.1475        251        640:  45%|████▍     | 361/809 [07:10<08:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255     0.4176     0.1475        146        640:  45%|████▍     | 362/809 [07:11<08:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.256     0.4177     0.1475        350        640:  45%|████▍     | 364/809 [07:14<08:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.255     0.4179     0.1474        198        640:  45%|████▌     | 366/809 [07:16<08:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.254     0.4182     0.1472        220        640:  45%|████▌     | 367/809 [07:17<08:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4186     0.1471        106        640:  46%|████▌     | 369/809 [07:20<08:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4191     0.1468        195        640:  46%|████▌     | 371/809 [07:22<08:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.25     0.4193     0.1467        130        640:  46%|████▌     | 373/809 [07:24<08:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4189     0.1472        512        640:  46%|████▋     | 375/809 [07:27<08:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251      0.419      0.147        222        640:  46%|████▋     | 376/809 [07:28<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.25     0.4191     0.1468        313        640:  47%|████▋     | 378/809 [07:30<08:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4194     0.1466        185        640:  47%|████▋     | 380/809 [07:33<08:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4192     0.1466        262        640:  47%|████▋     | 381/809 [07:34<08:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4187     0.1466        330        640:  47%|████▋     | 383/809 [07:36<08:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4188     0.1465        178        640:  48%|████▊     | 385/809 [07:39<08:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4188     0.1464         98        640:  48%|████▊     | 386/809 [07:41<08:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4184     0.1466        469        640:  48%|████▊     | 388/809 [07:42<08:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4185     0.1465        117        640:  48%|████▊     | 390/809 [07:45<08:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4185     0.1463        165        640:  48%|████▊     | 392/809 [07:47<08:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4183     0.1463        285        640:  49%|████▊     | 393/809 [07:48<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4186     0.1462        142        640:  49%|████▉     | 395/809 [07:50<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4182     0.1461        146        640:  49%|████▉     | 397/809 [07:53<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4183     0.1461         59        640:  49%|████▉     | 398/809 [07:54<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4182     0.1459        206        640:  49%|████▉     | 400/809 [07:56<07:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4182     0.1459        165        640:  50%|████▉     | 402/809 [07:59<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4183     0.1458        165        640:  50%|████▉     | 403/809 [08:00<07:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4183     0.1456        263        640:  50%|█████     | 405/809 [08:02<08:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4184     0.1454        185        640:  50%|█████     | 407/809 [08:05<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4185     0.1451        146        640:  51%|█████     | 409/809 [08:07<07:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4186     0.1451        244        640:  51%|█████     | 410/809 [08:08<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4186     0.1457        374        640:  51%|█████     | 412/809 [08:11<08:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4185     0.1457        328        640:  51%|█████     | 414/809 [08:13<07:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4184     0.1456        162        640:  51%|█████▏    | 415/809 [08:14<07:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4183     0.1461        297        640:  52%|█████▏    | 417/809 [08:16<07:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.254     0.4179     0.1462        287        640:  52%|█████▏    | 419/809 [08:19<07:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.254     0.4181     0.1461        169        640:  52%|█████▏    | 420/809 [08:20<07:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4185     0.1459        242        640:  52%|█████▏    | 422/809 [08:22<07:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.253     0.4184     0.1458        294        640:  52%|█████▏    | 424/809 [08:25<07:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4189     0.1456        101        640:  53%|█████▎    | 425/809 [08:26<07:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4192     0.1454        159        640:  53%|█████▎    | 427/809 [08:28<07:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4192     0.1457        116        640:  53%|█████▎    | 429/809 [08:31<07:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.252     0.4193     0.1456        179        640:  53%|█████▎    | 430/809 [08:32<07:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.251     0.4197     0.1454        246        640:  53%|█████▎    | 432/809 [08:34<07:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.25     0.4201     0.1453        285        640:  54%|█████▎    | 434/809 [08:37<07:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4205     0.1452        145        640:  54%|█████▍    | 435/809 [08:38<07:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248      0.421     0.1451        219        640:  54%|█████▍    | 437/809 [08:40<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4212      0.145        284        640:  54%|█████▍    | 439/809 [08:43<07:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4219     0.1447         91        640:  55%|█████▍    | 441/809 [08:45<07:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246      0.422     0.1448        318        640:  55%|█████▍    | 442/809 [08:46<07:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246      0.422     0.1447        341        640:  55%|█████▍    | 444/809 [08:49<07:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4224     0.1444        176        640:  55%|█████▌    | 446/809 [08:51<07:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4225     0.1444        283        640:  55%|█████▌    | 447/809 [08:52<07:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4224     0.1445        412        640:  56%|█████▌    | 449/809 [08:54<07:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4225     0.1444        339        640:  56%|█████▌    | 451/809 [08:57<07:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4225     0.1442        204        640:  56%|█████▌    | 452/809 [08:58<07:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4226     0.1441        344        640:  56%|█████▌    | 454/809 [09:00<07:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4222     0.1442        179        640:  56%|█████▋    | 456/809 [09:03<06:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245      0.422     0.1443        435        640:  56%|█████▋    | 457/809 [09:04<07:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245      0.422     0.1442        250        640:  57%|█████▋    | 459/809 [09:06<06:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4222     0.1441        320        640:  57%|█████▋    | 461/809 [09:09<06:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4219     0.1443        405        640:  57%|█████▋    | 462/809 [09:10<06:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4217     0.1443        282        640:  57%|█████▋    | 464/809 [09:12<06:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4218     0.1446        380        640:  58%|█████▊    | 466/809 [09:15<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246      0.422     0.1446        337        640:  58%|█████▊    | 468/809 [09:17<06:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246      0.422     0.1445        244        640:  58%|█████▊    | 469/809 [09:18<06:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4224     0.1444        257        640:  58%|█████▊    | 471/809 [09:21<06:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4226     0.1447        396        640:  58%|█████▊    | 473/809 [09:23<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4225     0.1448        318        640:  59%|█████▊    | 474/809 [09:24<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4223     0.1451        245        640:  59%|█████▉    | 476/809 [09:27<06:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4222      0.145        300        640:  59%|█████▉    | 478/809 [09:29<06:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4222      0.145        183        640:  59%|█████▉    | 479/809 [09:30<06:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4223      0.145        187        640:  59%|█████▉    | 481/809 [09:33<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4227     0.1449         89        640:  60%|█████▉    | 483/809 [09:35<06:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4227     0.1449        343        640:  60%|█████▉    | 484/809 [09:36<06:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4226     0.1448        332        640:  60%|██████    | 486/809 [09:39<06:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4228     0.1448        263        640:  60%|██████    | 488/809 [09:41<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1446         88        640:  61%|██████    | 490/809 [09:43<06:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4233     0.1445         91        640:  61%|██████    | 491/809 [09:44<06:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1445        246        640:  61%|██████    | 493/809 [09:47<06:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1445        262        640:  61%|██████    | 495/809 [09:49<06:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1444        192        640:  61%|██████▏   | 496/809 [09:50<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243      0.423     0.1443        464        640:  62%|██████▏   | 498/809 [09:53<06:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243      0.423     0.1445        289        640:  62%|██████▏   | 500/809 [09:55<06:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4228     0.1446        349        640:  62%|██████▏   | 501/809 [09:56<06:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4225     0.1445        331        640:  62%|██████▏   | 503/809 [09:59<06:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4228     0.1443        292        640:  62%|██████▏   | 505/809 [10:01<06:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4228     0.1443        203        640:  63%|██████▎   | 506/809 [10:02<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1441        129        640:  63%|██████▎   | 508/809 [10:05<06:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243      0.423     0.1442        257        640:  63%|██████▎   | 510/809 [10:07<05:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1443        262        640:  63%|██████▎   | 511/809 [10:08<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4231     0.1447        494        640:  63%|██████▎   | 513/809 [10:11<05:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4236     0.1445        243        640:  64%|██████▎   | 515/809 [10:13<05:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4239     0.1444         83        640:  64%|██████▍   | 516/809 [10:14<05:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4241     0.1442        164        640:  64%|██████▍   | 518/809 [10:17<05:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4244     0.1441        124        640:  64%|██████▍   | 520/809 [10:19<05:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4241     0.1445        636        640:  64%|██████▍   | 521/809 [10:20<05:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4242     0.1446        626        640:  65%|██████▍   | 523/809 [10:23<05:57,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4237     0.1452        679        640:  65%|██████▍   | 525/809 [10:25<05:59,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4239     0.1453         98        640:  65%|██████▌   | 527/809 [10:28<05:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242      0.424     0.1453        145        640:  65%|██████▌   | 528/809 [10:29<05:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4239     0.1452        153        640:  66%|██████▌   | 530/809 [10:31<05:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4234     0.1453        319        640:  66%|██████▌   | 532/809 [10:34<05:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4235     0.1452        183        640:  66%|██████▌   | 533/809 [10:35<05:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4235     0.1452        273        640:  66%|██████▌   | 535/809 [10:37<05:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4236     0.1451        264        640:  66%|██████▋   | 537/809 [10:40<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4236     0.1451        376        640:  67%|██████▋   | 538/809 [10:41<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4241     0.1451         46        640:  67%|██████▋   | 540/809 [10:43<05:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4242      0.145        217        640:  67%|██████▋   | 542/809 [10:46<05:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4245     0.1449         83        640:  67%|██████▋   | 543/809 [10:47<05:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4244      0.145        475        640:  67%|██████▋   | 545/809 [10:49<05:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4245     0.1449        227        640:  68%|██████▊   | 547/809 [10:52<05:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4245     0.1451        175        640:  68%|██████▊   | 548/809 [10:53<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4244      0.145        173        640:  68%|██████▊   | 550/809 [10:55<05:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4239     0.1453        430        640:  68%|██████▊   | 552/809 [10:58<05:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4239     0.1452        212        640:  68%|██████▊   | 554/809 [11:00<05:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4237     0.1452        354        640:  69%|██████▊   | 555/809 [11:01<05:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4238     0.1454        214        640:  69%|██████▉   | 557/809 [11:04<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4238     0.1454        209        640:  69%|██████▉   | 559/809 [11:06<04:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4241     0.1453        128        640:  69%|██████▉   | 560/809 [11:07<04:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241      0.424     0.1453        363        640:  69%|██████▉   | 562/809 [11:09<04:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4242     0.1452        413        640:  70%|██████▉   | 564/809 [11:12<04:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4243     0.1452        110        640:  70%|██████▉   | 565/809 [11:13<04:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4246     0.1451        233        640:  70%|███████   | 567/809 [11:15<04:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4247     0.1451        253        640:  70%|███████   | 569/809 [11:18<04:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4246     0.1452        386        640:  70%|███████   | 570/809 [11:19<04:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4251     0.1451        189        640:  71%|███████   | 572/809 [11:21<04:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4255     0.1449        151        640:  71%|███████   | 574/809 [11:24<04:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4253     0.1451        157        640:  71%|███████   | 575/809 [11:26<04:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4254     0.1451        205        640:  71%|███████▏  | 577/809 [11:27<04:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4252     0.1452        359        640:  72%|███████▏  | 579/809 [11:30<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4252     0.1451        206        640:  72%|███████▏  | 581/809 [11:32<04:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4253     0.1451        229        640:  72%|███████▏  | 582/809 [11:33<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4251     0.1454        496        640:  72%|███████▏  | 584/809 [11:36<04:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.425     0.1454        195        640:  72%|███████▏  | 586/809 [11:38<04:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4252     0.1455        167        640:  73%|███████▎  | 587/809 [11:39<04:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4252     0.1456        188        640:  73%|███████▎  | 589/809 [11:42<04:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.425     0.1455        293        640:  73%|███████▎  | 591/809 [11:44<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4248     0.1455        267        640:  73%|███████▎  | 592/809 [11:45<04:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.425     0.1455        202        640:  73%|███████▎  | 594/809 [11:47<04:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4246     0.1458        422        640:  74%|███████▎  | 596/809 [11:50<04:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4247     0.1458        275        640:  74%|███████▍  | 597/809 [11:51<04:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4249     0.1457        123        640:  74%|███████▍  | 599/809 [11:53<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.425     0.1456        230        640:  74%|███████▍  | 601/809 [11:56<04:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4248     0.1461        533        640:  74%|███████▍  | 602/809 [11:57<04:14,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24      0.425      0.146        271        640:  75%|███████▍  | 604/809 [12:00<04:21,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4251     0.1459        209        640:  75%|███████▍  | 605/809 [12:01<04:13,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4252     0.1458        184        640:  75%|███████▌  | 607/809 [12:03<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4249     0.1461        356        640:  75%|███████▌  | 609/809 [12:06<04:05,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4247     0.1464        740        640:  75%|███████▌  | 610/809 [12:07<04:14,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4249     0.1463        108        640:  76%|███████▌  | 612/809 [12:10<04:05,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4248     0.1463        455        640:  76%|███████▌  | 614/809 [12:12<03:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4247     0.1463        288        640:  76%|███████▌  | 615/809 [12:13<03:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4248     0.1464        148        640:  76%|███████▋  | 617/809 [12:16<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4246     0.1468        185        640:  77%|███████▋  | 619/809 [12:18<03:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4247      0.147        371        640:  77%|███████▋  | 620/809 [12:19<03:54,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4245     0.1471        802        640:  77%|███████▋  | 622/809 [12:22<03:58,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4243     0.1472        235        640:  77%|███████▋  | 623/809 [12:23<03:51,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4246     0.1471         99        640:  77%|███████▋  | 625/809 [12:25<03:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4247     0.1471        129        640:  78%|███████▊  | 627/809 [12:28<03:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4246     0.1472        255        640:  78%|███████▊  | 629/809 [12:30<03:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4247     0.1474        381        640:  78%|███████▊  | 630/809 [12:31<03:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241      0.425     0.1473        298        640:  78%|███████▊  | 632/809 [12:34<03:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4251     0.1474        266        640:  78%|███████▊  | 634/809 [12:36<03:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4254     0.1473        164        640:  78%|███████▊  | 635/809 [12:37<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4257     0.1471        127        640:  79%|███████▊  | 637/809 [12:40<03:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4254     0.1476        280        640:  79%|███████▉  | 639/809 [12:42<03:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4254     0.1475        176        640:  79%|███████▉  | 640/809 [12:43<03:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4255     0.1475        259        640:  79%|███████▉  | 642/809 [12:46<03:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4254     0.1475        113        640:  80%|███████▉  | 644/809 [12:48<03:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4252     0.1476        272        640:  80%|███████▉  | 645/809 [12:49<03:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242      0.425     0.1475        247        640:  80%|███████▉  | 647/809 [12:52<03:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4252     0.1475        134        640:  80%|████████  | 649/809 [12:54<03:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4252     0.1475        257        640:  80%|████████  | 650/809 [12:55<03:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244      0.425     0.1476        243        640:  81%|████████  | 652/809 [12:58<03:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4249     0.1477        312        640:  81%|████████  | 654/809 [13:00<03:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4251     0.1476        144        640:  81%|████████  | 655/809 [13:01<03:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4249     0.1479        238        640:  81%|████████  | 657/809 [13:04<03:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4248     0.1478        327        640:  81%|████████▏ | 659/809 [13:06<03:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4247     0.1479        249        640:  82%|████████▏ | 660/809 [13:07<03:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4246     0.1478        333        640:  82%|████████▏ | 662/809 [13:10<02:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4244      0.148        299        640:  82%|████████▏ | 664/809 [13:12<02:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4244     0.1479        239        640:  82%|████████▏ | 665/809 [13:13<02:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4241     0.1479        366        640:  82%|████████▏ | 667/809 [13:16<02:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4238     0.1485        613        640:  83%|████████▎ | 668/809 [13:17<02:59,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248      0.424     0.1484        273        640:  83%|████████▎ | 670/809 [13:20<02:52,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248      0.424     0.1483        305        640:  83%|████████▎ | 672/809 [13:22<02:49,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4239     0.1484        248        640:  83%|████████▎ | 674/809 [13:24<02:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4241     0.1483        161        640:  83%|████████▎ | 675/809 [13:26<02:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248      0.424     0.1483        297        640:  84%|████████▎ | 677/809 [13:28<02:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4238     0.1483        251        640:  84%|████████▍ | 679/809 [13:30<02:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4238     0.1483        406        640:  84%|████████▍ | 681/809 [13:33<02:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.249     0.4238     0.1483        258        640:  84%|████████▍ | 682/809 [13:34<02:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248      0.424     0.1481        158        640:  85%|████████▍ | 684/809 [13:36<02:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.248     0.4241     0.1482        163        640:  85%|████████▍ | 686/809 [13:39<02:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4242     0.1481        119        640:  85%|████████▍ | 687/809 [13:40<02:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4245      0.148        268        640:  85%|████████▌ | 689/809 [13:42<02:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4242     0.1481        306        640:  85%|████████▌ | 691/809 [13:45<02:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4243     0.1481        150        640:  86%|████████▌ | 692/809 [13:46<02:24,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.247     0.4244      0.148        129        640:  86%|████████▌ | 694/809 [13:48<02:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4247     0.1479        229        640:  86%|████████▌ | 696/809 [13:51<02:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4247      0.148        183        640:  86%|████████▋ | 698/809 [13:53<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246     0.4249     0.1479        141        640:  86%|████████▋ | 699/809 [13:54<02:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.246      0.425     0.1478        150        640:  87%|████████▋ | 701/809 [13:56<02:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245      0.425     0.1479        392        640:  87%|████████▋ | 703/809 [13:59<02:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4253     0.1479        177        640:  87%|████████▋ | 704/809 [14:00<02:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4253     0.1479        132        640:  87%|████████▋ | 706/809 [14:02<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.245     0.4254     0.1479        145        640:  88%|████████▊ | 708/809 [14:05<02:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4255     0.1479        227        640:  88%|████████▊ | 709/809 [14:06<01:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4256     0.1479        328        640:  88%|████████▊ | 711/809 [14:08<01:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243     0.4259     0.1477        255        640:  88%|████████▊ | 713/809 [14:11<01:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.244     0.4258     0.1478        336        640:  88%|████████▊ | 714/809 [14:12<01:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.243      0.426     0.1477        209        640:  89%|████████▊ | 716/809 [14:14<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4263     0.1476        150        640:  89%|████████▉ | 718/809 [14:17<01:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4265     0.1476        262        640:  89%|████████▉ | 719/809 [14:18<01:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.242     0.4266     0.1476        387        640:  89%|████████▉ | 721/809 [14:20<01:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241     0.4268     0.1476        209        640:  89%|████████▉ | 723/809 [14:23<01:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.241      0.427     0.1475        265        640:  90%|████████▉ | 725/809 [14:25<01:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4272     0.1475        139        640:  90%|████████▉ | 726/809 [14:26<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4274     0.1474        386        640:  90%|████████▉ | 728/809 [14:29<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4275     0.1472        194        640:  90%|█████████ | 730/809 [14:31<01:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4275     0.1473        331        640:  90%|█████████ | 731/809 [14:32<01:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4276     0.1472        276        640:  91%|█████████ | 733/809 [14:35<01:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4278     0.1471        179        640:  91%|█████████ | 735/809 [14:37<01:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4281      0.147        190        640:  91%|█████████ | 736/809 [14:38<01:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4282     0.1469        148        640:  91%|█████████ | 738/809 [14:40<01:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4283     0.1468        157        640:  91%|█████████▏| 740/809 [14:43<01:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4282     0.1469        580        640:  92%|█████████▏| 741/809 [14:44<01:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4281     0.1468        311        640:  92%|█████████▏| 743/809 [14:46<01:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237      0.428     0.1468        184        640:  92%|█████████▏| 745/809 [14:49<01:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4279     0.1469        241        640:  92%|█████████▏| 746/809 [14:50<01:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4278     0.1468        150        640:  92%|█████████▏| 748/809 [14:52<01:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4277     0.1468        306        640:  93%|█████████▎| 750/809 [14:55<01:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4276     0.1467        320        640:  93%|█████████▎| 751/809 [14:56<01:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4275     0.1467        144        640:  93%|█████████▎| 753/809 [14:58<01:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4272     0.1469        255        640:  93%|█████████▎| 755/809 [15:01<01:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4272     0.1468        361        640:  93%|█████████▎| 756/809 [15:02<01:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4271      0.147        375        640:  94%|█████████▎| 758/809 [15:04<01:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4271     0.1469        174        640:  94%|█████████▍| 760/809 [15:07<00:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4272     0.1469        131        640:  94%|█████████▍| 761/809 [15:08<00:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4273     0.1468        140        640:  94%|█████████▍| 763/809 [15:10<00:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4272     0.1469        338        640:  95%|█████████▍| 765/809 [15:13<00:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4271     0.1473        442        640:  95%|█████████▍| 766/809 [15:14<00:53,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4272     0.1472        246        640:  95%|█████████▍| 768/809 [15:17<00:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4272     0.1472        454        640:  95%|█████████▌| 770/809 [15:19<00:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.427     0.1475        571        640:  95%|█████████▌| 771/809 [15:20<00:46,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239      0.427     0.1475        417        640:  96%|█████████▌| 773/809 [15:23<00:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4269     0.1475        310        640:  96%|█████████▌| 775/809 [15:25<00:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4269     0.1475        313        640:  96%|█████████▌| 776/809 [15:26<00:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4269     0.1475        267        640:  96%|█████████▌| 778/809 [15:29<00:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4269     0.1477        238        640:  96%|█████████▋| 780/809 [15:31<00:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24     0.4268     0.1477        301        640:  97%|█████████▋| 781/809 [15:32<00:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G       1.24      0.427     0.1476        233        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4273     0.1475        236        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4274     0.1475        224        640:  97%|█████████▋| 786/809 [15:38<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.239     0.4274     0.1474        212        640:  97%|█████████▋| 788/809 [15:40<00:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4275     0.1473        397        640:  98%|█████████▊| 790/809 [15:43<00:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.238     0.4276     0.1472        255        640:  98%|█████████▊| 792/809 [15:45<00:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4278     0.1471        179        640:  98%|█████████▊| 793/809 [15:46<00:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4279     0.1471        360        640:  98%|█████████▊| 795/809 [15:49<00:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4282      0.147        103        640:  99%|█████████▊| 797/809 [15:51<00:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4281      0.147        267        640:  99%|█████████▊| 798/809 [15:52<00:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4282     0.1469        236        640:  99%|█████████▉| 800/809 [15:55<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.236     0.4282     0.1468        342        640:  99%|█████████▉| 802/809 [15:57<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237      0.428      0.147        195        640:  99%|█████████▉| 803/809 [15:58<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237      0.428     0.1469        194        640: 100%|█████████▉| 805/809 [16:01<00:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237     0.4282     0.1468        332        640: 100%|█████████▉| 807/809 [16:03<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     30/300        13G      1.237      0.428     0.1469        284        640: 100%|██████████| 809/809 [16:05<00:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  12%|█▏        | 8/69 [00:01<00:14,  4.08it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  22%|██▏       | 15/69 [00:03<00:13,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  33%|███▎      | 23/69 [00:05<00:11,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  45%|████▍     | 31/69 [00:07<00:09,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  57%|█████▋    | 39/69 [00:09<00:07,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.64it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  78%|███████▊  | 54/69 [00:13<00:03,  3.81it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:16<00:01,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|████���█████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.422      0.286       0.28      0.143
-     31/300      13.2G      1.026     0.4687    0.08351        268        640:   0%|          | 1/809 [00:01<15:56,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.043     0.4975     0.1002        200        640:   0%|          | 3/809 [00:03<16:09,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.078     0.4782     0.1047        145        640:   0%|          | 4/809 [00:04<15:50,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.244     0.4262      0.129        424        640:   1%|          | 6/809 [00:07<15:50,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.365     0.3924     0.1512        157        640:   1%|          | 8/809 [00:09<16:08,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.398     0.3785     0.1497        334        640:   1%|          | 9/809 [00:10<16:31,  1.24s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.389     0.3877     0.1427        166        640:   1%|▏         | 11/809 [00:13<16:12,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.313     0.4124     0.1354        115        640:   2%|▏         | 13/809 [00:15<16:16,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.31     0.4106     0.1336        225        640:   2%|▏         | 14/809 [00:16<15:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.274     0.4171     0.1311        103        640:   2%|▏         | 16/809 [00:19<15:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.273     0.4183     0.1339        323        640:   2%|▏         | 18/809 [00:21<16:26,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.255     0.4278     0.1346        202        640:   2%|▏         | 19/809 [00:23<16:25,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4347     0.1334        272        640:   3%|▎         | 21/809 [00:25<15:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.246     0.4337     0.1342        322        640:   3%|▎         | 23/809 [00:27<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.255     0.4318     0.1399        264        640:   3%|▎         | 25/809 [00:30<15:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4351     0.1395        260        640:   3%|▎         | 26/809 [00:31<15:36,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.254     0.4316     0.1428        406        640:   3%|▎         | 28/809 [00:33<15:49,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.25     0.4328     0.1416        337        640:   4%|▎         | 30/809 [00:36<15:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.255     0.4307     0.1425        488        640:   4%|▍         | 31/809 [00:37<15:43,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.256     0.4296      0.143        560        640:   4%|▍         | 33/809 [00:39<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.256     0.4297     0.1429        427        640:   4%|▍         | 35/809 [00:42<15:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.25     0.4306     0.1415        170        640:   4%|▍         | 36/809 [00:43<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.245     0.4315     0.1419        392        640:   5%|▍         | 38/809 [00:45<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4308     0.1419        263        640:   5%|▍         | 40/809 [00:48<15:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4312     0.1406        285        640:   5%|▌         | 41/809 [00:49<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4346     0.1388        141        640:   5%|▌         | 43/809 [00:51<15:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.214      0.438     0.1365         96        640:   6%|▌         | 45/809 [00:54<15:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.216     0.4376     0.1356        225        640:   6%|▌         | 47/809 [00:56<14:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.212     0.4377      0.135        187        640:   6%|▌         | 48/809 [00:57<14:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.214     0.4373     0.1341        140        640:   6%|▌         | 50/809 [00:59<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.203     0.4398     0.1324        146        640:   6%|▋         | 52/809 [01:02<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.204     0.4394     0.1324        314        640:   7%|▋         | 53/809 [01:03<14:48,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.199     0.4425     0.1318        233        640:   7%|▋         | 55/809 [01:05<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.199      0.442     0.1319        164        640:   7%|▋         | 57/809 [01:08<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.195     0.4428     0.1314        209        640:   7%|▋         | 58/809 [01:09<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.192     0.4437     0.1314        118        640:   7%|▋         | 60/809 [01:11<14:59,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.196     0.4437     0.1327         28        640:   8%|▊         | 62/809 [01:14<14:54,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.198     0.4422     0.1324        267        640:   8%|▊         | 63/809 [01:15<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.203     0.4404     0.1327        209        640:   8%|▊         | 65/809 [01:17<14:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.211     0.4374     0.1332        386        640:   8%|▊         | 67/809 [01:20<14:55,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.21     0.4372     0.1327        189        640:   9%|▊         | 69/809 [01:22<14:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.207     0.4375     0.1324        277        640:   9%|▊         | 70/809 [01:23<14:27,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.201     0.4384     0.1313        286        640:   9%|▉         | 72/809 [01:26<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.199     0.4388     0.1319        268        640:   9%|▉         | 74/809 [01:28<14:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.199     0.4385     0.1317        303        640:   9%|▉         | 75/809 [01:29<14:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.199      0.438     0.1322        167        640:  10%|▉         | 77/809 [01:32<14:42,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.194     0.4388     0.1314        195        640:  10%|▉         | 79/809 [01:34<14:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.192      0.439      0.131        173        640:  10%|▉         | 80/809 [01:35<14:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.187     0.4406     0.1307        340        640:  10%|█         | 82/809 [01:37<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.188     0.4401      0.131        155        640:  10%|█         | 84/809 [01:40<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.192     0.4384     0.1316        531        640:  11%|█         | 85/809 [01:41<14:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.194     0.4383     0.1327        367        640:  11%|█         | 87/809 [01:44<14:35,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.198      0.437     0.1332        505        640:  11%|█         | 89/809 [01:46<14:30,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.193     0.4388     0.1328        125        640:  11%|█         | 90/809 [01:47<14:23,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.203     0.4364     0.1376        340        640:  11%|█▏        | 92/809 [01:50<14:42,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.208     0.4345     0.1382        172        640:  12%|█▏        | 94/809 [01:52<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.209     0.4343     0.1381        332        640:  12%|█▏        | 95/809 [01:53<14:13,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.211     0.4342     0.1385        205        640:  12%|█▏        | 97/809 [01:56<14:11,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.215      0.433      0.139        430        640:  12%|█▏        | 98/809 [01:57<14:25,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.215     0.4332     0.1397        190        640:  12%|█▏        | 100/809 [01:59<14:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.219     0.4326     0.1419        455        640:  13%|█▎        | 102/809 [02:02<14:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.217     0.4326     0.1416        304        640:  13%|█▎        | 104/809 [02:04<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.215     0.4334     0.1411        181        640:  13%|█▎        | 105/809 [02:05<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.214     0.4338      0.141        224        640:  13%|█▎        | 107/809 [02:08<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.212     0.4344     0.1404        222        640:  13%|█▎        | 109/809 [02:10<13:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.212     0.4345     0.1402        232        640:  14%|█▎        | 110/809 [02:11<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.211      0.435     0.1399        254        640:  14%|█▍        | 112/809 [02:13<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.212     0.4347     0.1398        286        640:  14%|█▍        | 114/809 [02:16<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.213     0.4343     0.1401        303        640:  14%|█▍        | 115/809 [02:17<14:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.212     0.4344       0.14        182        640:  14%|█▍        | 117/809 [02:19<13:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.221     0.4315     0.1435        326        640:  15%|█▍        | 119/809 [02:22<14:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4305      0.144        313        640:  15%|█▍        | 120/809 [02:23<13:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225       0.43     0.1441        353        640:  15%|█▌        | 122/809 [02:25<13:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4291     0.1442        366        640:  15%|█▌        | 124/809 [02:28<13:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4291      0.144        170        640:  15%|█▌        | 125/809 [02:29<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4277     0.1443        300        640:  16%|█▌        | 127/809 [02:32<13:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4268     0.1455        198        640:  16%|█▌        | 129/809 [02:34<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4272     0.1453        314        640:  16%|█▌        | 131/809 [02:36<13:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.427     0.1449        116        640:  16%|█▋        | 132/809 [02:37<13:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4263     0.1452        238        640:  17%|█▋        | 134/809 [02:40<13:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236      0.426      0.145        110        640:  17%|█▋        | 136/809 [02:42<13:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4266     0.1454        130        640:  17%|█▋        | 137/809 [02:43<13:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24      0.425     0.1463        289        640:  17%|█▋        | 139/809 [02:46<13:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24      0.425     0.1474        401        640:  17%|█▋        | 141/809 [02:48<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4256     0.1473         82        640:  18%|█▊        | 142/809 [02:49<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4252     0.1473        259        640:  18%|█▊        | 144/809 [02:52<13:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24      0.425     0.1484        148        640:  18%|█▊        | 146/809 [02:54<13:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241      0.425     0.1487        389        640:  18%|█▊        | 147/809 [02:55<13:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4246     0.1489        176        640:  18%|█▊        | 149/809 [02:58<13:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4245     0.1485        199        640:  19%|█▊        | 151/809 [03:00<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4246     0.1481        282        640:  19%|█▉        | 152/809 [03:01<13:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4247      0.148        314        640:  19%|█▉        | 154/809 [03:04<12:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4253     0.1478        294        640:  19%|█▉        | 156/809 [03:06<13:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4267     0.1474         87        640:  20%|█▉        | 158/809 [03:08<12:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4272     0.1472        201        640:  20%|█▉        | 159/809 [03:10<12:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4276     0.1469        325        640:  20%|█▉        | 161/809 [03:12<12:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4281     0.1469        480        640:  20%|██        | 163/809 [03:14<12:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4286     0.1466        210        640:  20%|██        | 164/809 [03:15<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4292     0.1473        329        640:  21%|██        | 166/809 [03:18<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4293     0.1473        334        640:  21%|██        | 168/809 [03:20<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232      0.429      0.148        329        640:  21%|██        | 169/809 [03:21<12:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4292     0.1487        202        640:  21%|██        | 171/809 [03:24<12:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4295     0.1483        254        640:  21%|██▏       | 173/809 [03:26<12:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4296      0.148        135        640:  22%|██▏       | 174/809 [03:27<12:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4305     0.1475        199        640:  22%|██▏       | 176/809 [03:30<12:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4314     0.1469        131        640:  22%|██▏       | 178/809 [03:32<12:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.222      0.432     0.1466        176        640:  22%|██▏       | 179/809 [03:33<12:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4317     0.1468        258        640:  22%|██▏       | 181/809 [03:36<12:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4309     0.1482        361        640:  23%|██▎       | 183/809 [03:38<12:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4316     0.1479         97        640:  23%|██▎       | 184/809 [03:39<12:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.22     0.4332     0.1473        146        640:  23%|██▎       | 186/809 [03:42<12:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.217     0.4335      0.147        275        640:  23%|██▎       | 188/809 [03:44<12:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.216     0.4337     0.1465        185        640:  23%|██▎       | 190/809 [03:46<12:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.215     0.4339     0.1465        338        640:  24%|██▎       | 192/809 [03:49<12:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.215     0.4338     0.1464        219        640:  24%|██▍       | 193/809 [03:50<11:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.219     0.4328     0.1476        279        640:  24%|██▍       | 195/809 [03:52<12:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.218     0.4336     0.1477        234        640:  24%|██▍       | 197/809 [03:55<12:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.219     0.4333     0.1478         95        640:  24%|██▍       | 198/809 [03:56<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.22     0.4325     0.1488         92        640:  25%|██▍       | 200/809 [03:58<12:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4313     0.1491        281        640:  25%|██▍       | 202/809 [04:01<12:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4307     0.1494        235        640:  25%|██▌       | 203/809 [04:02<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4306     0.1496        157        640:  25%|██▌       | 205/809 [04:04<12:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.222     0.4313     0.1492        216        640:  26%|██▌       | 207/809 [04:07<11:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4304     0.1497        434        640:  26%|██▌       | 208/809 [04:09<11:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4305     0.1495        244        640:  26%|██▌       | 210/809 [04:10<11:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.219     0.4313     0.1489        124        640:  26%|██▌       | 212/809 [04:13<11:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4305     0.1514        641        640:  26%|██▋       | 213/809 [04:14<12:26,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.221     0.4312     0.1512        219        640:  27%|██▋       | 215/809 [04:16<11:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4309     0.1525        754        640:  27%|██▋       | 217/809 [04:19<12:30,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4305     0.1526        212        640:  27%|██▋       | 218/809 [04:20<12:10,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4302     0.1532        338        640:  27%|██▋       | 220/809 [04:23<11:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4299     0.1537        392        640:  27%|██▋       | 222/809 [04:25<11:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4294      0.155        568        640:  28%|██▊       | 223/809 [04:26<12:16,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4298     0.1546        274        640:  28%|██▊       | 225/809 [04:29<11:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4303     0.1542        204        640:  28%|██▊       | 227/809 [04:31<11:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4301     0.1544        399        640:  28%|██▊       | 228/809 [04:32<11:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4304     0.1542        273        640:  28%|██▊       | 230/809 [04:35<11:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4306     0.1553        179        640:  29%|██▊       | 232/809 [04:37<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4311     0.1551        210        640:  29%|██▉       | 233/809 [04:38<11:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.222     0.4319     0.1546        199        640:  29%|██▉       | 235/809 [04:41<11:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.221     0.4324     0.1544        164        640:  29%|██▉       | 237/809 [04:43<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.221     0.4324     0.1543        312        640:  29%|██▉       | 238/809 [04:44<11:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.221     0.4323     0.1545        269        640:  30%|██▉       | 240/809 [04:46<11:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4312     0.1601        530        640:  30%|██▉       | 241/809 [04:48<12:04,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4305     0.1604        207        640:  30%|███       | 243/809 [04:50<11:55,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231       0.43     0.1649        479        640:  30%|███       | 245/809 [04:53<12:03,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4299     0.1649        390        640:  30%|███       | 246/809 [04:54<11:55,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4292     0.1657        200        640:  31%|███       | 248/809 [04:57<12:16,  1.31s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4297     0.1653        147        640:  31%|███       | 249/809 [04:58<11:47,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4299     0.1652        237        640:  31%|███       | 250/809 [04:59<11:25,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4299     0.1648        350        640:  31%|███▏      | 253/809 [05:03<11:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4299     0.1646        203        640:  31%|███▏      | 254/809 [05:04<11:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4296     0.1643        182        640:  32%|███▏      | 256/809 [05:07<11:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4294     0.1649        110        640:  32%|███▏      | 258/809 [05:09<11:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4297     0.1648        287        640:  32%|███▏      | 259/809 [05:10<11:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4295     0.1644        382        640:  32%|███▏      | 261/809 [05:13<11:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4301      0.164        167        640:  33%|███▎      | 263/809 [05:15<10:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4296     0.1638        181        640:  33%|███▎      | 264/809 [05:16<10:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4292      0.164        289        640:  33%|███▎      | 266/809 [05:19<11:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232      0.429     0.1639        358        640:  33%|███▎      | 268/809 [05:21<10:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4288     0.1644        274        640:  33%|███▎      | 269/809 [05:22<10:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4289     0.1647        319        640:  33%|███▎      | 271/809 [05:25<10:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4292     0.1645        220        640:  34%|███▎      | 273/809 [05:27<10:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4292     0.1644        209        640:  34%|███▍      | 274/809 [05:28<10:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4292     0.1641        218        640:  34%|███▍      | 276/809 [05:31<10:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4286      0.164        269        640:  34%|███▍      | 278/809 [05:33<10:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4287     0.1639        257        640:  34%|███▍      | 279/809 [05:34<10:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4294     0.1634        155        640:  35%|███▍      | 281/809 [05:37<10:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4291     0.1633        463        640:  35%|███▍      | 283/809 [05:39<10:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4291     0.1634        260        640:  35%|███▌      | 284/809 [05:40<10:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4289     0.1631        212        640:  35%|███▌      | 286/809 [05:43<10:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4287     0.1626        413        640:  36%|███▌      | 288/809 [05:45<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4288     0.1625        299        640:  36%|███▌      | 289/809 [05:46<10:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4289     0.1629        290        640:  36%|███▌      | 291/809 [05:49<10:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4287      0.163        347        640:  36%|███▌      | 293/809 [05:51<10:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4285     0.1631        316        640:  36%|███▋      | 294/809 [05:52<10:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4289     0.1626        189        640:  37%|███▋      | 296/809 [05:55<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4283      0.163        236        640:  37%|███▋      | 298/809 [05:57<10:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4283     0.1628        221        640:  37%|███▋      | 299/809 [05:58<10:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4292     0.1623        122        640:  37%|███▋      | 301/809 [06:01<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4289     0.1621        255        640:  37%|███▋      | 303/809 [06:03<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4295     0.1618        114        640:  38%|███▊      | 305/809 [06:05<09:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4295     0.1617        413        640:  38%|███▊      | 306/809 [06:06<09:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4295     0.1616        197        640:  38%|███▊      | 308/809 [06:09<10:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4298     0.1614        167        640:  38%|███▊      | 309/809 [06:10<09:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4304     0.1609        172        640:  38%|███▊      | 311/809 [06:12<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4302     0.1616        338        640:  39%|███▊      | 313/809 [06:15<09:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4306     0.1615        299        640:  39%|███▉      | 315/809 [06:17<09:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4306     0.1616        159        640:  39%|███▉      | 316/809 [06:18<09:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4307     0.1613        264        640:  39%|███▉      | 318/809 [06:21<09:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4307     0.1611        309        640:  40%|███▉      | 320/809 [06:23<09:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4309     0.1609        223        640:  40%|███▉      | 321/809 [06:24<09:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23      0.432     0.1605        137        640:  40%|███▉      | 323/809 [06:27<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4325     0.1602        173        640:  40%|████      | 325/809 [06:29<09:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4332     0.1599         67        640:  40%|████      | 327/809 [06:31<09:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4335     0.1598        145        640:  41%|████      | 328/809 [06:32<09:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4334     0.1601        465        640:  41%|████      | 330/809 [06:35<09:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4332     0.1603        325        640:  41%|████      | 332/809 [06:37<09:47,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4331     0.1603        277        640:  41%|████      | 333/809 [06:39<09:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4331     0.1601        286        640:  41%|████▏     | 335/809 [06:41<09:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4325     0.1616        834        640:  42%|████▏     | 336/809 [06:42<10:08,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4325     0.1615        347        640:  42%|████▏     | 338/809 [06:45<09:49,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4326     0.1613        180        640:  42%|████▏     | 340/809 [06:47<09:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4324     0.1614        208        640:  42%|████▏     | 342/809 [06:50<09:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4323     0.1613        445        640:  42%|████▏     | 343/809 [06:51<09:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4326      0.161        139        640:  43%|████▎     | 345/809 [06:53<09:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4326     0.1608        327        640:  43%|████▎     | 347/809 [06:56<09:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4329     0.1606        165        640:  43%|████▎     | 348/809 [06:57<09:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4335     0.1602         78        640:  43%|████▎     | 350/809 [06:59<09:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225      0.434     0.1598        251        640:  44%|████▎     | 352/809 [07:01<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4339     0.1599        324        640:  44%|████▍     | 354/809 [07:04<09:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4339     0.1597        366        640:  44%|████▍     | 355/809 [07:05<09:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227      0.433     0.1596        319        640:  44%|████▍     | 357/809 [07:08<09:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4322     0.1596        474        640:  44%|████▍     | 359/809 [07:10<08:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4319     0.1597        153        640:  44%|████▍     | 360/809 [07:11<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4311     0.1595        345        640:  45%|████▍     | 362/809 [07:13<08:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4304     0.1598        341        640:  45%|████▍     | 364/809 [07:16<09:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4304     0.1598        366        640:  45%|████▌     | 365/809 [07:17<08:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4305     0.1595        158        640:  45%|████▌     | 367/809 [07:20<08:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4308     0.1593        242        640:  46%|████▌     | 369/809 [07:22<08:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4309     0.1591        317        640:  46%|████▌     | 370/809 [07:23<08:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4308      0.159        345        640:  46%|████▌     | 372/809 [07:26<08:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234       0.43     0.1591        447        640:  46%|████▌     | 374/809 [07:28<08:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4298      0.159        129        640:  46%|████▋     | 375/809 [07:29<08:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4297     0.1589        278        640:  47%|████▋     | 377/809 [07:31<08:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4292     0.1589        223        640:  47%|████▋     | 379/809 [07:34<08:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237      0.429     0.1593        366        640:  47%|████▋     | 380/809 [07:35<08:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4298      0.159         88        640:  47%|████▋     | 382/809 [07:37<08:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235       0.43      0.159        242        640:  47%|████▋     | 384/809 [07:40<08:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4306     0.1587        271        640:  48%|████▊     | 386/809 [07:42<08:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4307     0.1587        245        640:  48%|████▊     | 387/809 [07:43<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233      0.431     0.1585        128        640:  48%|████▊     | 389/809 [07:46<08:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4308      0.159        561        640:  48%|████▊     | 390/809 [07:47<08:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4309     0.1593        155        640:  48%|████▊     | 392/809 [07:49<08:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.431     0.1596        157        640:  49%|████▊     | 394/809 [07:52<08:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.431     0.1594        176        640:  49%|████▉     | 396/809 [07:54<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4306     0.1597        426        640:  49%|████▉     | 397/809 [07:55<08:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4303     0.1596        187        640:  49%|████▉     | 399/809 [07:58<08:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4298     0.1595        230        640:  50%|████▉     | 401/809 [08:00<08:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4296     0.1595        252        640:  50%|████▉     | 402/809 [08:01<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4296     0.1593        232        640:  50%|████▉     | 404/809 [08:04<08:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236       0.43     0.1591        156        640:  50%|█████     | 406/809 [08:06<07:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4301     0.1589        197        640:  50%|█████     | 407/809 [08:07<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4306     0.1587        225        640:  51%|█████     | 409/809 [08:09<07:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4306     0.1588        295        640:  51%|█████     | 411/809 [08:12<08:11,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4308     0.1587        130        640:  51%|█████     | 412/809 [08:13<08:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4312     0.1584         75        640:  51%|█████     | 414/809 [08:15<07:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4313     0.1584        652        640:  51%|█████▏    | 416/809 [08:18<08:19,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4313     0.1584        262        640:  52%|█████▏    | 417/809 [08:19<08:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4311     0.1585        397        640:  52%|█████▏    | 419/809 [08:22<08:08,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4313     0.1585        235        640:  52%|█████▏    | 420/809 [08:23<07:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4312     0.1586        322        640:  52%|█████▏    | 422/809 [08:25<07:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4309     0.1585        206        640:  52%|█████▏    | 424/809 [08:28<07:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4308     0.1584        283        640:  53%|█████▎    | 426/809 [08:30<07:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4309     0.1583        168        640:  53%|█████▎    | 427/809 [08:31<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4307     0.1582        159        640:  53%|█████▎    | 429/809 [08:34<07:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4303     0.1582        274        640:  53%|█████▎    | 431/809 [08:36<07:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234       0.43     0.1582        337        640:  53%|█████▎    | 432/809 [08:37<07:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4298     0.1581        244        640:  54%|█████▎    | 434/809 [08:40<07:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4299     0.1579        205        640:  54%|█████▍    | 436/809 [08:42<07:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4302     0.1578        105        640:  54%|█████▍    | 437/809 [08:43<07:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4301     0.1578        262        640:  54%|█████▍    | 439/809 [08:46<07:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4303     0.1576        115        640:  55%|█████▍    | 441/809 [08:48<07:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4306     0.1575        189        640:  55%|█████▍    | 442/809 [08:49<07:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4305     0.1575        360        640:  55%|█████▍    | 444/809 [08:52<07:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4305     0.1574        182        640:  55%|█████▌    | 446/809 [08:54<07:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4304     0.1573        208        640:  55%|█████▌    | 447/809 [08:55<07:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4304     0.1572        352        640:  56%|█████▌    | 449/809 [08:58<07:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4304      0.157        317        640:  56%|█████▌    | 450/809 [09:00<07:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4302      0.157        408        640:  56%|█████▌    | 452/809 [09:01<07:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233       0.43     0.1569        318        640:  56%|█████▌    | 454/809 [09:04<07:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233       0.43     0.1568        336        640:  56%|█████▋    | 456/809 [09:06<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4301     0.1566        107        640:  56%|█████▋    | 457/809 [09:07<06:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4299     0.1564        182        640:  57%|█████▋    | 459/809 [09:10<06:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4301     0.1562        255        640:  57%|█████▋    | 461/809 [09:12<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4301     0.1561        257        640:  57%|█████▋    | 462/809 [09:13<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233       0.43     0.1563        413        640:  57%|█████▋    | 464/809 [09:16<06:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4297     0.1563        327        640:  58%|█████▊    | 466/809 [09:18<06:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4297     0.1562        306        640:  58%|█████▊    | 467/809 [09:19<06:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4298     0.1561        171        640:  58%|█████▊    | 469/809 [09:22<06:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4299     0.1559        235        640:  58%|█████▊    | 471/809 [09:24<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4296     0.1559        311        640:  58%|█████▊    | 472/809 [09:25<06:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4294     0.1559        541        640:  59%|█████▊    | 474/809 [09:28<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4292     0.1563        730        640:  59%|█████▉    | 476/809 [09:30<07:00,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4294     0.1562        126        640:  59%|█████▉    | 477/809 [09:31<06:50,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4296      0.156        248        640:  59%|█████▉    | 479/809 [09:34<06:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4296     0.1559        290        640:  59%|█████▉    | 481/809 [09:36<06:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4297     0.1558        199        640:  60%|█████▉    | 482/809 [09:37<06:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4298     0.1557        323        640:  60%|█████▉    | 484/809 [09:40<06:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4297     0.1557        359        640:  60%|██████    | 486/809 [09:42<06:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4295     0.1556        239        640:  60%|██████    | 487/809 [09:43<06:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4297     0.1554        247        640:  60%|██████    | 489/809 [09:46<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4295     0.1553        101        640:  61%|██████    | 491/809 [09:48<06:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4292     0.1555        604        640:  61%|██████    | 492/809 [09:50<06:31,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4297     0.1554        296        640:  61%|██████    | 494/809 [09:52<06:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4298     0.1552        196        640:  61%|██████▏   | 496/809 [09:54<06:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4296     0.1553        479        640:  61%|██████▏   | 497/809 [09:55<06:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4298     0.1551        229        640:  62%|██████▏   | 499/809 [09:58<06:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4295     0.1553        233        640:  62%|██████▏   | 501/809 [10:00<06:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4294     0.1552        256        640:  62%|██████▏   | 502/809 [10:01<06:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4295     0.1552        267        640:  62%|██████▏   | 504/809 [10:04<06:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4293     0.1552        470        640:  63%|██████▎   | 506/809 [10:06<05:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234     0.4291     0.1552        273        640:  63%|██████▎   | 507/809 [10:07<06:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.429     0.1552        167        640:  63%|██████▎   | 510/809 [10:11<05:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4287     0.1552        175        640:  63%|██████▎   | 511/809 [10:12<06:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4283     0.1552        409        640:  63%|██████▎   | 513/809 [10:15<05:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4279     0.1561        368        640:  64%|██████▎   | 514/809 [10:16<06:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4277     0.1561        200        640:  64%|██████▍   | 516/809 [10:18<05:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4278      0.156        227        640:  64%|██████▍   | 518/809 [10:21<05:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4281     0.1559        126        640:  64%|██████▍   | 519/809 [10:22<05:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4279     0.1563        542        640:  64%|██████▍   | 521/809 [10:24<05:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4277     0.1562        684        640:  65%|██████▍   | 523/809 [10:27<05:52,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4275     0.1562        254        640:  65%|██████▍   | 524/809 [10:28<05:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4275      0.156         86        640:  65%|██████▌   | 526/809 [10:30<05:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4271     0.1561        364        640:  65%|██████▌   | 528/809 [10:33<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4267      0.156        327        640:  66%|██████▌   | 530/809 [10:35<05:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4264      0.156        247        640:  66%|██████▌   | 531/809 [10:36<05:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4261     0.1562        395        640:  66%|██████▌   | 533/809 [10:39<05:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243      0.426     0.1561        212        640:  66%|██████▌   | 535/809 [10:41<05:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4259     0.1561        260        640:  66%|██████▋   | 536/809 [10:42<05:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.245     0.4254     0.1565        290        640:  67%|██████▋   | 538/809 [10:45<05:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.244     0.4258     0.1564        332        640:  67%|██████▋   | 540/809 [10:47<05:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.244     0.4257     0.1563        350        640:  67%|██████▋   | 541/809 [10:48<05:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.244     0.4259     0.1564        444        640:  67%|██████▋   | 543/809 [10:51<05:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243      0.426     0.1564        266        640:  67%|██████▋   | 545/809 [10:53<05:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4262     0.1563        325        640:  67%|██████▋   | 546/809 [10:54<05:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4262     0.1561        163        640:  68%|██████▊   | 548/809 [10:57<05:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262      0.156        236        640:  68%|██████▊   | 550/809 [10:59<05:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262      0.156        116        640:  68%|██████▊   | 551/809 [11:00<05:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4265     0.1557        133        640:  68%|██████▊   | 553/809 [11:03<05:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4264     0.1558        357        640:  69%|██████▊   | 555/809 [11:05<05:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4265     0.1557        249        640:  69%|██████▊   | 556/809 [11:06<05:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4268     0.1558        309        640:  69%|██████▉   | 558/809 [11:09<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24      0.427     0.1558        288        640:  69%|██████▉   | 560/809 [11:11<04:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4269     0.1558        285        640:  69%|██████▉   | 561/809 [11:12<04:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4271     0.1556        267        640:  70%|██████▉   | 563/809 [11:14<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4273     0.1554        180        640:  70%|██████▉   | 565/809 [11:17<04:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4276     0.1553        354        640:  70%|███████   | 567/809 [11:19<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4277     0.1552        140        640:  70%|███████   | 568/809 [11:20<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4276     0.1556        447        640:  70%|███████   | 570/809 [11:23<04:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238      0.428     0.1556        137        640:  71%|███████   | 572/809 [11:25<04:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4278     0.1556        355        640:  71%|███████   | 573/809 [11:26<04:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4279     0.1554        196        640:  71%|███████   | 575/809 [11:29<04:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4278     0.1556        341        640:  71%|███████▏  | 577/809 [11:31<04:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4279     0.1555        379        640:  71%|███████▏  | 578/809 [11:32<04:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4279     0.1554        215        640:  72%|███████▏  | 580/809 [11:35<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4281     0.1552        196        640:  72%|███████▏  | 582/809 [11:37<04:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237      0.428     0.1552        400        640:  72%|███████▏  | 583/809 [11:38<04:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4282      0.155        233        640:  72%|███████▏  | 585/809 [11:41<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4282     0.1549        278        640:  73%|███████▎  | 587/809 [11:43<04:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4279     0.1551        203        640:  73%|███████▎  | 588/809 [11:44<04:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4275     0.1555        260        640:  73%|███████▎  | 590/809 [11:47<04:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4276     0.1554        186        640:  73%|███████▎  | 592/809 [11:49<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4277     0.1554        220        640:  73%|███████▎  | 593/809 [11:50<04:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4273     0.1557        360        640:  74%|███████▎  | 595/809 [11:53<04:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4272     0.1557        260        640:  74%|███████▍  | 597/809 [11:55<04:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4271     0.1556        201        640:  74%|███████▍  | 598/809 [11:56<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4271     0.1554        238        640:  74%|███████▍  | 600/809 [11:59<04:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4269     0.1555        258        640:  74%|███████▍  | 602/809 [12:01<04:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4269     0.1554        228        640:  75%|███████▍  | 603/809 [12:02<04:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4267     0.1554        315        640:  75%|███████▍  | 605/809 [12:05<04:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4265     0.1555        310        640:  75%|███████▌  | 607/809 [12:07<04:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4262     0.1563        606        640:  75%|███████▌  | 608/809 [12:08<04:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243      0.426     0.1562        225        640:  75%|███████▌  | 610/809 [12:11<04:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4261     0.1562        422        640:  76%|███████▌  | 612/809 [12:13<03:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243      0.426     0.1561        277        640:  76%|███████▌  | 613/809 [12:14<03:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.244     0.4259     0.1563        178        640:  76%|███████▌  | 615/809 [12:17<03:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4261     0.1562        177        640:  76%|███████▋  | 617/809 [12:19<03:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4263      0.156        167        640:  77%|███████▋  | 619/809 [12:21<03:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4261     0.1565        568        640:  77%|███████▋  | 620/809 [12:23<03:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4263     0.1564        148        640:  77%|███████▋  | 622/809 [12:25<03:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4264     0.1563        438        640:  77%|███████▋  | 624/809 [12:27<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4264     0.1563        399        640:  77%|███████▋  | 625/809 [12:28<03:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4263     0.1564        377        640:  78%|███████▊  | 627/809 [12:31<03:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4263     0.1563        251        640:  78%|███████▊  | 628/809 [12:32<03:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262     0.1562        275        640:  78%|███████▊  | 630/809 [12:35<03:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262     0.1563        367        640:  78%|███████▊  | 632/809 [12:37<03:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262     0.1563        369        640:  78%|███████▊  | 634/809 [12:39<03:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4262     0.1562        296        640:  78%|███████▊  | 635/809 [12:41<03:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.243     0.4261     0.1566        225        640:  79%|███████▊  | 637/809 [12:43<03:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4264     0.1565        224        640:  79%|███████▉  | 639/809 [12:45<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4263     0.1565        231        640:  79%|███████▉  | 640/809 [12:46<03:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.242     0.4265     0.1564        307        640:  79%|███████▉  | 642/809 [12:49<03:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4265     0.1563        150        640:  80%|███████▉  | 644/809 [12:51<03:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4265     0.1562        374        640:  80%|███████▉  | 645/809 [12:52<03:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4266     0.1562        301        640:  80%|███████▉  | 647/809 [12:55<03:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4268     0.1561        316        640:  80%|████████  | 649/809 [12:57<03:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4267     0.1561        192        640:  80%|████████  | 650/809 [12:58<03:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4267     0.1561        239        640:  81%|████████  | 652/809 [13:01<03:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4268     0.1561        216        640:  81%|████████  | 654/809 [13:03<03:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.241     0.4269      0.156        261        640:  81%|████████  | 655/809 [13:05<03:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4272     0.1558        227        640:  81%|████████  | 657/809 [13:07<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4274     0.1557        180        640:  81%|████████▏ | 659/809 [13:09<02:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.24     0.4274     0.1558        381        640:  82%|████████▏ | 660/809 [13:11<02:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4278     0.1556         75        640:  82%|████████▏ | 662/809 [13:13<02:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4277     0.1556        480        640:  82%|████████▏ | 664/809 [13:15<02:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.239     0.4277     0.1556        217        640:  82%|████████▏ | 665/809 [13:16<02:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4278     0.1555        226        640:  82%|████████▏ | 667/809 [13:19<02:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4279     0.1553        122        640:  83%|████████▎ | 669/809 [13:21<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4281     0.1552        323        640:  83%|████████▎ | 671/809 [13:23<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238      0.428     0.1553        425        640:  83%|████████▎ | 672/809 [13:25<02:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.238     0.4277     0.1553        381        640:  83%|████████▎ | 674/809 [13:27<02:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237      0.428     0.1552        287        640:  84%|████████▎ | 676/809 [13:29<02:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.237     0.4281     0.1551        196        640:  84%|████████▎ | 677/809 [13:31<02:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4284      0.155        336        640:  84%|████████▍ | 679/809 [13:33<02:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4285      0.155        159        640:  84%|████████▍ | 681/809 [13:35<02:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.236     0.4285      0.155        314        640:  84%|████████▍ | 682/809 [13:37<02:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4287     0.1547        213        640:  85%|████████▍ | 684/809 [13:39<02:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4288     0.1548        165        640:  85%|████████�� | 686/809 [13:41<02:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.235     0.4288     0.1548        450        640:  85%|████████▍ | 687/809 [13:43<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.429     0.1547         95        640:  85%|████████▌ | 689/809 [13:45<02:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.234      0.429     0.1547        168        640:  85%|████████▌ | 691/809 [13:47<02:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.233     0.4293     0.1545        105        640:  86%|████████▌ | 693/809 [13:50<02:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4296     0.1545        121        640:  86%|████████▌ | 694/809 [13:51<02:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4298     0.1543        123        640:  86%|████████▌ | 696/809 [13:53<02:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232     0.4299     0.1542        359        640:  86%|████████▋ | 698/809 [13:55<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.232       0.43     0.1542        100        640:  87%|████████▋ | 700/809 [13:58<02:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4301      0.154        100        640:  87%|████████▋ | 701/809 [13:59<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231       0.43      0.154        384        640:  87%|████████▋ | 703/809 [14:01<02:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4302     0.1539        190        640:  87%|████████▋ | 705/809 [14:04<02:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231       0.43      0.154        350        640:  87%|████████▋ | 707/809 [14:06<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4302     0.1539        102        640:  88%|████████▊ | 708/809 [14:07<02:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4302     0.1538        194        640:  88%|████████▊ | 710/809 [14:10<01:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.231     0.4301     0.1536        219        640:  88%|████████▊ | 712/809 [14:12<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1535        157        640:  88%|████████▊ | 714/809 [14:14<01:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1534        239        640:  88%|████████▊ | 715/809 [14:15<01:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1534        292        640:  89%|████████▊ | 717/809 [14:18<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1533        252        640:  89%|████████▉ | 719/809 [14:20<01:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1534        275        640:  89%|████████▉ | 720/809 [14:21<01:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1533        453        640:  89%|████████▉ | 722/809 [14:24<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4303     0.1532        155        640:  89%|████████▉ | 724/809 [14:26<01:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4305     0.1531         77        640:  90%|████████▉ | 725/809 [14:27<01:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4304     0.1532        383        640:  90%|████████▉ | 727/809 [14:30<01:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4308     0.1531        206        640:  90%|█████████ | 729/809 [14:32<01:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228      0.431      0.153        139        640:  90%|█████████ | 730/809 [14:33<01:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4309     0.1531        132        640:  90%|█████████ | 732/809 [14:36<01:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4307     0.1538        287        640:  91%|█████████ | 734/809 [14:38<01:32,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4304     0.1539        561        640:  91%|█████████ | 735/809 [14:39<01:33,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4306     0.1538        177        640:  91%|█████████ | 737/809 [14:42<01:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4307     0.1537        156        640:  91%|█████████▏| 739/809 [14:44<01:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4307     0.1537        401        640:  91%|█████████▏| 740/809 [14:45<01:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4305      0.154        465        640:  92%|█████████▏| 742/809 [14:48<01:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4305      0.154        214        640:  92%|█████████▏| 744/809 [14:50<01:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4306      0.154        233        640:  92%|█████████▏| 745/809 [14:51<01:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4307      0.154         76        640:  92%|█████████▏| 747/809 [14:54<01:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4306     0.1539        281        640:  93%|█████████▎| 749/809 [14:56<01:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4306      0.154        309        640:  93%|█████████▎| 750/809 [14:57<01:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4303     0.1542        320        640:  93%|█████████▎| 752/809 [15:00<01:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4303     0.1541        162        640:  93%|█████████▎| 754/809 [15:02<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4302     0.1541        174        640:  93%|█████████▎| 756/809 [15:04<01:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4301     0.1541        372        640:  94%|█████████▎| 757/809 [15:06<01:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4299     0.1543        438        640:  94%|█████████▍| 759/809 [15:08<01:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4301     0.1543        195        640:  94%|█████████▍| 760/809 [15:09<00:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23       0.43     0.1542        131        640:  94%|█████████▍| 762/809 [15:12<00:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23       0.43     0.1543        466        640:  94%|█████████▍| 764/809 [15:14<00:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229       0.43     0.1543        141        640:  95%|█████████▍| 765/809 [15:15<00:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4299     0.1543        250        640:  95%|█████████▍| 767/809 [15:18<00:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4296     0.1545        484        640:  95%|█████████▌| 769/809 [15:20<00:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G       1.23     0.4298     0.1545         99        640:  95%|█████████▌| 770/809 [15:21<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4299     0.1545        102        640:  95%|█████████▌| 772/809 [15:24<00:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229       0.43     0.1544        280        640:  96%|█████████▌| 774/809 [15:26<00:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.229     0.4301     0.1543        219        640:  96%|█████████▌| 776/809 [15:28<00:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4302     0.1543        202        640:  96%|█████████▌| 777/809 [15:30<00:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4303     0.1542         99        640:  96%|█████████▋| 779/809 [15:32<00:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.228     0.4304     0.1541        208        640:  97%|█████████▋| 781/809 [15:34<00:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.227     0.4306      0.154        331        640:  97%|█████████▋| 782/809 [15:36<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4309     0.1539         81        640:  97%|█████████▋| 784/809 [15:38<00:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.226     0.4311     0.1538        191        640:  97%|█████████▋| 786/809 [15:40<00:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.225     0.4313     0.1538         95        640:  97%|█████████▋| 787/809 [15:42<00:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4316     0.1536        217        640:  98%|█████████▊| 789/809 [15:44<00:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4318     0.1536        257        640:  98%|█████████▊| 791/809 [15:46<00:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4318     0.1535        192        640:  98%|█████████▊| 792/809 [15:47<00:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4317     0.1535        212        640:  98%|█████████▊| 794/809 [15:50<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4318     0.1534        254        640:  98%|█████████▊| 796/809 [15:52<00:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4317     0.1534        208        640:  99%|█████████▊| 797/809 [15:53<00:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.224     0.4316     0.1533        262        640:  99%|█████████▉| 799/809 [15:56<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4319     0.1532        200        640:  99%|█████████▉| 801/809 [15:58<00:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4318     0.1532        256        640:  99%|█████████▉| 803/809 [16:00<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4317     0.1531        236        640:  99%|█████████▉| 804/809 [16:02<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.223     0.4319      0.153        139        640: 100%|█████████▉| 806/809 [16:04<00:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     31/300      13.2G      1.222     0.4319     0.1529        122        640: 100%|█████████▉| 808/809 [16:06<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:01<00:15,  4.07it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:02<00:14,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:05<00:12,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:11,  3.72it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  3.84it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:10<00:06,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:12<00:05,  3.70it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  84%|████████▍ | 58/69 [00:14<00:02,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4633     0.1516        389        640:   0%|          | 2/809 [00:02<16:36,  1.24s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.092     0.5007     0.1235        121        640:   0%|          | 4/809 [00:04<16:02,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.096     0.4971     0.1245        331        640:   1%|          | 5/809 [00:05<15:52,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.118     0.4814     0.1199        160        640:   1%|          | 7/809 [00:08<15:40,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.098     0.4767     0.1207        322        640:   1%|          | 9/809 [00:10<15:37,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.072     0.4857     0.1167        141        640:   1%|          | 10/809 [00:11<15:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.092     0.4837     0.1216        265        640:   1%|▏         | 12/809 [00:14<15:41,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.098     0.4852     0.1198        259        640:   2%|▏         | 14/809 [00:16<15:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.107     0.4869     0.1213        250        640:   2%|▏         | 16/809 [00:18<15:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.109     0.4865     0.1228        203        640:   2%|▏         | 17/809 [00:20<15:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.115     0.4872     0.1282        278        640:   2%|▏         | 19/809 [00:22<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.114     0.4851     0.1273        309        640:   3%|▎         | 21/809 [00:24<15:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.104     0.4849     0.1254         98        640:   3%|▎         | 22/809 [00:26<16:06,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.107     0.4832     0.1263        246        640:   3%|▎         | 24/809 [00:28<15:52,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.134     0.4766     0.1457        503        640:   3%|▎         | 26/809 [00:31<16:21,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.14     0.4737     0.1452        154        640:   3%|▎         | 27/809 [00:32<15:54,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.181     0.4577     0.1499        436        640:   4%|▎         | 29/809 [00:35<16:44,  1.29s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.193     0.4526     0.1503        439        640:   4%|▍         | 31/809 [00:37<16:06,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.191     0.4533     0.1487        112        640:   4%|▍         | 32/809 [00:38<15:51,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4448     0.1549        321        640:   4%|▍         | 34/809 [00:41<15:43,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.225     0.4431     0.1582        195        640:   4%|▍         | 36/809 [00:43<15:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4456     0.1577        173        640:   5%|▍         | 37/809 [00:44<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4434     0.1571        444        640:   5%|▍         | 39/809 [00:46<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4433     0.1565        158        640:   5%|▌         | 41/809 [00:49<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4439     0.1553        419        640:   5%|▌         | 42/809 [00:50<15:33,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.225     0.4387     0.1552        250        640:   5%|▌         | 44/809 [00:52<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.235     0.4343     0.1544        106        640:   6%|▌         | 46/809 [00:55<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.253     0.4293     0.1587        416        640:   6%|▌         | 47/809 [00:56<15:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.264     0.4251     0.1578        268        640:   6%|▌         | 49/809 [00:58<15:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.269     0.4229     0.1568        275        640:   6%|▋         | 51/809 [01:01<15:19,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.266     0.4231     0.1556        264        640:   6%|▋         | 52/809 [01:02<15:28,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.247     0.4296     0.1532        159        640:   7%|▋         | 54/809 [01:04<14:59,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.242     0.4304     0.1518        248        640:   7%|▋         | 56/809 [01:07<14:50,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.246     0.4289     0.1527        308        640:   7%|▋         | 57/809 [01:08<14:57,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.245      0.429     0.1526        202        640:   7%|▋         | 59/809 [01:10<14:58,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.238     0.4322     0.1514        153        640:   8%|▊         | 61/809 [01:13<15:08,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.241     0.4313     0.1511        300        640:   8%|▊         | 62/809 [01:14<14:53,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.249     0.4288     0.1527        273        640:   8%|▊         | 64/809 [01:16<14:53,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.255     0.4268     0.1525        122        640:   8%|▊         | 66/809 [01:19<14:55,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.248     0.4297     0.1517        112        640:   8%|▊         | 67/809 [01:20<14:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.243     0.4301     0.1516         74        640:   9%|▊         | 69/809 [01:22<14:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.236     0.4308     0.1506        147        640:   9%|▉         | 71/809 [01:25<14:29,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.233     0.4321       0.15        201        640:   9%|▉         | 73/809 [01:27<14:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.236     0.4311     0.1506        365        640:   9%|▉         | 74/809 [01:28<14:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.23     0.4338     0.1495        108        640:   9%|▉         | 76/809 [01:31<14:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.241     0.4306     0.1526        381        640:  10%|▉         | 78/809 [01:33<14:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.243     0.4297     0.1525        257        640:  10%|▉         | 79/809 [01:34<14:32,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.244      0.429     0.1526        376        640:  10%|█         | 81/809 [01:37<14:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.244     0.4284     0.1518        332        640:  10%|█         | 83/809 [01:39<14:32,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.243      0.428     0.1519        285        640:  10%|█         | 84/809 [01:40<14:32,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.245     0.4273     0.1527        187        640:  11%|█         | 86/809 [01:43<14:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.239     0.4282     0.1518        127        640:  11%|█         | 88/809 [01:45<14:32,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.24      0.428     0.1515        321        640:  11%|█         | 89/809 [01:46<14:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.234     0.4295     0.1503        238        640:  11%|█         | 91/809 [01:49<14:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.232       0.43     0.1499        482        640:  11%|█▏        | 93/809 [01:51<14:13,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.227     0.4317     0.1494        142        640:  12%|█▏        | 94/809 [01:52<14:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.221     0.4333     0.1482        228        640:  12%|█▏        | 96/809 [01:55<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4347     0.1475        129        640:  12%|█▏        | 98/809 [01:57<14:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217      0.435     0.1472        205        640:  12%|█▏        | 99/809 [01:58<13:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348     0.1475        311        640:  12%|█▏        | 101/809 [02:00<14:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4366     0.1471        294        640:  13%|█▎        | 103/809 [02:03<14:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.203      0.438     0.1462         98        640:  13%|█▎        | 105/809 [02:05<13:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.205     0.4378     0.1465        427        640:  13%|█▎        | 106/809 [02:06<14:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.202     0.4393     0.1459        275        640:  13%|█▎        | 108/809 [02:09<13:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.197     0.4405     0.1452        178        640:  14%|█▎        | 110/809 [02:11<13:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.193     0.4418     0.1446        192        640:  14%|█▎        | 111/809 [02:12<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.188     0.4426     0.1437        169        640:  14%|█▍        | 113/809 [02:15<13:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.187     0.4429     0.1434        195        640:  14%|█▍        | 115/809 [02:17<13:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.185     0.4432      0.143        318        640:  14%|█▍        | 116/809 [02:18<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.187     0.4419     0.1429        362        640:  15%|█▍        | 118/809 [02:21<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.181     0.4429     0.1423         83        640:  15%|█▍        | 120/809 [02:23<13:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186     0.4416     0.1473        499        640:  15%|█▍        | 121/809 [02:24<14:32,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.187     0.4409     0.1471        462        640:  15%|█▌        | 123/809 [02:27<14:17,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.188     0.4409     0.1476        374        640:  15%|█▌        | 125/809 [02:29<13:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4401      0.148        577        640:  16%|█▌        | 126/809 [02:30<13:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.192     0.4392     0.1479        202        640:  16%|█▌        | 128/809 [02:33<13:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.195     0.4381     0.1486        469        640:  16%|█▌        | 130/809 [02:35<13:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.194     0.4383     0.1482        177        640:  16%|█▌        | 131/809 [02:36<13:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.192     0.4389     0.1478        215        640:  16%|█▋        | 133/809 [02:39<13:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.194      0.438     0.1475        342        640:  17%|█▋        | 135/809 [02:41<13:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.193     0.4381     0.1476        323        640:  17%|█▋        | 136/809 [02:42<13:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4397     0.1472        208        640:  17%|█▋        | 138/809 [02:45<13:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.189     0.4404     0.1467        341        640:  17%|█▋        | 140/809 [02:47<13:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.188     0.4406     0.1465        258        640:  17%|█▋        | 141/809 [02:48<13:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.191     0.4399     0.1467        283        640:  18%|█▊        | 143/809 [02:51<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.194     0.4391     0.1469        229        640:  18%|█▊        | 145/809 [02:53<13:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4402     0.1463        143        640:  18%|█▊        | 147/809 [02:55<13:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4402     0.1462        303        640:  18%|█▊        | 148/809 [02:57<13:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.195      0.439     0.1474        773        640:  19%|█▊        | 150/809 [02:59<13:36,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.198     0.4377     0.1475        278        640:  19%|█▊        | 151/809 [03:00<13:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.198     0.4377     0.1474        297        640:  19%|█▉        | 153/809 [03:03<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.201     0.4371     0.1485        351        640:  19%|█▉        | 155/809 [03:05<13:37,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2     0.4372     0.1489        325        640:  19%|█▉        | 156/809 [03:06<13:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.199     0.4373     0.1486        252        640:  20%|█▉        | 158/809 [03:09<13:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.202     0.4361     0.1485        293        640:  20%|█▉        | 160/809 [03:11<12:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2     0.4368     0.1483        269        640:  20%|█▉        | 161/809 [03:12<12:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.201     0.4369     0.1487        169        640:  20%|██        | 163/809 [03:15<13:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.202     0.4367     0.1485        296        640:  20%|██        | 165/809 [03:17<12:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.201     0.4363     0.1481        171        640:  21%|██        | 166/809 [03:19<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.205     0.4353     0.1507        256        640:  21%|██        | 168/809 [03:21<12:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.204     0.4358     0.1504        378        640:  21%|██        | 170/809 [03:23<12:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.202     0.4367     0.1502         97        640:  21%|██        | 171/809 [03:25<12:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2     0.4376     0.1494        292        640:  21%|██▏       | 173/809 [03:27<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2      0.438     0.1491        163        640:  22%|██▏       | 175/809 [03:29<12:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2     0.4381     0.1489        309        640:  22%|██▏       | 176/809 [03:31<12:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.199     0.4383     0.1484        171        640:  22%|██▏       | 178/809 [03:33<12:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.199      0.438      0.148        395        640:  22%|██▏       | 180/809 [03:35<12:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.196      0.439     0.1474        199        640:  22%|██▏       | 182/809 [03:38<12:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.198     0.4387     0.1473        207        640:  23%|██▎       | 184/809 [03:40<12:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.196      0.439     0.1469        294        640:  23%|██▎       | 185/809 [03:41<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.193     0.4399     0.1463        258        640:  23%|██▎       | 187/809 [03:44<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.195     0.4397     0.1464        321        640:  23%|██▎       | 189/809 [03:46<12:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.194       0.44     0.1461        255        640:  23%|██▎       | 190/809 [03:47<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.193     0.4405     0.1458        233        640:  24%|██▎       | 192/809 [03:49<12:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4416     0.1453        159        640:  24%|██▍       | 194/809 [03:52<12:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4416      0.145        232        640:  24%|██▍       | 195/809 [03:53<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.19     0.4417      0.145        289        640:  24%|██▍       | 197/809 [03:55<11:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.191     0.4416     0.1457        211        640:  25%|██▍       | 199/809 [03:58<12:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.191     0.4413     0.1455        358        640:  25%|██▍       | 200/809 [03:59<12:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.188     0.4424     0.1451        130        640:  25%|██▍       | 202/809 [04:01<12:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186      0.443      0.145        159        640:  25%|██▌       | 204/809 [04:04<11:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.185     0.4434     0.1445        149        640:  25%|██▌       | 206/809 [04:06<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.187      0.443     0.1448        498        640:  26%|██▌       | 207/809 [04:07<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186     0.4438     0.1449        210        640:  26%|██▌       | 209/809 [04:10<11:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.189     0.4433     0.1453        366        640:  26%|██▌       | 211/809 [04:12<12:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.189     0.4433     0.1454        313        640:  26%|██▌       | 212/809 [04:13<12:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.188     0.4433     0.1451        223        640:  26%|██▋       | 214/809 [04:16<11:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186     0.4435     0.1448        176        640:  27%|██▋       | 216/809 [04:18<11:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186     0.4433     0.1448        164        640:  27%|██▋       | 217/809 [04:19<11:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.185     0.4437     0.1444        183        640:  27%|██▋       | 219/809 [04:22<11:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.185     0.4439     0.1444        388        640:  27%|██▋       | 221/809 [04:24<11:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.184     0.4442     0.1441         73        640:  27%|██▋       | 222/809 [04:25<11:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.185     0.4442     0.1448        113        640:  28%|██▊       | 224/809 [04:27<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.186     0.4439      0.145        369        640:  28%|██▊       | 226/809 [04:30<11:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.189     0.4431     0.1466        361        640:  28%|██▊       | 227/809 [04:31<12:05,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.192     0.4421     0.1466        275        640:  28%|██▊       | 229/809 [04:34<11:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.195     0.4411     0.1473        425        640:  29%|██▊       | 231/809 [04:36<11:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.196     0.4407     0.1473        310        640:  29%|██▊       | 232/809 [04:37<11:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.194     0.4414     0.1469        145        640:  29%|██▉       | 234/809 [04:40<11:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.196     0.4411     0.1474        337        640:  29%|██▉       | 236/809 [04:42<11:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.196      0.441     0.1472        255        640:  29%|██▉       | 237/809 [04:43<11:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.198     0.4401     0.1472        188        640:  30%|██▉       | 239/809 [04:45<11:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G        1.2     0.4398     0.1483        127        640:  30%|██▉       | 241/809 [04:48<11:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.201     0.4398     0.1486        262        640:  30%|██▉       | 242/809 [04:49<11:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.204     0.4391     0.1491        635        640:  30%|███       | 244/809 [04:52<11:40,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.204     0.4391     0.1493        227        640:  30%|███       | 245/809 [04:53<11:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4382     0.1504        434        640:  31%|███       | 247/809 [04:55<11:34,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4381     0.1516        351        640:  31%|███       | 249/809 [04:58<11:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4382     0.1512         82        640:  31%|███       | 251/809 [05:00<11:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4378     0.1512        245        640:  31%|███       | 252/809 [05:01<11:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4368     0.1512        330        640:  31%|███▏      | 254/809 [05:04<11:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1511        342        640:  32%|███▏      | 256/809 [05:06<11:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4363     0.1509         97        640:  32%|███▏      | 257/809 [05:07<11:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4362     0.1507        210        640:  32%|███▏      | 259/809 [05:10<10:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4358     0.1505        314        640:  32%|███▏      | 261/809 [05:12<10:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1505        388        640:  32%|███▏      | 262/809 [05:13<10:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4359     0.1503        263        640:  33%|███▎      | 264/809 [05:16<10:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4353     0.1502        290        640:  33%|███▎      | 266/809 [05:18<11:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4353     0.1502        265        640:  33%|███▎      | 267/809 [05:19<10:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4354     0.1501        174        640:  33%|███▎      | 269/809 [05:22<10:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4357     0.1498        201        640:  33%|███▎      | 271/809 [05:24<10:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.436     0.1496        157        640:  34%|███▎      | 272/809 [05:25<10:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4351       0.15        399        640:  34%|███▍      | 274/809 [05:28<10:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4359     0.1496        275        640:  34%|███▍      | 276/809 [05:30<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4352     0.1505        374        640:  34%|███▍      | 277/809 [05:31<10:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4351     0.1505        266        640:  34%|███▍      | 279/809 [05:34<10:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.221     0.4345     0.1504        328        640:  35%|███▍      | 281/809 [05:36<10:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.221     0.4345     0.1504        286        640:  35%|███▍      | 282/809 [05:37<10:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.221     0.4346     0.1502        206        640:  35%|███▌      | 284/809 [05:40<10:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4356     0.1498        134        640:  35%|███▌      | 286/809 [05:42<10:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4356     0.1508        140        640:  36%|███▌      | 288/809 [05:44<10:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4354     0.1508        185        640:  36%|███▌      | 289/809 [05:45<10:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4356     0.1505        169        640:  36%|███▌      | 291/809 [05:48<10:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4355     0.1506        292        640:  36%|███▌      | 293/809 [05:50<10:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.221     0.4351     0.1506        187        640:  36%|███▋      | 294/809 [05:51<10:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4352     0.1505        201        640:  37%|███▋      | 296/809 [05:54<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.222     0.4346     0.1508        364        640:  37%|███▋      | 298/809 [05:56<10:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.222     0.4348     0.1507        215        640:  37%|███▋      | 299/809 [05:57<10:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4355     0.1503        120        640:  37%|███▋      | 301/809 [06:00<09:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4363     0.1499        203        640:  37%|███▋      | 303/809 [06:02<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4364     0.1498        202        640:  38%|███▊      | 305/809 [06:04<09:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.437     0.1496        120        640:  38%|███▊      | 306/809 [06:06<09:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4365       0.15        205        640:  38%|███▊      | 308/809 [06:08<09:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.437     0.1498        339        640:  38%|███▊      | 310/809 [06:10<09:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4371     0.1497        217        640:  38%|███▊      | 311/809 [06:12<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4372     0.1494        113        640:  39%|███▊      | 313/809 [06:14<09:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4372     0.1491        327        640:  39%|███▉      | 315/809 [06:16<09:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4374     0.1489        220        640:  39%|███▉      | 316/809 [06:17<09:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4382     0.1486        189        640:  39%|███▉      | 318/809 [06:20<09:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4386     0.1487        232        640:  40%|███▉      | 320/809 [06:22<09:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4387     0.1487        224        640:  40%|███▉      | 321/809 [06:23<09:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4383      0.149        485        640:  40%|███▉      | 323/809 [06:26<09:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4381     0.1489        403        640:  40%|████      | 325/809 [06:28<09:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4381     0.1488        135        640:  40%|████      | 327/809 [06:30<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4378      0.149        362        640:  41%|████      | 328/809 [06:32<09:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4377      0.149        211        640:  41%|████      | 330/809 [06:34<09:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4374     0.1489        109        640:  41%|████      | 332/809 [06:36<09:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4374     0.1489        171        640:  41%|████      | 333/809 [06:38<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4368     0.1491        261        640:  41%|████▏     | 335/809 [06:40<09:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4368      0.149        312        640:  42%|████▏     | 336/809 [06:41<09:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4364     0.1491        186        640:  42%|████▏     | 338/809 [06:44<09:45,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4366     0.1494        158        640:  42%|████▏     | 340/809 [06:46<09:50,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4364     0.1494        344        640:  42%|████▏     | 341/809 [06:48<09:40,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4363     0.1492        305        640:  42%|████▏     | 343/809 [06:50<09:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4358     0.1493        588        640:  43%|████▎     | 345/809 [06:53<09:35,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4358     0.1491        264        640:  43%|████▎     | 347/809 [06:55<09:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.436      0.149        139        640:  43%|████▎     | 348/809 [06:56<09:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217      0.436     0.1492        274        640:  43%|████▎     | 350/809 [06:58<09:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4359     0.1492        311        640:  44%|████▎     | 352/809 [07:01<09:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4358     0.1491        363        640:  44%|████▎     | 353/809 [07:02<09:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4361     0.1491        184        640:  44%|████▍     | 355/809 [07:04<09:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4366     0.1489        204        640:  44%|████▍     | 357/809 [07:07<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4363      0.149        420        640:  44%|████▍     | 358/809 [07:08<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4361     0.1491        288        640:  44%|████▍     | 360/809 [07:10<08:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4361     0.1489        151        640:  45%|████▍     | 362/809 [07:13<08:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4361      0.149        353        640:  45%|████▍     | 363/809 [07:14<08:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.436     0.1488        247        640:  45%|████▌     | 365/809 [07:16<08:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4365     0.1486        200        640:  45%|████▌     | 367/809 [07:19<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4365     0.1486        357        640:  46%|████▌     | 369/809 [07:21<08:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214      0.437     0.1484        151        640:  46%|████▌     | 370/809 [07:22<08:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4373     0.1487        517        640:  46%|████▌     | 372/809 [07:25<08:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.437     0.1488        392        640:  46%|████▌     | 374/809 [07:27<08:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4369     0.1487        357        640:  46%|████▋     | 375/809 [07:28<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.437     0.1486        205        640:  47%|████▋     | 377/809 [07:31<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4372     0.1484        300        640:  47%|████▋     | 379/809 [07:33<08:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4373     0.1484        263        640:  47%|████▋     | 380/809 [07:34<08:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4376     0.1481        144        640:  47%|████▋     | 382/809 [07:37<08:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4373     0.1487        252        640:  47%|████▋     | 384/809 [07:39<08:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4375     0.1487        197        640:  48%|████▊     | 385/809 [07:40<08:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4379     0.1485        185        640:  48%|████▊     | 387/809 [07:43<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4383     0.1483        166        640:  48%|████▊     | 389/809 [07:45<08:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4382     0.1482        329        640:  48%|████▊     | 390/809 [07:47<08:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4381     0.1482        206        640:  48%|████▊     | 392/809 [07:48<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4378     0.1481        335        640:  49%|████▊     | 394/809 [07:51<08:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4381      0.148        351        640:  49%|████▉     | 396/809 [07:53<08:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4382     0.1479        299        640:  49%|████▉     | 397/809 [07:54<08:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4382     0.1478        284        640:  49%|████▉     | 399/809 [07:57<08:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212      0.438     0.1476        232        640:  50%|████▉     | 401/809 [07:59<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4382     0.1475        181        640:  50%|████▉     | 402/809 [08:00<08:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4382     0.1475        150        640:  50%|████▉     | 404/809 [08:03<07:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4379     0.1475        417        640:  50%|█████     | 406/809 [08:05<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4379     0.1475        263        640:  50%|█████     | 407/809 [08:06<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4377     0.1479        287        640:  51%|█████     | 409/809 [08:09<08:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4378     0.1478        181        640:  51%|█████     | 411/809 [08:11<07:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4381     0.1476        206        640:  51%|█████     | 413/809 [08:13<07:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4377     0.1477        256        640:  51%|█████     | 414/809 [08:14<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4377     0.1479        122        640:  51%|█████▏    | 416/809 [08:17<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4376     0.1485        299        640:  52%|█████▏    | 418/809 [08:19<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4377     0.1484        362        640:  52%|█████▏    | 419/809 [08:20<07:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4378     0.1482        339        640:  52%|█████▏    | 421/809 [08:23<07:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4377     0.1486        406        640:  52%|█████▏    | 423/809 [08:25<07:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4375     0.1494        324        640:  52%|█████▏    | 424/809 [08:26<07:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4372     0.1493        150        640:  53%|█████▎    | 426/809 [08:29<07:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4367     0.1499        163        640:  53%|█████▎    | 428/809 [08:31<07:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4369     0.1498         94        640:  53%|█████▎    | 429/809 [08:32<07:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4365     0.1498        329        640:  53%|█████▎    | 431/809 [08:35<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4359     0.1513        231        640:  54%|█████▎    | 433/809 [08:37<07:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4356     0.1522        371        640:  54%|█████▎    | 434/809 [08:39<07:51,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4356      0.152        397        640:  54%|█████▍    | 436/809 [08:41<07:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4354     0.1521        431        640:  54%|█████▍    | 438/809 [08:43<07:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4352     0.1522        381        640:  54%|█████▍    | 439/809 [08:45<07:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4347     0.1524        384        640:  55%|█████▍    | 441/809 [08:47<07:30,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4346     0.1525        414        640:  55%|█████▍    | 442/809 [08:48<07:37,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4344     0.1526        310        640:  55%|█████▍    | 444/809 [08:51<07:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.22     0.4346     0.1527        390        640:  55%|█████▌    | 446/809 [08:53<07:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4348     0.1525        178        640:  55%|█████▌    | 448/809 [08:55<07:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4346     0.1526        326        640:  56%|█████▌    | 449/809 [08:57<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4345     0.1526        382        640:  56%|█████▌    | 451/809 [08:59<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4344     0.1525        128        640:  56%|█████▌    | 453/809 [09:01<07:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.219     0.4345     0.1525        308        640:  56%|█████▌    | 454/809 [09:03<06:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4348     0.1523        266        640:  56%|█████▋    | 456/809 [09:05<06:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4348     0.1525        185        640:  57%|█████▋    | 458/809 [09:07<07:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4348     0.1524        259        640:  57%|█████▋    | 459/809 [09:08<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4349     0.1522        273        640:  57%|█████▋    | 461/809 [09:11<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1522        300        640:  57%|█████▋    | 463/809 [09:13<06:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4352      0.152        247        640:  57%|█████▋    | 464/809 [09:14<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4353      0.152        305        640:  58%|█████▊    | 466/809 [09:17<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4352     0.1519        197        640:  58%|█████▊    | 468/809 [09:19<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.435     0.1519        236        640:  58%|█████▊    | 470/809 [09:21<06:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348      0.152        438        640:  58%|█████▊    | 471/809 [09:23<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4346      0.152        280        640:  58%|█████▊    | 473/809 [09:25<06:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4345     0.1518        271        640:  59%|█████▊    | 475/809 [09:28<06:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4345     0.1518        311        640:  59%|█████▉    | 476/809 [09:29<06:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1517        154        640:  59%|█████▉    | 478/809 [09:31<06:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348     0.1518        237        640:  59%|█████▉    | 480/809 [09:33<06:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4347     0.1519        280        640:  59%|█████▉    | 481/809 [09:35<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4351     0.1517        192        640:  60%|█████▉    | 483/809 [09:37<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.435     0.1516        256        640:  60%|█████▉    | 485/809 [09:39<06:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4352     0.1515        240        640:  60%|██████    | 486/809 [09:40<06:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4351     0.1518        186        640:  60%|██████    | 488/809 [09:43<06:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1522        228        640:  61%|██████    | 490/809 [09:46<06:38,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1522        295        640:  61%|██████    | 491/809 [09:47<06:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4348     0.1523        229        640:  61%|██████    | 493/809 [09:49<06:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4347     0.1523        197        640:  61%|██████    | 495/809 [09:52<06:30,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348     0.1522        123        640:  61%|██████▏   | 496/809 [09:53<06:26,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1522        279        640:  62%|██████▏   | 498/809 [09:55<06:21,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.435      0.152        260        640:  62%|██████▏   | 500/809 [09:58<06:21,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4351     0.1519        213        640:  62%|██████▏   | 502/809 [10:00<06:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4351     0.1518        331        640:  62%|██████▏   | 503/809 [10:01<06:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4352     0.1516        109        640:  62%|██████▏   | 505/809 [10:04<05:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4346     0.1517        216        640:  63%|██████▎   | 507/809 [10:06<06:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4347     0.1517        154        640:  63%|██████▎   | 508/809 [10:07<06:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4348     0.1517        243        640:  63%|██████▎   | 510/809 [10:10<05:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4346     0.1515        259        640:  63%|██████▎   | 512/809 [10:12<05:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4347     0.1515        178        640:  63%|██████▎   | 513/809 [10:13<05:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4346     0.1514        150        640:  64%|██████▎   | 515/809 [10:16<05:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4348     0.1513        182        640:  64%|██████▍   | 517/809 [10:18<05:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4351     0.1513        146        640:  64%|██████▍   | 518/809 [10:19<05:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4354     0.1512        247        640:  64%|██████▍   | 520/809 [10:22<05:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4355      0.151        236        640:  65%|██████▍   | 522/809 [10:24<05:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4355     0.1511        224        640:  65%|██████▍   | 523/809 [10:25<05:56,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4354      0.151        233        640:  65%|██████▍   | 525/809 [10:28<05:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4352      0.151        141        640:  65%|██████▌   | 527/809 [10:30<05:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4352      0.151        106        640:  65%|██████▌   | 528/809 [10:31<05:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4348     0.1511        340        640:  66%|██████▌   | 530/809 [10:34<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4348      0.151        241        640:  66%|██████▌   | 532/809 [10:36<05:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4347      0.151        308        640:  66%|██████▌   | 533/809 [10:37<05:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4349     0.1508        237        640:  66%|██████▌   | 535/809 [10:40<05:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348     0.1512        357        640:  66%|██████▋   | 537/809 [10:42<05:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4346     0.1511        250        640:  67%|██████▋   | 538/809 [10:43<05:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4348     0.1509        202        640:  67%|██████▋   | 540/809 [10:46<05:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4351     0.1507        199        640:  67%|██████▋   | 542/809 [10:48<05:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1508        582        640:  67%|██████▋   | 544/809 [10:50<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.435     0.1507        226        640:  67%|██████▋   | 545/809 [10:51<05:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4352     0.1508        128        640:  68%|██████▊   | 547/809 [10:54<05:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4352     0.1507        418        640:  68%|██████▊   | 548/809 [10:55<05:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.435     0.1506        247        640:  68%|██████▊   | 550/809 [10:57<05:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4352     0.1505        130        640:  68%|██████▊   | 552/809 [11:00<05:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216      0.435     0.1506        133        640:  68%|██████▊   | 554/809 [11:02<05:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4349     0.1506        372        640:  69%|██████▊   | 555/809 [11:03<05:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4347     0.1507        509        640:  69%|██████▉   | 557/809 [11:06<04:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4347     0.1507        229        640:  69%|██████▉   | 559/809 [11:08<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4346     0.1507        455        640:  69%|██████▉   | 560/809 [11:09<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4345     0.1506        499        640:  69%|██████▉   | 562/809 [11:12<04:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4347     0.1505        252        640:  70%|██████▉   | 564/809 [11:14<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4346     0.1504        258        640:  70%|██████▉   | 565/809 [11:15<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4348     0.1505        197        640:  70%|███████   | 567/809 [11:18<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4348     0.1503        140        640:  70%|███████   | 569/809 [11:20<04:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4348     0.1503        239        640:  70%|███████   | 570/809 [11:21<04:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4346     0.1502        224        640:  71%|███████   | 572/809 [11:24<04:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4349     0.1501        222        640:  71%|███████   | 574/809 [11:26<04:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4349       0.15        248        640:  71%|███████   | 576/809 [11:28<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217      0.435     0.1499         95        640:  71%|███████▏  | 577/809 [11:29<04:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4349     0.1498        316        640:  72%|███████▏  | 579/809 [11:32<04:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217      0.435     0.1496        299        640:  72%|███████▏  | 581/809 [11:34<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4351     0.1496        187        640:  72%|███████▏  | 582/809 [11:35<04:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4354     0.1494        181        640:  72%|███████▏  | 584/809 [11:38<04:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4353     0.1493        188        640:  72%|███████▏  | 586/809 [11:40<04:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4355     0.1493        129        640:  73%|███████▎  | 587/809 [11:41<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4357     0.1491        183        640:  73%|███████▎  | 589/809 [11:44<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1489        242        640:  73%|███████▎  | 591/809 [11:46<04:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4357     0.1488        372        640:  73%|███████▎  | 593/809 [11:48<04:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4357     0.1488        399        640:  73%|███████▎  | 594/809 [11:50<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4357      0.149        312        640:  74%|███████▎  | 596/809 [11:52<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4359     0.1488        297        640:  74%|███████▍  | 598/809 [11:54<04:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4361     0.1487        208        640:  74%|███████▍  | 599/809 [11:55<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4361     0.1487        398        640:  74%|███████▍  | 601/809 [11:58<04:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4363     0.1485        282        640:  75%|███████▍  | 603/809 [12:00<04:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4364     0.1485        170        640:  75%|███████▍  | 604/809 [12:02<04:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4362     0.1487        326        640:  75%|███████▍  | 606/809 [12:04<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4362     0.1485        140        640:  75%|███████▌  | 608/809 [12:06<03:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4361     0.1485        403        640:  75%|███████▌  | 609/809 [12:08<03:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4362     0.1484        190        640:  76%|███████▌  | 611/809 [12:10<03:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215      0.436     0.1484        283        640:  76%|███████▌  | 613/809 [12:12<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1483        370        640:  76%|███████▌  | 614/809 [12:13<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4357     0.1484        221        640:  76%|███████▌  | 616/809 [12:16<03:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4355     0.1484        123        640:  76%|███████▋  | 618/809 [12:18<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4357     0.1482        222        640:  77%|███████▋  | 620/809 [12:20<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4358     0.1481        199        640:  77%|███████▋  | 621/809 [12:22<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1481        177        640:  77%|███████▋  | 623/809 [12:24<03:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214      0.436      0.148        134        640:  77%|███████▋  | 625/809 [12:26<03:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214      0.436     0.1479        209        640:  77%|███████▋  | 626/809 [12:28<03:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4362     0.1478        227        640:  78%|███████▊  | 628/809 [12:30<03:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4361     0.1477        187        640:  78%|███████▊  | 630/809 [12:32<03:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1476        108        640:  78%|███████▊  | 631/809 [12:33<03:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4362     0.1476        288        640:  78%|███████▊  | 633/809 [12:36<03:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4362     0.1476        185        640:  78%|███████▊  | 635/809 [12:38<03:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4364     0.1475        306        640:  79%|███████▊  | 637/809 [12:41<03:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4363     0.1475        158        640:  79%|███████▉  | 638/809 [12:42<03:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1474         68        640:  79%|███████▉  | 640/809 [12:44<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4367     0.1473         79        640:  79%|███████▉  | 641/809 [12:45<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4369     0.1472        190        640:  79%|███████▉  | 643/809 [12:48<03:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4369     0.1472        234        640:  80%|███████▉  | 645/809 [12:50<03:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21      0.437      0.147        149        640:  80%|███████▉  | 647/809 [12:52<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4371      0.147        191        640:  80%|████████  | 648/809 [12:54<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4373     0.1469        260        640:  80%|████████  | 650/809 [12:56<03:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4375     0.1468        243        640:  81%|████████  | 652/809 [12:58<03:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4374     0.1468        176        640:  81%|████████  | 654/809 [13:01<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4374     0.1468        297        640:  81%|████████  | 656/809 [13:03<03:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4376     0.1468         87        640:  81%|████████  | 657/809 [13:04<03:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.209     0.4376     0.1469        452        640:  81%|████████▏ | 659/809 [13:07<03:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4379     0.1468        176        640:  82%|████████▏ | 661/809 [13:09<02:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4379     0.1468        243        640:  82%|████████▏ | 662/809 [13:10<02:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4379     0.1469        420        640:  82%|████████▏ | 664/809 [13:13<02:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4382     0.1467        237        640:  82%|████████▏ | 666/809 [13:15<02:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4381      0.147        496        640:  82%|████████▏ | 667/809 [13:16<02:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4384     0.1469         63        640:  83%|████████▎ | 669/809 [13:19<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4382      0.147        326        640:  83%|████████▎ | 671/809 [13:21<02:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386     0.1469        146        640:  83%|████████▎ | 672/809 [13:22<02:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4382     0.1472        286        640:  83%|████████▎ | 674/809 [13:25<02:46,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4381     0.1475        473        640:  83%|████████▎ | 675/809 [13:26<02:47,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208      0.438     0.1474        406        640:  84%|████████▎ | 677/809 [13:29<02:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208      0.438     0.1477        285        640:  84%|████████▍ | 679/809 [13:31<02:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4381     0.1477        186        640:  84%|████████▍ | 680/809 [13:32<02:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386     0.1475         56        640:  84%|████████▍ | 682/809 [13:35<02:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386     0.1474        255        640:  85%|████████▍ | 684/809 [13:37<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386     0.1474        313        640:  85%|████████▍ | 685/809 [13:38<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4385      0.148         98        640:  85%|████████▍ | 687/809 [13:41<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4385      0.148        165        640:  85%|████████▌ | 689/809 [13:43<02:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4384     0.1481        298        640:  85%|████████▌ | 690/809 [13:44<02:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4382     0.1481        356        640:  86%|████████▌ | 692/809 [13:46<02:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4384      0.148        120        640:  86%|████████▌ | 694/809 [13:49<02:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4383      0.148        365        640:  86%|████████▌ | 695/809 [13:50<02:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386     0.1479        295        640:  86%|████████▌ | 697/809 [13:52<02:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4387      0.148        239        640:  86%|████████▋ | 699/809 [13:55<02:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4385     0.1481        217        640:  87%|████████▋ | 701/809 [13:57<02:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4384     0.1481        301        640:  87%|████████▋ | 702/809 [13:58<02:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4382     0.1482        250        640:  87%|████████▋ | 704/809 [14:01<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208      0.438     0.1484        165        640:  87%|████████▋ | 706/809 [14:03<02:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208      0.438     0.1483        293        640:  87%|████████▋ | 707/809 [14:04<02:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4382     0.1482        158        640:  88%|████████▊ | 709/809 [14:07<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4385     0.1481        208        640:  88%|████████▊ | 711/809 [14:09<01:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.206     0.4386      0.148        198        640:  88%|████████▊ | 712/809 [14:10<01:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.207     0.4384     0.1481        426        640:  88%|████████▊ | 714/809 [14:13<01:58,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208      0.438     0.1481        360        640:  89%|████████▊ | 716/809 [14:15<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.208     0.4378     0.1481        296        640:  89%|████████▊ | 717/809 [14:16<01:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4374     0.1482        253        640:  89%|████████▉ | 719/809 [14:19<01:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4371     0.1482        174        640:  89%|████████▉ | 721/809 [14:21<01:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21     0.4369     0.1481        250        640:  89%|████████▉ | 722/809 [14:22<01:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G       1.21      0.437      0.148        360        640:  89%|████████▉ | 724/809 [14:25<01:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211      0.437      0.148        266        640:  90%|████████▉ | 726/809 [14:27<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4368     0.1482        425        640:  90%|████████▉ | 727/809 [14:28<01:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4369     0.1482        196        640:  90%|█████████ | 729/809 [14:31<01:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4368     0.1482        233        640:  90%|█████████ | 731/809 [14:33<01:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1482        318        640:  91%|█████████ | 733/809 [14:35<01:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4364     0.1483        239        640:  91%|████��████ | 734/809 [14:37<01:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4361     0.1491        445        640:  91%|█████████ | 736/809 [14:39<01:31,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1494        336        640:  91%|█████████ | 737/809 [14:40<01:29,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1493        207        640:  91%|█████████▏| 739/809 [14:43<01:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4362     0.1492        231        640:  92%|█████████▏| 741/809 [14:45<01:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1491        149        640:  92%|█████████▏| 742/809 [14:46<01:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4368     0.1489         74        640:  92%|█████████▏| 744/809 [14:49<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4368     0.1488        258        640:  92%|█████████▏| 746/809 [14:51<01:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4367     0.1488        198        640:  92%|█████████▏| 748/809 [14:53<01:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4368     0.1487        118        640:  93%|█████████▎| 749/809 [14:55<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4364     0.1488        630        640:  93%|█████████▎| 751/809 [14:57<01:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1488        164        640:  93%|█████████▎| 752/809 [14:58<01:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4361     0.1488        208        640:  93%|█████████▎| 754/809 [15:01<01:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1487        188        640:  93%|█████████▎| 756/809 [15:03<01:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1487        122        640:  94%|█████████▎| 758/809 [15:05<00:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4362     0.1488        324        640:  94%|█████████▍| 759/809 [15:07<00:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4365     0.1486         57        640:  94%|█████████▍| 761/809 [15:09<00:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1485        198        640:  94%|█████████▍| 763/809 [15:11<00:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1485        154        640:  94%|█████████▍| 764/809 [15:13<00:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4367     0.1485         96        640:  95%|█████████▍| 766/809 [15:15<00:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1484        201        640:  95%|█████████▍| 768/809 [15:17<00:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.211     0.4367     0.1483        217        640:  95%|█████████▌| 769/809 [15:18<00:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.212     0.4366     0.1482        200        640:  95%|█████████▌| 771/809 [15:21<00:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.213     0.4363     0.1485        402        640:  96%|█████████▌| 773/809 [15:23<00:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214      0.436     0.1486        206        640:  96%|█████████▌| 775/809 [15:25<00:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4357     0.1489        326        640:  96%|█████████▌| 776/809 [15:27<00:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4353      0.149        300        640:  96%|█████████▌| 778/809 [15:29<00:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4357     0.1488        175        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4356     0.1491        513        640:  97%|█████████▋| 781/809 [15:33<00:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4357     0.1491        296        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214      0.436     0.1489        197        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4363     0.1488         90        640:  97%|█████████▋| 786/809 [15:39<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4362     0.1488        257        640:  97%|█████████▋| 788/809 [15:41<00:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4362     0.1487         63        640:  98%|█████████▊| 790/809 [15:43<00:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.214     0.4361     0.1487        293        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4359     0.1489        390        640:  98%|█████████▊| 793/809 [15:47<00:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.215     0.4358      0.149        370        640:  98%|█████████▊| 795/809 [15:49<00:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4356      0.149        395        640:  98%|█████████▊| 796/809 [15:51<00:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4356     0.1491         98        640:  99%|█████████▊| 798/809 [15:53<00:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4354     0.1492        359        640:  99%|█████████▉| 800/809 [15:56<00:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4354     0.1491        119        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.216     0.4353     0.1492        461        640:  99%|█████████▉| 803/809 [15:59<00:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4351     0.1495        132        640: 100%|█████████▉| 805/809 [16:02<00:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.217     0.4351     0.1495        235        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     32/300      13.1G      1.218     0.4349     0.1502        598        640: 100%|█████████▉| 808/809 [16:05<00:01,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:01<00:15,  4.09it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:03<00:14,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:05<00:12,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:11,  3.65it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:09<00:08,  3.84it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:11<00:06,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:12<00:05,  3.72it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  84%|████████▍ | 58/69 [00:15<00:02,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       0.95     0.4876     0.1088         97        640:   0%|          | 2/809 [00:02<15:43,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G     0.9975     0.4848     0.1075        354        640:   0%|          | 3/809 [00:03<15:50,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.049     0.4757      0.108        329        640:   1%|          | 5/809 [00:05<15:47,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.095       0.46     0.1066        228        640:   1%|          | 7/809 [00:08<15:43,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.117     0.4551     0.1122        291        640:   1%|          | 8/809 [00:09<15:53,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4402     0.1184         99        640:   1%|          | 10/809 [00:11<16:05,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.144     0.4482     0.1155        263        640:   1%|▏         | 12/809 [00:14<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.163     0.4426     0.1207        125        640:   2%|▏         | 14/809 [00:16<15:56,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205      0.434     0.1327        409        640:   2%|▏         | 15/809 [00:18<16:22,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197      0.437     0.1281        277        640:   2%|▏         | 17/809 [00:20<16:13,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4442     0.1276        168        640:   2%|▏         | 19/809 [00:22<15:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19     0.4396     0.1269        159        640:   3%|▎         | 21/809 [00:25<15:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4419     0.1252        149        640:   3%|▎         | 22/809 [00:26<15:17,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.164     0.4468     0.1222        259        640:   3%|▎         | 24/809 [00:28<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.157     0.4483     0.1228        200        640:   3%|▎         | 26/809 [00:30<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.147     0.4505      0.122        203        640:   3%|▎         | 27/809 [00:32<15:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.14     0.4515     0.1213        174        640:   4%|▎         | 29/809 [00:34<15:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.171     0.4433     0.1277        330        640:   4%|▍         | 31/809 [00:36<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4392     0.1305        464        640:   4%|▍         | 32/809 [00:38<15:41,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4343     0.1345        575        640:   4%|▍         | 34/809 [00:40<15:53,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4355     0.1334         82        640:   4%|▍         | 36/809 [00:42<15:24,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4354     0.1324        157        640:   5%|▍         | 37/809 [00:44<15:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4347     0.1303        212        640:   5%|▍         | 39/809 [00:46<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4361     0.1298        155        640:   5%|▌         | 41/809 [00:48<15:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4354     0.1296        155        640:   5%|▌         | 43/809 [00:51<15:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.194     0.4332      0.137        407        640:   5%|▌         | 44/809 [00:52<15:26,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4371     0.1358        153        640:   6%|▌         | 46/809 [00:54<14:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.173     0.4385     0.1342         87        640:   6%|▌         | 48/809 [00:57<14:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4378     0.1346        165        640:   6%|▌         | 49/809 [00:58<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.166     0.4422     0.1329        200        640:   6%|▋         | 51/809 [01:00<15:07,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.169     0.4405     0.1326        155        640:   7%|▋         | 53/809 [01:03<14:54,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.166     0.4418     0.1322        213        640:   7%|▋         | 54/809 [01:04<14:44,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.167     0.4415     0.1315        314        640:   7%|▋         | 56/809 [01:06<14:42,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.167     0.4416      0.132        389        640:   7%|▋         | 58/809 [01:08<14:52,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.162     0.4446     0.1313        115        640:   7%|▋         | 60/809 [01:11<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.162     0.4446     0.1308        174        640:   8%|▊         | 61/809 [01:12<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.167     0.4439     0.1312        129        640:   8%|▊         | 63/809 [01:14<14:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.17     0.4436     0.1312        183        640:   8%|▊         | 65/809 [01:17<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.164     0.4449     0.1306        210        640:   8%|▊         | 66/809 [01:18<14:32,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.166     0.4439     0.1308        299        640:   8%|▊         | 68/809 [01:20<14:27,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.168     0.4442     0.1323         94        640:   9%|▊         | 70/809 [01:23<15:07,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.17     0.4436     0.1332        352        640:   9%|▉         | 71/809 [01:24<14:57,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.164     0.4458     0.1323        130        640:   9%|▉         | 73/809 [01:26<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.169     0.4442     0.1322        309        640:   9%|▉         | 75/809 [01:29<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.172      0.443     0.1322        189        640:   9%|▉         | 76/809 [01:30<14:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4417     0.1314         95        640:  10%|▉         | 78/809 [01:32<14:08,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.439     0.1342        274        640:  10%|▉         | 80/809 [01:34<14:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4387     0.1342        358        640:  10%|█         | 81/809 [01:36<15:10,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4403     0.1336        165        640:  10%|█         | 83/809 [01:38<14:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4391     0.1332        248        640:  11%|█         | 85/809 [01:41<14:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4392      0.133        281        640:  11%|█         | 86/809 [01:42<14:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187      0.438     0.1328        321        640:  11%|█         | 88/809 [01:44<14:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4385     0.1322        308        640:  11%|█         | 90/809 [01:46<14:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4382     0.1316        345        640:  11%|█▏        | 92/809 [01:49<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4369     0.1318        175        640:  11%|█▏        | 93/809 [01:50<14:25,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4366     0.1316        264        640:  12%|█▏        | 95/809 [01:52<14:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4372     0.1311        237        640:  12%|█▏        | 97/809 [01:55<14:21,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.191     0.4359     0.1325        774        640:  12%|█▏        | 98/809 [01:56<14:51,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4365      0.132        159        640:  12%|█▏        | 100/809 [01:59<14:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.191      0.436     0.1325        287        640:  12%|█▏        | 101/809 [02:00<14:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4363     0.1322        182        640:  13%|█▎        | 103/809 [02:02<14:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.192     0.4361     0.1323        137        640:  13%|█▎        | 105/809 [02:05<14:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.191     0.4369     0.1321        152        640:  13%|█▎        | 107/809 [02:07<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.194     0.4362     0.1331        369        640:  13%|█▎        | 108/809 [02:08<14:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187      0.438     0.1319        146        640:  14%|█▎        | 110/809 [02:10<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4408     0.1309        109        640:  14%|█▍        | 112/809 [02:13<13:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4415     0.1305        237        640:  14%|█▍        | 113/809 [02:14<13:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4423     0.1302        234        640:  14%|█▍        | 115/809 [02:16<13:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.175     0.4424       0.13        261        640:  14%|█▍        | 116/809 [02:18<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4423     0.1296        378        640:  15%|█▍        | 118/809 [02:20<14:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4414     0.1296        167        640:  15%|█▍        | 120/809 [02:23<13:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4405     0.1302        338        640:  15%|█▌        | 122/809 [02:25<13:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4411     0.1301        224        640:  15%|█▌        | 123/809 [02:26<13:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.173     0.4421     0.1297        194        640:  15%|█▌        | 125/809 [02:29<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4417     0.1297        276        640:  16%|█▌        | 126/809 [02:30<13:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4414     0.1317        329        640:  16%|█▌        | 128/809 [02:32<13:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4413     0.1318        163        640:  16%|█▌        | 130/809 [02:35<13:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4419     0.1317        138        640:  16%|█▋        | 132/809 [02:37<13:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4415     0.1322        292        640:  16%|█▋        | 133/809 [02:38<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4413     0.1324        388        640:  17%|█▋        | 135/809 [02:41<13:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4417     0.1324        156        640:  17%|█▋        | 137/809 [02:43<13:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4409     0.1337        412        640:  17%|█▋        | 138/809 [02:44<13:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4417     0.1336        206        640:  17%|█▋        | 140/809 [02:47<13:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4417     0.1334        137        640:  18%|█▊        | 142/809 [02:49<12:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4423     0.1333        203        640:  18%|█▊        | 143/809 [02:50<12:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181      0.442     0.1349        350        640:  18%|█▊        | 145/809 [02:52<13:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4418     0.1348        278        640:  18%|█▊        | 147/809 [02:55<13:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4413     0.1348        320        640:  18%|█▊        | 148/809 [02:56<13:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4416     0.1345        130        640:  19%|█▊        | 150/809 [02:58<13:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4417     0.1344        184        640:  19%|█▉        | 152/809 [03:01<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182      0.442     0.1341        358        640:  19%|█▉        | 154/809 [03:03<12:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4419     0.1339        170        640:  19%|█▉        | 155/809 [03:04<12:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4407     0.1344        191        640:  19%|█▉        | 157/809 [03:07<12:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4391     0.1352        319        640:  20%|█▉        | 159/809 [03:09<12:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187     0.4402     0.1348        178        640:  20%|█▉        | 161/809 [03:11<12:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187     0.4405     0.1347        130        640:  20%|██        | 162/809 [03:13<12:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4401     0.1349        315        640:  20%|██        | 164/809 [03:15<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4408     0.1342        166        640:  21%|██        | 166/809 [03:17<12:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.441      0.134        251        640:  21%|██        | 167/809 [03:18<12:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4411     0.1335        182        640:  21%|██        | 169/809 [03:21<12:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4417     0.1332         57        640:  21%|██        | 171/809 [03:23<12:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.175     0.4427     0.1328        173        640:  21%|██▏       | 173/809 [03:26<12:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4432     0.1327        244        640:  22%|██▏       | 174/809 [03:27<12:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174      0.443     0.1328        383        640:  22%|██▏       | 176/809 [03:29<12:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.172     0.4438     0.1325        195        640:  22%|██▏       | 178/809 [03:32<12:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.17     0.4448     0.1324        143        640:  22%|██▏       | 179/809 [03:33<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.169     0.4454     0.1322        221        640:  22%|██▏       | 181/809 [03:35<12:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.168     0.4458      0.132        273        640:  23%|██▎       | 183/809 [03:37<12:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.167      0.446      0.132        230        640:  23%|██▎       | 184/809 [03:39<12:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.166     0.4464     0.1318        149        640:  23%|██▎       | 186/809 [03:41<12:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.168     0.4457     0.1317        316        640:  23%|██▎       | 188/809 [03:43<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.167     0.4459     0.1316        185        640:  23%|██▎       | 189/809 [03:44<12:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.169     0.4456     0.1315        298        640:  24%|██▎       | 191/809 [03:47<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.171     0.4448     0.1319        254        640:  24%|██▍       | 193/809 [03:49<12:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.171     0.4446     0.1318        254        640:  24%|██▍       | 195/809 [03:52<12:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.171     0.4448     0.1318        243        640:  24%|██▍       | 196/809 [03:53<12:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.172     0.4445     0.1316        318        640:  24%|██▍       | 198/809 [03:55<11:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.17     0.4455     0.1314         93        640:  25%|██▍       | 200/809 [03:58<11:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.172     0.4451     0.1322        312        640:  25%|██▍       | 201/809 [03:59<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4445     0.1323        163        640:  25%|██▌       | 203/809 [04:01<12:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176      0.444     0.1324        440        640:  25%|██▌       | 205/809 [04:04<12:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4441     0.1324        330        640:  25%|██▌       | 206/809 [04:05<11:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4442     0.1323        248        640:  26%|██▌       | 208/809 [04:07<11:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4444     0.1321        291        640:  26%|██▌       | 210/809 [04:10<11:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177      0.444     0.1325        298        640:  26%|██▌       | 211/809 [04:11<11:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4437     0.1326        155        640:  26%|██▋       | 213/809 [04:13<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4435     0.1328        173        640:  27%|██▋       | 215/809 [04:15<11:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4431     0.1332        379        640:  27%|██▋       | 216/809 [04:17<11:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4432     0.1329        244        640:  27%|██▋       | 218/809 [04:19<11:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18      0.443     0.1331        239        640:  27%|██▋       | 220/809 [04:21<11:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4431     0.1332        173        640:  27%|██▋       | 222/809 [04:24<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4429     0.1332        137        640:  28%|██▊       | 223/809 [04:25<11:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4437     0.1331        108        640:  28%|██▊       | 225/809 [04:27<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4436     0.1336        463        640:  28%|██▊       | 227/809 [04:30<11:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4442     0.1337        216        640:  28%|██▊       | 228/809 [04:31<11:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4441     0.1343        396        640:  28%|██▊       | 230/809 [04:33<11:54,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4437     0.1343        194        640:  29%|██▊       | 232/809 [04:36<11:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.176     0.4444      0.134        149        640:  29%|██▉       | 233/809 [04:37<11:47,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4439     0.1341        285        640:  29%|██▉       | 235/809 [04:39<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4433     0.1341        294        640:  29%|██▉       | 237/809 [04:42<11:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4428     0.1343        342        640:  29%|██▉       | 238/809 [04:43<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4421     0.1344        155        640:  30%|██▉       | 240/809 [04:45<11:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4424     0.1341        201        640:  30%|██▉       | 242/809 [04:48<11:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4421     0.1342        215        640:  30%|███       | 243/809 [04:49<11:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4415     0.1354        449        640:  30%|███       | 245/809 [04:51<11:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.441     0.1354        241        640:  31%|███       | 247/809 [04:54<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4409     0.1354        204        640:  31%|███       | 248/809 [04:55<11:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183      0.441     0.1352        260        640:  31%|███       | 250/809 [04:57<11:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4404     0.1355        330        640:  31%|███       | 252/809 [05:00<11:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187     0.4399     0.1361        522        640:  31%|███▏      | 253/809 [05:01<11:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4395     0.1365        185        640:  32%|███▏      | 255/809 [05:03<11:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4396     0.1365        211        640:  32%|███▏      | 257/809 [05:06<11:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19     0.4393     0.1371        416        640:  32%|███▏      | 258/809 [05:07<11:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4398     0.1368        165        640:  32%|███▏      | 260/809 [05:09<10:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19     0.4397     0.1381        458        640:  32%|███▏      | 262/809 [05:12<11:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19     0.4394     0.1381        221        640:  33%|███▎      | 263/809 [05:13<10:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4397      0.138        163        640:  33%|███▎      | 265/809 [05:15<10:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4404     0.1376        190        640:  33%|███▎      | 267/809 [05:18<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4402     0.1378        422        640:  33%|███▎      | 268/809 [05:19<10:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4404     0.1374        259        640:  33%|███▎      | 270/809 [05:21<10:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4407      0.138        142        640:  34%|███▎      | 272/809 [05:24<10:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4411     0.1377         82        640:  34%|███▎      | 273/809 [05:25<10:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4414     0.1376        189        640:  34%|███▍      | 275/809 [05:27<10:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4417     0.1374        228        640:  34%|███▍      | 277/809 [05:30<10:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178      0.442     0.1372        128        640:  34%|███▍      | 279/809 [05:32<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4421      0.137        220        640:  35%|███▍      | 280/809 [05:33<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.175     0.4428     0.1366        162        640:  35%|███▍      | 282/809 [05:35<10:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4431     0.1364        238        640:  35%|███▌      | 284/809 [05:38<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.175     0.4424     0.1367        354        640:  35%|███▌      | 285/809 [05:39<10:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4427     0.1364         85        640:  35%|███▌      | 287/809 [05:41<10:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174      0.443     0.1366        171        640:  36%|███▌      | 289/809 [05:44<10:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.174     0.4428     0.1366        250        640:  36%|███▌      | 290/809 [05:45<10:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.175     0.4427     0.1365        223        640:  36%|███▌      | 292/809 [05:47<10:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4421     0.1366        201        640:  36%|███▋      | 294/809 [05:50<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4417     0.1369        324        640:  36%|███▋      | 295/809 [05:51<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4409      0.137        355        640:  37%|███▋      | 297/809 [05:53<10:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18      0.441     0.1372        301        640:  37%|███▋      | 299/809 [05:56<10:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4411      0.137        202        640:  37%|███▋      | 300/809 [05:57<10:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179      0.441     0.1371        287        640:  37%|███▋      | 302/809 [05:59<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4409     0.1369        251        640:  38%|███▊      | 304/809 [06:02<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178      0.441     0.1368        309        640:  38%|███▊      | 305/809 [06:03<10:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4407     0.1368        418        640:  38%|███▊      | 307/809 [06:05<10:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4406      0.137        185        640:  38%|███▊      | 309/809 [06:08<10:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4409     0.1368        151        640:  38%|███▊      | 310/809 [06:09<09:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4407     0.1377        218        640:  39%|███▊      | 312/809 [06:11<09:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4408     0.1376        168        640:  39%|███▉      | 314/809 [06:14<10:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4412     0.1373        300        640:  39%|███▉      | 316/809 [06:16<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4412     0.1372        252        640:  39%|███▉      | 317/809 [06:17<09:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178      0.441     0.1371        414        640:  39%|███▉      | 319/809 [06:20<09:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4406     0.1374        324        640:  40%|███▉      | 321/809 [06:22<09:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4406     0.1374        420        640:  40%|███▉      | 322/809 [06:24<09:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4402     0.1376        189        640:  40%|████      | 324/809 [06:26<09:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4402     0.1375        200        640:  40%|████      | 326/809 [06:28<09:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4404     0.1374        281        640:  41%|████      | 328/809 [06:31<09:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.181     0.4406     0.1374        154        640:  41%|████      | 329/809 [06:32<09:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18      0.441     0.1372        186        640:  41%|████      | 331/809 [06:34<09:45,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4412     0.1371        211        640:  41%|████      | 332/809 [06:36<09:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4414     0.1371         74        640:  41%|████▏     | 334/809 [06:38<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4414      0.137        188        640:  42%|████▏     | 336/809 [06:40<09:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4417     0.1369        198        640:  42%|████▏     | 338/809 [06:43<09:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4416      0.137        256        640:  42%|████▏     | 339/809 [06:44<09:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4419     0.1369        220        640:  42%|████▏     | 341/809 [06:46<09:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4415     0.1368        326        640:  42%|████▏     | 343/809 [06:49<09:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179     0.4415     0.1368         77        640:  43%|████▎     | 344/809 [06:50<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178      0.442     0.1366         95        640:  43%|████▎     | 346/809 [06:52<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4416     0.1366        323        640:  43%|████▎     | 348/809 [06:54<09:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4419     0.1364        142        640:  43%|████▎     | 350/809 [06:57<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.177     0.4423     0.1363        142        640:  43%|████▎     | 351/809 [06:58<08:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.178     0.4422     0.1368        246        640:  44%|████▎     | 353/809 [07:00<08:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179      0.442     0.1377        282        640:  44%|████▍     | 355/809 [07:03<08:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.179      0.442     0.1376        327        640:  44%|████▍     | 356/809 [07:04<08:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.18     0.4418     0.1382        144        640:  44%|████▍     | 358/809 [07:06<09:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4412     0.1387        327        640:  44%|████▍     | 360/809 [07:09<09:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4412     0.1388        342        640:  45%|████▍     | 361/809 [07:10<09:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4412     0.1386        269        640:  45%|████▍     | 363/809 [07:12<09:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4412     0.1385        154        640:  45%|████▌     | 365/809 [07:15<08:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4408     0.1385        300        640:  45%|████▌     | 366/809 [07:16<09:07,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4414     0.1383        151        640:  45%|████▌     | 368/809 [07:19<09:07,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4409     0.1387        579        640:  46%|████▌     | 369/809 [07:20<09:21,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4409     0.1386        295        640:  46%|████▌     | 371/809 [07:22<09:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.441     0.1383        285        640:  46%|████▌     | 373/809 [07:25<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4415     0.1382        110        640:  46%|████▌     | 374/809 [07:26<08:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.182     0.4419      0.138        261        640:  46%|████▋     | 376/809 [07:28<08:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4415     0.1381        216        640:  47%|████▋     | 378/809 [07:31<08:48,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4414     0.1381        186        640:  47%|████▋     | 379/809 [07:32<08:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.441     0.1383        395        640:  47%|████▋     | 381/809 [07:34<08:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185      0.441     0.1383        223        640:  47%|████▋     | 383/809 [07:37<08:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.186     0.4405     0.1384        480        640:  47%|████▋     | 384/809 [07:38<08:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4407     0.1382        167        640:  48%|████▊     | 386/809 [07:40<08:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184      0.441      0.138        129        640:  48%|████▊     | 388/809 [07:42<08:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4412     0.1379        252        640:  48%|████▊     | 390/809 [07:45<08:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4413     0.1378        209        640:  48%|████▊     | 391/809 [07:46<08:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4413     0.1378        337        640:  49%|████▊     | 393/809 [07:48<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4413     0.1378        294        640:  49%|████▉     | 395/809 [07:51<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.183     0.4412     0.1378        300        640:  49%|████▉     | 396/809 [07:52<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4412      0.138        321        640:  49%|████▉     | 398/809 [07:54<08:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4412     0.1379        284        640:  49%|████▉     | 400/809 [07:57<08:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185      0.441     0.1383        355        640:  50%|████▉     | 401/809 [07:58<08:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.184     0.4409     0.1381        218        640:  50%|████▉     | 403/809 [08:00<08:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4407      0.138        154        640:  50%|█████     | 405/809 [08:03<07:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.185     0.4406     0.1381        301        640:  50%|████��     | 406/809 [08:04<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187     0.4401     0.1389        342        640:  50%|█████     | 408/809 [08:06<08:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4399     0.1389        261        640:  51%|█████     | 410/809 [08:09<08:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.187     0.4401     0.1388        160        640:  51%|█████     | 411/809 [08:10<08:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4397     0.1389        291        640:  51%|█████     | 413/809 [08:13<08:05,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4394      0.139        276        640:  51%|█████▏    | 415/809 [08:15<07:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19      0.439     0.1399        399        640:  51%|█████▏    | 416/809 [08:16<08:19,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19      0.439     0.1399        278        640:  52%|█████▏    | 418/809 [08:19<08:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4393     0.1397        218        640:  52%|█████▏    | 419/809 [08:20<07:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189      0.439       0.14        555        640:  52%|█████▏    | 421/809 [08:22<07:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19      0.439     0.1399        305        640:  52%|█████▏    | 423/809 [08:25<07:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189      0.439     0.1398        157        640:  52%|█████▏    | 424/809 [08:26<07:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.188     0.4395     0.1399        117        640:  53%|█████▎    | 426/809 [08:28<07:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.189     0.4391     0.1399        321        640:  53%|█████▎    | 428/809 [08:31<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.19     0.4387       0.14        463        640:  53%|█████▎    | 429/809 [08:32<07:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.192      0.438     0.1404        338        640:  53%|█████▎    | 431/809 [08:34<07:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.192     0.4378     0.1403        267        640:  54%|█████▎    | 433/809 [08:37<07:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.192     0.4376     0.1403        125        640:  54%|█████▎    | 434/809 [08:38<07:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.192     0.4375     0.1401        117        640:  54%|█████▍    | 436/809 [08:40<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.194      0.437     0.1411        371        640:  54%|█████▍    | 438/809 [08:43<07:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.194      0.437     0.1411        235        640:  54%|█████▍    | 439/809 [08:44<07:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.194      0.437     0.1409        319        640:  55%|█████▍    | 441/809 [08:46<07:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.193     0.4371     0.1407        158        640:  55%|█████▍    | 443/809 [08:49<07:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.193     0.4368     0.1407        293        640:  55%|█████▍    | 444/809 [08:50<07:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.193     0.4369     0.1406        136        640:  55%|█████▌    | 446/809 [08:52<07:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.193     0.4368     0.1406        209        640:  55%|█████▌    | 448/809 [08:55<07:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.193     0.4368     0.1405        208        640:  56%|█████▌    | 449/809 [08:56<07:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.195     0.4361     0.1406        367        640:  56%|█████▌    | 451/809 [08:58<07:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.195      0.436     0.1405        262        640:  56%|█████▌    | 453/809 [09:01<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.196     0.4355     0.1406        445        640:  56%|█████▌    | 454/809 [09:02<07:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.198      0.435     0.1406        214        640:  56%|█████▋    | 456/809 [09:04<07:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4354     0.1404        196        640:  57%|█████▋    | 458/809 [09:07<06:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4352     0.1404        351        640:  57%|█████▋    | 460/809 [09:09<06:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.196     0.4355     0.1403        260        640:  57%|█████▋    | 461/809 [09:10<06:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4355     0.1405        178        640:  57%|█████▋    | 463/809 [09:13<06:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4355     0.1404        299        640:  57%|█████▋    | 465/809 [09:15<06:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.196     0.4357     0.1404         90        640:  58%|█████▊    | 466/809 [09:16<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4356     0.1402        226        640:  58%|█████▊    | 468/809 [09:19<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4354     0.1403        184        640:  58%|█████▊    | 470/809 [09:21<06:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.197     0.4354     0.1402        182        640:  58%|█████▊    | 471/809 [09:22<06:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.198     0.4352     0.1403        444        640:  58%|█████▊    | 473/809 [09:25<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.199     0.4349     0.1405        470        640:  59%|█████▊    | 475/809 [09:27<06:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4347     0.1405        214        640:  59%|█████▉    | 477/809 [09:29<06:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4345     0.1404        203        640:  59%|█████▉    | 479/809 [09:32<06:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4344     0.1403        134        640:  59%|█████▉    | 480/809 [09:33<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4345     0.1403        310        640:  60%|█████▉    | 482/809 [09:35<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4344     0.1403        236        640:  60%|█████▉    | 484/809 [09:37<06:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4342      0.141        344        640:  60%|█████▉    | 485/809 [09:39<06:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4344      0.141        176        640:  60%|██████    | 487/809 [09:41<06:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G        1.2     0.4346      0.141        185        640:  60%|██████    | 489/809 [09:44<06:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4344      0.141        255        640:  61%|██████    | 490/809 [09:45<06:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.202      0.434     0.1415        147        640:  61%|██████    | 492/809 [09:47<06:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4343     0.1415         90        640:  61%|██████    | 494/809 [09:49<06:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.202     0.4341     0.1416        456        640:  61%|██████    | 495/809 [09:51<06:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203      0.434     0.1418        232        640:  61%|██████▏   | 497/809 [09:53<06:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4339     0.1418        416        640:  62%|██████▏   | 499/809 [09:55<06:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203      0.434     0.1417        123        640:  62%|██████▏   | 500/809 [09:57<06:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.202     0.4342     0.1416         89        640:  62%|██████▏   | 502/809 [09:59<06:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.202     0.4343     0.1416        267        640:  62%|██████▏   | 504/809 [10:01<06:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.201     0.4344     0.1415        265        640:  63%|██████▎   | 506/809 [10:04<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4341      0.142        550        640:  63%|██████▎   | 507/809 [10:05<06:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4342      0.142        393        640:  63%|██████▎   | 509/809 [10:07<06:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.202     0.4345     0.1419        274        640:  63%|██████▎   | 511/809 [10:10<05:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4345     0.1422        198        640:  63%|██████▎   | 512/809 [10:11<05:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4344     0.1424        188        640:  64%|██████▎   | 514/809 [10:13<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.203     0.4344     0.1424        391        640:  64%|██████▎   | 515/809 [10:15<06:13,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205      0.434     0.1436        520        640:  64%|██████▍   | 517/809 [10:17<06:05,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4342     0.1435        195        640:  64%|██████▍   | 518/809 [10:18<05:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4339     0.1436        402        640:  64%|██████▍   | 520/809 [10:21<06:04,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4339     0.1438        209        640:  65%|██████▍   | 522/809 [10:24<06:00,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4339     0.1438        231        640:  65%|██████▍   | 523/809 [10:25<05:52,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205      0.434     0.1438        238        640:  65%|██████▍   | 525/809 [10:27<05:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4342     0.1437        274        640:  65%|██████▌   | 527/809 [10:30<05:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4342     0.1438        308        640:  65%|██████▌   | 528/809 [10:31<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4342     0.1438        285        640:  66%|██████▌   | 530/809 [10:33<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.204     0.4345     0.1438        458        640:  66%|██████▌   | 532/809 [10:36<05:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.204     0.4345     0.1438        438        640:  66%|██████▌   | 533/809 [10:37<05:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4341     0.1445        679        640:  66%|██████▌   | 535/809 [10:39<05:41,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206      0.434     0.1445        180        640:  66%|██████▋   | 537/809 [10:42<05:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206      0.434     0.1445        265        640:  67%|██████▋   | 538/809 [10:43<05:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4339     0.1444        409        640:  67%|██████▋   | 540/809 [10:45<05:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4339     0.1445        479        640:  67%|██████▋   | 542/809 [10:48<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4338     0.1448        551        640:  67%|██████▋   | 543/809 [10:49<05:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4336     0.1447        243        640:  67%|██████▋   | 545/809 [10:51<05:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4338     0.1446        235        640:  68%|██████▊   | 547/809 [10:54<05:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206      0.434     0.1444        132        640:  68%|██████▊   | 548/809 [10:55<05:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4338     0.1444        230        640:  68%|██████▊   | 550/809 [10:57<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4339     0.1443        258        640:  68%|██████▊   | 552/809 [11:00<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4339     0.1443        144        640:  68%|██████▊   | 553/809 [11:01<05:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4338     0.1442        343        640:  69%|██████▊   | 555/809 [11:03<05:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4337     0.1442        186        640:  69%|██████▉   | 557/809 [11:06<04:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4337      0.144        169        640:  69%|██████▉   | 559/809 [11:08<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4335     0.1441        462        640:  69%|██████▉   | 560/809 [11:09<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4334     0.1441        183        640:  69%|██████▉   | 562/809 [11:12<04:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4333      0.144        240        640:  70%|██████▉   | 564/809 [11:14<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4334      0.144        308        640:  70%|██████▉   | 565/809 [11:15<04:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.205     0.4335     0.1439        184        640:  70%|███████   | 567/809 [11:17<04:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4333     0.1439        342        640:  70%|███████   | 569/809 [11:20<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.206     0.4331     0.1439        335        640:  70%|███████   | 570/809 [11:21<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4328     0.1441        185        640:  71%|███████   | 572/809 [11:23<04:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4328      0.144         99        640:  71%|███████   | 574/809 [11:26<04:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4326     0.1441        246        640:  71%|███████   | 575/809 [11:27<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4325     0.1441        201        640:  71%|███████▏  | 577/809 [11:29<04:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.209     0.4322     0.1444        547        640:  72%|███████▏  | 579/809 [11:32<04:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4324     0.1443        134        640:  72%|███████▏  | 580/809 [11:33<04:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4325     0.1441        120        640:  72%|███████▏  | 582/809 [11:35<04:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4324     0.1441        321        640:  72%|███████▏  | 584/809 [11:38<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.207     0.4325      0.144        269        640:  72%|███████▏  | 586/809 [11:40<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4323     0.1444        523        640:  73%|███████▎  | 587/809 [11:41<04:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4323     0.1445        192        640:  73%|███████▎  | 589/809 [11:44<04:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4322     0.1445        220        640:  73%|███████▎  | 591/809 [11:46<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4321     0.1445        340        640:  73%|███████▎  | 592/809 [11:47<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.209     0.4319     0.1449        438        640:  73%|███████▎  | 594/809 [11:50<04:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.209     0.4319     0.1449        334        640:  74%|███████▎  | 596/809 [11:52<04:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4321     0.1448        159        640:  74%|███████▍  | 597/809 [11:53<04:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4322      0.145        133        640:  74%|███████▍  | 599/809 [11:56<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.209     0.4321     0.1453        206        640:  74%|███████▍  | 601/809 [11:58<04:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4322     0.1452        239        640:  74%|███████▍  | 602/809 [11:59<04:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4323     0.1452        218        640:  75%|███████▍  | 604/809 [12:02<04:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4323     0.1451        252        640:  75%|███████▍  | 606/809 [12:04<04:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4321     0.1451        281        640:  75%|███████▌  | 607/809 [12:05<04:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4323      0.145        124        640:  75%|███████▌  | 609/809 [12:08<04:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.208     0.4321     0.1451        230        640:  76%|███████▌  | 611/809 [12:10<03:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.209      0.432     0.1451        209        640:  76%|███████▌  | 612/809 [12:11<03:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.21     0.4316     0.1453        358        640:  76%|███████▌  | 614/809 [12:14<03:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.211     0.4313     0.1454        146        640:  76%|███████▌  | 616/809 [12:16<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4311     0.1456        194        640:  76%|███████▋  | 617/809 [12:17<03:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212      0.431     0.1455        254        640:  77%|███████▋  | 619/809 [12:20<03:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.211     0.4312     0.1454        185        640:  77%|███████▋  | 621/809 [12:22<03:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4311     0.1455        168        640:  77%|███████▋  | 623/809 [12:24<03:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.211     0.4313     0.1454        212        640:  77%|███████▋  | 624/809 [12:25<03:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4312     0.1455        461        640:  77%|███████▋  | 626/809 [12:28<03:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.213      0.431     0.1456        167        640:  78%|███████▊  | 628/809 [12:30<03:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4311     0.1455        175        640:  78%|███████▊  | 630/809 [12:33<03:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4312     0.1454        149        640:  78%|███████▊  | 631/809 [12:34<03:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.211     0.4314     0.1453        161        640:  78%|███████▊  | 633/809 [12:36<03:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4314     0.1453        265        640:  78%|███████▊  | 635/809 [12:39<03:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4313     0.1453        257        640:  79%|███████▊  | 636/809 [12:40<03:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4312     0.1452        248        640:  79%|███████▉  | 638/809 [12:42<03:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.212     0.4313     0.1452        182        640:  79%|███████▉  | 640/809 [12:44<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.213     0.4312     0.1452        258        640:  79%|███████▉  | 641/809 [12:46<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.214     0.4308     0.1454        372        640:  79%|███████▉  | 643/809 [12:48<03:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.214     0.4307     0.1453        276        640:  80%|███████▉  | 645/809 [12:50<03:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.213     0.4308     0.1452         84        640:  80%|███████▉  | 647/809 [12:53<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.213     0.4308     0.1452        144        640:  80%|████████  | 648/809 [12:54<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.214     0.4306     0.1451        281        640:  80%|████████  | 650/809 [12:56<03:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4304     0.1456        465        640:  80%|████████  | 651/809 [12:58<03:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4303     0.1455        233        640:  81%|████████  | 653/809 [13:00<03:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4303     0.1457        265        640:  81%|████████  | 655/809 [13:02<03:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4302     0.1456        226        640:  81%|████████  | 656/809 [13:04<03:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4302     0.1455        240        640:  81%|████████▏ | 658/809 [13:06<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216       0.43     0.1457        274        640:  82%|████████▏ | 660/809 [13:08<03:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216       0.43     0.1457        357        640:  82%|████████▏ | 661/809 [13:10<03:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4299     0.1459        260        640:  82%|████████▏ | 663/809 [13:12<02:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297      0.146        314        640:  82%|████████▏ | 665/809 [13:15<02:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1459        136        640:  82%|████████▏ | 666/809 [13:16<02:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297      0.146        230        640:  83%|████████▎ | 668/809 [13:18<02:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4296     0.1459        272        640:  83%|████████▎ | 670/809 [13:20<02:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297     0.1459        123        640:  83%|████████▎ | 672/809 [13:23<02:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1458        124        640:  83%|████████▎ | 673/809 [13:24<02:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297     0.1458        438        640:  83%|████████▎ | 675/809 [13:26<02:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1457        181        640:  84%|████████▎ | 677/809 [13:29<02:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1457        264        640:  84%|████████▍ | 678/809 [13:30<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4299     0.1456        100        640:  84%|████████▍ | 680/809 [13:32<02:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4299     0.1455        224        640:  84%|████████▍ | 682/809 [13:35<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297     0.1455        221        640:  84%|████████▍ | 683/809 [13:36<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4296     0.1456         49        640:  85%|████████▍ | 685/809 [13:38<02:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4294     0.1455        274        640:  85%|████████▍ | 687/809 [13:41<02:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4293     0.1455        225        640:  85%|████████▌ | 688/809 [13:42<02:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4293     0.1454        292        640:  85%|████████▌ | 690/809 [13:45<02:43,  1.38s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4292     0.1455        334        640:  85%|████████▌ | 691/809 [13:46<02:36,  1.32s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4293     0.1454        140        640:  86%|████████▌ | 693/809 [13:48<02:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4294     0.1453        307        640:  86%|████████▌ | 695/809 [13:51<02:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4293     0.1453        364        640:  86%|████████▌ | 696/809 [13:52<02:32,  1.35s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4294     0.1455        178        640:  86%|████████▋ | 698/809 [13:55<02:19,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4297     0.1454         81        640:  86%|████████▋ | 699/809 [13:56<02:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217       0.43     0.1453        148        640:  87%|████████▋ | 701/809 [13:58<02:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4302     0.1452         95        640:  87%|████████▋ | 703/809 [14:01<02:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4303     0.1451        259        640:  87%|████████▋ | 704/809 [14:02<02:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4304     0.1451        205        640:  87%|████████▋ | 706/809 [14:04<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4305     0.1451        357        640:  88%|████████▊ | 708/809 [14:07<02:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4305     0.1451        436        640:  88%|████████▊ | 709/809 [14:08<02:06,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4305      0.145        245        640:  88%|████████▊ | 711/809 [14:10<02:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4306      0.145        439        640:  88%|████████▊ | 713/809 [14:13<01:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.215     0.4306     0.1449        309        640:  88%|████████▊ | 714/809 [14:14<01:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.214     0.4307     0.1448        111        640:  89%|████████▊ | 716/809 [14:16<01:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4302      0.145        290        640:  89%|████████▉ | 718/809 [14:19<01:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4302      0.145        176        640:  89%|████████▉ | 719/809 [14:20<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217       0.43     0.1451        262        640:  89%|████████▉ | 721/809 [14:22<01:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4301      0.145        147        640:  89%|████████▉ | 723/809 [14:25<01:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4302     0.1449        193        640:  89%|████████▉ | 724/809 [14:26<01:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216     0.4301     0.1448        188        640:  90%|████████▉ | 726/809 [14:28<01:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.216       0.43     0.1449        152        640:  90%|████████▉ | 728/809 [14:31<01:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4299     0.1452        432        640:  90%|█████████ | 729/809 [14:32<01:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1451        148        640:  90%|█████████ | 731/809 [14:34<01:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4296     0.1451        175        640:  91%|█████████ | 733/809 [14:37<01:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4294     0.1451        130        640:  91%|█████████ | 735/809 [14:39<01:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4292     0.1451        302        640:  91%|█████████ | 736/809 [14:40<01:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4294     0.1451        192        640:  91%|█████████ | 738/809 [14:43<01:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4296      0.145        130        640:  91%|█████████▏| 740/809 [14:45<01:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4295      0.145        230        640:  92%|█████████▏| 741/809 [14:46<01:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4298     0.1449        124        640:  92%|█████████▏| 743/809 [14:48<01:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217       0.43     0.1448        127        640:  92%|█████████▏| 745/809 [14:51<01:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.217     0.4299     0.1448        230        640:  92%|█████████▏| 746/809 [14:52<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4297     0.1448        188        640:  92%|█████████▏| 748/809 [14:54<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4295     0.1448        267        640:  93%|█████████▎| 750/809 [14:57<01:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4294     0.1448        241        640:  93%|█████████▎| 752/809 [14:59<01:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4294     0.1448        221        640:  93%|█████████▎| 753/809 [15:00<01:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4295     0.1447        180        640:  93%|█████████▎| 755/809 [15:02<01:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4293     0.1446        277        640:  94%|█████████▎| 757/809 [15:05<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.218     0.4292     0.1446        391        640:  94%|█████████▎| 758/809 [15:06<01:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4289     0.1447        194        640:  94%|█████████▍| 760/809 [15:08<00:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219      0.429     0.1446        169        640:  94%|█████████▍| 762/809 [15:11<00:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4287     0.1447        433        640:  94%|█████████▍| 763/809 [15:12<00:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4286     0.1447        250        640:  95%|█████████▍| 765/809 [15:15<00:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.221     0.4285     0.1446        189        640:  95%|█████████▍| 767/809 [15:17<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.221     0.4284     0.1447        363        640:  95%|█████████▍| 768/809 [15:18<00:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.221     0.4285     0.1447        327        640:  95%|█████████▌| 770/809 [15:20<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.221     0.4286     0.1446        191        640:  95%|█████████▌| 772/809 [15:23<00:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4287     0.1445        297        640:  96%|█████████▌| 773/809 [15:24<00:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219      0.429     0.1444        124        640:  96%|█████████▌| 775/809 [15:26<00:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4291     0.1443        215        640:  96%|█████████▌| 777/809 [15:29<00:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4291     0.1443        434        640:  96%|█████████▌| 778/809 [15:30<00:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4291     0.1442        260        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4292     0.1442        295        640:  97%|█████████▋| 782/809 [15:35<00:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4293     0.1441        159        640:  97%|█████████▋| 784/809 [15:37<00:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4295     0.1441        123        640:  97%|█████████▋| 785/809 [15:38<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4294      0.144        171        640:  97%|█████████▋| 787/809 [15:41<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4292      0.144        247        640:  98%|█████████▊| 789/809 [15:43<00:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4291      0.144         72        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.221     0.4288     0.1442        179        640:  98%|█████████▊| 793/809 [15:48<00:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4288     0.1442        240        640:  98%|█████████▊| 794/809 [15:49<00:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289     0.1442        391        640:  98%|█████████▊| 796/809 [15:51<00:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289     0.1442        303        640:  99%|█████████▊| 798/809 [15:54<00:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289     0.1441        167        640:  99%|█████████▉| 799/809 [15:55<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289     0.1442        139        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22      0.429     0.1441        179        640:  99%|█████████▉| 803/809 [15:59<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289     0.1441        374        640:  99%|█████████▉| 804/809 [16:01<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G       1.22     0.4289      0.144        182        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     33/300      13.2G      1.219     0.4292     0.1439        107        640: 100%|█████████▉| 808/809 [16:05<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.02it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  41%|████      | 28/69 [00:07<00:10,  3.84it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:09<00:08,  3.81it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:11<00:06,  3.81it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:04,  3.78it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.3936     0.1239        204        640:   0%|          | 2/809 [00:02<15:36,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.183     0.4169     0.1202        235        640:   0%|          | 4/809 [00:04<15:32,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4086     0.1224        198        640:   1%|          | 5/809 [00:05<15:38,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.322     0.3827     0.1619        376        640:   1%|          | 7/809 [00:08<16:02,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.341     0.3765     0.1637        356        640:   1%|          | 8/809 [00:09<16:15,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.328     0.3844     0.1561        250        640:   1%|          | 10/809 [00:11<15:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.355     0.3747     0.1805        132        640:   1%|▏         | 12/809 [00:14<15:54,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.32     0.3853     0.1698        259        640:   2%|▏         | 14/809 [00:16<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.327     0.3847     0.1757        458        640:   2%|▏         | 15/809 [00:17<15:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.296     0.3911     0.1676        308        640:   2%|▏         | 17/809 [00:20<16:15,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.284     0.3918     0.1644        300        640:   2%|▏         | 18/809 [00:21<16:30,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.3984     0.1587        394        640:   2%|▏         | 20/809 [00:24<16:01,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4082     0.1542        112        640:   3%|▎         | 22/809 [00:26<15:39,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.252     0.4038     0.1604        485        640:   3%|▎         | 23/809 [00:27<16:03,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254      0.402     0.1613        326        640:   3%|▎         | 25/809 [00:30<15:54,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4007     0.1597        379        640:   3%|▎         | 27/809 [00:32<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.261     0.3985     0.1589        300        640:   3%|▎         | 28/809 [00:33<15:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.27     0.3949     0.1584        300        640:   4%|▎         | 30/809 [00:36<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.283     0.3901     0.1591        352        640:   4%|▍         | 32/809 [00:38<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.285     0.3901     0.1595        327        640:   4%|▍         | 33/809 [00:39<15:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.277     0.3937     0.1582        305        640:   4%|▍         | 35/809 [00:42<15:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.281     0.3948     0.1603        264        640:   5%|▍         | 37/809 [00:44<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.281     0.3957     0.1597        387        640:   5%|▍         | 38/809 [00:45<15:23,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.274     0.3963     0.1573        315        640:   5%|▍         | 40/809 [00:48<15:31,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.269     0.3978      0.155        163        640:   5%|▌         | 42/809 [00:50<15:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4005     0.1538        104        640:   5%|▌         | 43/809 [00:51<15:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.261     0.3995      0.154        240        640:   6%|▌         | 45/809 [00:54<15:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.275     0.3964     0.1591        698        640:   6%|▌         | 47/809 [00:56<15:37,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.274     0.3977     0.1586        291        640:   6%|▌         | 48/809 [00:57<15:39,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.286     0.3949     0.1614        371        640:   6%|▌         | 50/809 [01:00<15:23,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.298     0.3909     0.1618        250        640:   6%|▋         | 52/809 [01:02<15:01,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.305     0.3895     0.1622        193        640:   7%|▋         | 53/809 [01:03<14:56,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.325     0.3835     0.1651        363        640:   7%|▋         | 55/809 [01:06<15:06,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.32     0.3838     0.1646        135        640:   7%|▋         | 57/809 [01:08<14:59,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.312     0.3861     0.1627        215        640:   7%|▋         | 59/809 [01:10<14:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.319     0.3845     0.1652        566        640:   7%|▋         | 60/809 [01:12<15:11,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.327     0.3822     0.1666        311        640:   8%|▊         | 62/809 [01:14<15:06,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.319     0.3848     0.1654        169        640:   8%|▊         | 63/809 [01:15<14:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.315     0.3868     0.1648        211        640:   8%|▊         | 65/809 [01:18<14:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.311     0.3883     0.1638        186        640:   8%|▊         | 67/809 [01:20<14:51,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.306     0.3901     0.1631        160        640:   8%|▊         | 68/809 [01:21<14:55,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.301     0.3929      0.162        263        640:   9%|▊         | 70/809 [01:24<14:51,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.295      0.396      0.161        400        640:   9%|▉         | 72/809 [01:26<14:53,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.29     0.3974       0.16        139        640:   9%|▉         | 73/809 [01:27<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.286     0.3993       0.16         92        640:   9%|▉         | 75/809 [01:30<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.289     0.3978     0.1618        722        640:  10%|▉         | 77/809 [01:32<15:06,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.283     0.3993     0.1609        157        640:  10%|▉         | 78/809 [01:33<14:46,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.29     0.3977     0.1649        511        640:  10%|▉         | 80/809 [01:36<15:41,  1.29s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.289     0.3974     0.1643        373        640:  10%|█         | 82/809 [01:38<14:56,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.286     0.3985     0.1634        142        640:  10%|█         | 83/809 [01:40<14:47,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.287     0.3982     0.1629        146        640:  11%|█         | 85/809 [01:42<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.289      0.397     0.1627        340        640:  11%|█         | 86/809 [01:43<14:38,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G        1.3     0.3944     0.1651        194        640:  11%|█         | 88/809 [01:46<14:56,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.294     0.3969     0.1639        205        640:  11%|█         | 90/809 [01:48<14:26,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.294     0.3964     0.1635        319        640:  11%|█▏        | 92/809 [01:50<14:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.296     0.3962     0.1633        314        640:  11%|█▏        | 93/809 [01:52<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.287     0.3983     0.1619        112        640:  12%|█▏        | 95/809 [01:54<14:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.283     0.3997     0.1609        133        640:  12%|█▏        | 97/809 [01:56<14:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.282     0.4001     0.1606        143        640:  12%|█▏        | 98/809 [01:58<13:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.286     0.3993     0.1617        186        640:  12%|█▏        | 100/809 [02:00<14:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.29     0.3983     0.1622        319        640:  13%|█▎        | 102/809 [02:02<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.289     0.3985     0.1619        204        640:  13%|█▎        | 103/809 [02:04<13:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.29      0.398     0.1613        298        640:  13%|█▎        | 105/809 [02:06<13:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.285     0.3996     0.1605         98        640:  13%|█▎        | 107/809 [02:08<13:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.285     0.3995     0.1604        302        640:  13%|█▎        | 108/809 [02:09<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.281     0.4006     0.1594        122        640:  14%|█▎        | 110/809 [02:12<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.276     0.4017     0.1581        120        640:  14%|█▍        | 112/809 [02:14<14:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.279     0.4011     0.1587        425        640:  14%|█▍        | 113/809 [02:15<13:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.282     0.4002      0.159        328        640:  14%|█▍        | 115/809 [02:18<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.287     0.3984     0.1594        259        640:  14%|█▍        | 117/809 [02:20<13:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.292     0.3968     0.1606        295        640:  15%|█▍        | 119/809 [02:22<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.295      0.396     0.1608        322        640:  15%|█▍        | 120/809 [02:24<13:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.289     0.3989     0.1597         85        640:  15%|█▌        | 122/809 [02:26<13:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.283     0.4007      0.159        253        640:  15%|█▌        | 124/809 [02:28<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.282     0.4012     0.1585        236        640:  15%|█▌        | 125/809 [02:30<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.271      0.405     0.1571        117        640:  16%|█▌        | 128/809 [02:33<13:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.268     0.4062     0.1567        133        640:  16%|█▌        | 129/809 [02:34<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.267     0.4065     0.1562        289        640:  16%|█▌        | 131/809 [02:37<13:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.268     0.4059     0.1559        159        640:  16%|█▋        | 133/809 [02:39<13:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.267     0.4061     0.1556        190        640:  17%|█▋        | 134/809 [02:40<13:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.267     0.4057     0.1556        425        640:  17%|█▋        | 136/809 [02:42<13:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.269     0.4055     0.1555        235        640:  17%|█▋        | 138/809 [02:45<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.269     0.4057     0.1554        264        640:  17%|█▋        | 139/809 [02:46<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.267     0.4063     0.1552         83        640:  17%|█▋        | 141/809 [02:48<13:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.264     0.4074     0.1544        259        640:  18%|█▊        | 143/809 [02:51<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.262     0.4075     0.1538        207        640:  18%|█▊        | 145/809 [02:53<13:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.261     0.4083     0.1536        113        640:  18%|█▊        | 146/809 [02:54<12:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.263     0.4076     0.1535        255        640:  18%|█▊        | 148/809 [02:57<12:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.264     0.4073     0.1535        206        640:  19%|█▊        | 150/809 [02:59<13:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.266     0.4066     0.1543        492        640:  19%|█▊        | 151/809 [03:00<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.265     0.4073     0.1537        255        640:  19%|█▉        | 153/809 [03:03<13:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.266     0.4072     0.1535        396        640:  19%|█▉        | 155/809 [03:05<13:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.268     0.4067     0.1533        227        640:  19%|█▉        | 156/809 [03:06<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.269     0.4069     0.1543        443        640:  20%|█▉        | 158/809 [03:09<12:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.271     0.4067     0.1558        497        640:  20%|█▉        | 160/809 [03:11<13:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.272     0.4065     0.1561        436        640:  20%|█▉        | 161/809 [03:12<13:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.27     0.4074     0.1553        177        640:  20%|██        | 163/809 [03:15<12:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.268     0.4082     0.1547        251        640:  20%|██        | 165/809 [03:17<12:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.266     0.4089     0.1542        271        640:  21%|██        | 166/809 [03:18<12:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.265     0.4092     0.1548        155        640:  21%|██        | 168/809 [03:21<12:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.266      0.409     0.1551        308        640:  21%|██        | 170/809 [03:23<12:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.267      0.409     0.1551        335        640:  21%|██        | 171/809 [03:24<12:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.265     0.4096     0.1547        411        640:  21%|██▏       | 173/809 [03:27<12:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.262     0.4105      0.154        109        640:  22%|██▏       | 175/809 [03:29<12:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.259     0.4114     0.1537        148        640:  22%|██▏       | 176/809 [03:30<12:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.259     0.4114     0.1534        149        640:  22%|██▏       | 178/809 [03:33<12:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.257     0.4118     0.1529        236        640:  22%|██▏       | 180/809 [03:35<12:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4123     0.1526        263        640:  22%|██▏       | 182/809 [03:37<12:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4124     0.1523        166        640:  23%|██▎       | 183/809 [03:38<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.4126     0.1519        280        640:  23%|██▎       | 185/809 [03:41<12:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4123     0.1518        236        640:  23%|██▎       | 187/809 [03:43<12:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4121     0.1518        276        640:  23%|██▎       | 188/809 [03:44<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4122     0.1522        571        640:  23%|██▎       | 190/809 [03:47<12:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4128     0.1521        145        640:  24%|██▎       | 192/809 [03:49<12:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4128     0.1521        334        640:  24%|██▍       | 193/809 [03:50<12:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4133     0.1522        414        640:  24%|██▍       | 195/809 [03:53<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.252     0.4135     0.1519        216        640:  24%|██▍       | 197/809 [03:55<11:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.252     0.4137     0.1519        268        640:  25%|██▍       | 199/809 [03:57<11:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.252     0.4138     0.1516        255        640:  25%|██▍       | 200/809 [03:58<12:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.251      0.414     0.1514        137        640:  25%|██▍       | 202/809 [04:01<11:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.251      0.414     0.1511        163        640:  25%|██▌       | 203/809 [04:02<11:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.252     0.4136      0.151        196        640:  25%|██▌       | 205/809 [04:05<12:36,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4129     0.1511        333        640:  26%|██▌       | 207/809 [04:07<12:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.4126     0.1516        343        640:  26%|██▌       | 208/809 [04:08<12:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4129     0.1512        116        640:  26%|██▌       | 210/809 [04:11<12:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.253     0.4131     0.1509        304        640:  26%|██▌       | 212/809 [04:13<11:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.251     0.4134     0.1508        280        640:  26%|██▋       | 213/809 [04:14<11:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.253      0.413     0.1514        333        640:  27%|██▋       | 215/809 [04:17<11:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.253     0.4131     0.1512        232        640:  27%|██▋       | 217/809 [04:19<11:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.4124      0.152        489        640:  27%|██▋       | 218/809 [04:20<11:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256      0.412     0.1521        387        640:  27%|██▋       | 220/809 [04:23<11:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.258     0.4109     0.1527        197        640:  27%|██▋       | 222/809 [04:25<11:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.258     0.4106     0.1526        249        640:  28%|██▊       | 223/809 [04:26<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.259     0.4106     0.1532        253        640:  28%|██▊       | 225/809 [04:29<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.259     0.4108     0.1531        244        640:  28%|██▊       | 227/809 [04:31<11:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.258     0.4112     0.1532        339        640:  28%|██��       | 229/809 [04:33<11:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.258     0.4113     0.1531        250        640:  28%|██▊       | 230/809 [04:35<11:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4122     0.1527        180        640:  29%|██▊       | 232/809 [04:37<11:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.256     0.4122     0.1528        345        640:  29%|██▉       | 234/809 [04:39<11:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.4129     0.1525        123        640:  29%|██▉       | 235/809 [04:40<11:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.255     0.4129     0.1526        316        640:  29%|██▉       | 237/809 [04:43<11:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4134     0.1525        199        640:  30%|██▉       | 239/809 [04:45<11:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4133     0.1525        213        640:  30%|██▉       | 240/809 [04:46<11:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.254     0.4135     0.1526         84        640:  30%|██▉       | 242/809 [04:49<11:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.251     0.4142     0.1523        227        640:  30%|███       | 244/809 [04:51<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.25     0.4148     0.1518        300        640:  30%|███       | 246/809 [04:53<10:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.251     0.4145      0.152        417        640:  31%|███       | 247/809 [04:55<11:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.249     0.4152     0.1517        278        640:  31%|███       | 249/809 [04:57<11:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.248     0.4157     0.1517        277        640:  31%|███       | 251/809 [04:59<10:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.247      0.416     0.1514        186        640:  31%|███       | 252/809 [05:00<10:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.247     0.4157     0.1513        309        640:  31%|███▏      | 254/809 [05:03<10:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.245     0.4161     0.1509        316        640:  32%|███▏      | 256/809 [05:05<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.244     0.4168     0.1508        142        640:  32%|███▏      | 257/809 [05:07<11:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.242     0.4176     0.1504        263        640:  32%|███▏      | 259/809 [05:09<10:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.242     0.4174     0.1503        573        640:  32%|███▏      | 261/809 [05:11<10:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.241     0.4178     0.1501        283        640:  32%|███▏      | 262/809 [05:12<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.24     0.4182     0.1499         96        640:  33%|███▎      | 264/809 [05:15<10:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237      0.419     0.1495        234        640:  33%|███▎      | 266/809 [05:17<10:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.235       0.42     0.1492        107        640:  33%|███▎      | 268/809 [05:20<10:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.236     0.4196     0.1496        617        640:  33%|███▎      | 269/809 [05:21<10:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4203     0.1491        152        640:  33%|███▎      | 271/809 [05:23<10:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237     0.4196     0.1502        487        640:  34%|███▎      | 272/809 [05:25<11:03,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4206     0.1497        229        640:  34%|███▍      | 274/809 [05:27<10:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.232     0.4213     0.1494        222        640:  34%|███▍      | 276/809 [05:29<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.231     0.4218     0.1492        189        640:  34%|███▍      | 277/809 [05:30<10:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.233     0.4212     0.1501        384        640:  34%|███▍      | 279/809 [05:33<10:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4209     0.1502        280        640:  35%|███▍      | 281/809 [05:35<10:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4211     0.1502        172        640:  35%|███▍      | 283/809 [05:38<10:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.232     0.4218     0.1499        244        640:  35%|███▌      | 285/809 [05:40<10:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.231     0.4221     0.1498        189        640:  35%|███▌      | 286/809 [05:41<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.232     0.4219     0.1501        305        640:  36%|███▌      | 288/809 [05:44<10:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4212     0.1503        529        640:  36%|███▌      | 290/809 [05:46<10:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.235     0.4209     0.1503        243        640:  36%|███▌      | 291/809 [05:47<10:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.236     0.4204     0.1502        198        640:  36%|███▌      | 293/809 [05:50<10:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237     0.4203     0.1501        173        640:  36%|███▋      | 295/809 [05:52<10:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4197     0.1503        598        640:  37%|███▋      | 296/809 [05:53<10:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4197     0.1503        129        640:  37%|███▋      | 298/809 [05:55<10:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.241     0.4193     0.1503        472        640:  37%|███▋      | 300/809 [05:58<10:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.242      0.419     0.1503        240        640:  37%|███▋      | 301/809 [05:59<10:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.242     0.4189       0.15        348        640:  37%|███▋      | 303/809 [06:01<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.24     0.4196     0.1496        234        640:  38%|███▊      | 305/809 [06:04<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4199     0.1493        362        640:  38%|███▊      | 307/809 [06:06<09:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239       0.42     0.1491        225        640:  38%|███▊      | 308/809 [06:07<09:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4199     0.1489        286        640:  38%|███▊      | 310/809 [06:10<09:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.24     0.4197     0.1493        465        640:  38%|███▊      | 311/809 [06:11<10:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4199      0.149        218        640:  39%|███▊      | 313/809 [06:13<10:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.241     0.4195     0.1501        208        640:  39%|███▉      | 315/809 [06:16<10:05,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.242     0.4193     0.1501        315        640:  39%|███▉      | 316/809 [06:17<09:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.241     0.4195     0.1497        193        640:  39%|███▉      | 318/809 [06:19<09:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.241     0.4196     0.1495        199        640:  40%|███▉      | 320/809 [06:22<09:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.239     0.4204     0.1492        157        640:  40%|███▉      | 322/809 [06:24<09:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.238     0.4211      0.149         74        640:  40%|███▉      | 323/809 [06:25<09:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237     0.4213     0.1489        164        640:  40%|████      | 325/809 [06:28<09:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.236     0.4216     0.1487        272        640:  40%|████      | 327/809 [06:30<09:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237     0.4212     0.1487        313        640:  41%|████      | 328/809 [06:31<09:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.237     0.4214     0.1488        163        640:  41%|████      | 330/809 [06:34<09:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.235     0.4217     0.1486        246        640:  41%|████      | 332/809 [06:36<09:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4221     0.1484         66        640:  41%|████      | 333/809 [06:37<09:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.234     0.4221     0.1484        299        640:  41%|████▏     | 335/809 [06:40<09:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.232     0.4225     0.1481         66        640:  42%|████▏     | 337/809 [06:42<09:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.231     0.4229     0.1479        317        640:  42%|████▏     | 338/809 [06:43<09:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.227     0.4238     0.1474        104        640:  42%|████▏     | 340/809 [06:46<09:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.226     0.4241     0.1472        307        640:  42%|████▏     | 342/809 [06:48<09:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.225     0.4246      0.147        105        640:  42%|████▏     | 343/809 [06:49<09:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.224     0.4249     0.1468        148        640:  43%|████▎     | 345/809 [06:52<09:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222     0.4256     0.1466        138        640:  43%|████▎     | 347/809 [06:54<09:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222     0.4258     0.1465        236        640:  43%|████▎     | 349/809 [06:56<08:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222     0.4259     0.1463        197        640:  43%|█���██▎     | 350/809 [06:57<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4263     0.1462        141        640:  44%|████▎     | 352/809 [07:00<09:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4264      0.146        262        640:  44%|████▍     | 354/809 [07:02<08:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.219     0.4268     0.1459        168        640:  44%|████▍     | 355/809 [07:03<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.218     0.4269     0.1457        298        640:  44%|████▍     | 357/809 [07:06<08:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.217     0.4274     0.1455        245        640:  44%|████▍     | 359/809 [07:08<08:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4278     0.1453        244        640:  45%|████▍     | 361/809 [07:10<08:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4279     0.1452        140        640:  45%|████▍     | 362/809 [07:12<08:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.217     0.4277     0.1453        219        640:  45%|████▍     | 364/809 [07:14<08:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4281     0.1451        172        640:  45%|████▌     | 366/809 [07:16<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4282      0.145        184        640:  45%|████▌     | 367/809 [07:17<08:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4284     0.1447        203        640:  46%|████▌     | 369/809 [07:20<08:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4288     0.1445        112        640:  46%|████▌     | 371/809 [07:22<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4286     0.1445        416        640:  46%|████▌     | 372/809 [07:23<08:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4289     0.1445        206        640:  46%|████▌     | 374/809 [07:26<08:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4291     0.1444        250        640:  46%|████▋     | 376/809 [07:28<08:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4291     0.1443        298        640:  47%|████▋     | 377/809 [07:29<08:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4298     0.1441        224        640:  47%|████▋     | 379/809 [07:32<08:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4302      0.144        102        640:  47%|████▋     | 381/809 [07:34<08:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4301     0.1441        505        640:  47%|████▋     | 382/809 [07:35<08:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4302     0.1438        224        640:  47%|████▋     | 384/809 [07:38<08:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4304     0.1438        153        640:  48%|████▊     | 386/809 [07:40<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4304     0.1438        232        640:  48%|████▊     | 388/809 [07:42<08:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4306     0.1437        355        640:  48%|████▊     | 389/809 [07:44<08:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4312     0.1435        237        640:  48%|████▊     | 391/809 [07:46<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4317     0.1432        336        640:  49%|████▊     | 393/809 [07:48<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4314     0.1432        562        640:  49%|████▊     | 394/809 [07:50<08:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4311     0.1438        459        640:  49%|████▉     | 396/809 [07:52<08:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4311     0.1437        347        640:  49%|████▉     | 398/809 [07:54<08:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4307     0.1446        325        640:  49%|████▉     | 399/809 [07:56<08:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4308     0.1444        243        640:  50%|████▉     | 401/809 [07:58<08:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4306     0.1445        129        640:  50%|████▉     | 403/809 [08:00<08:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4307     0.1444        178        640:  50%|████▉     | 404/809 [08:02<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1444        229        640:  50%|█████     | 406/809 [08:04<07:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4308     0.1441        183        640:  50%|█████     | 408/809 [08:06<07:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4308     0.1441        301        640:  51%|█████     | 409/809 [08:08<07:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4311      0.144        294        640:  51%|█████     | 411/809 [08:10<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4314     0.1439        214        640:  51%|█████     | 413/809 [08:12<07:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4315     0.1438        252        640:  51%|█████     | 414/809 [08:13<07:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4315     0.1439        418        640:  51%|█████▏    | 416/809 [08:16<07:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4317     0.1437        187        640:  52%|█████▏    | 418/809 [08:18<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4319     0.1434        177        640:  52%|█████▏    | 420/809 [08:21<07:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4321     0.1434        247        640:  52%|█████▏    | 421/809 [08:22<07:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4319     0.1434        287        640:  52%|█████▏    | 423/809 [08:24<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4322     0.1432        135        640:  53%|█████▎    | 425/809 [08:27<07:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4319     0.1437        287        640:  53%|█████▎    | 427/809 [08:29<07:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4322     0.1436        174        640:  53%|█████▎    | 428/809 [08:30<07:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4323     0.1435        335        640:  53%|█████▎    | 430/809 [08:32<07:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4323     0.1434        162        640:  53%|█████▎    | 432/809 [08:35<07:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4321     0.1434        327        640:  54%|█████▎    | 433/809 [08:36<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4324     0.1432        133        640:  54%|█████▍    | 435/809 [08:38<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4328      0.143        200        640:  54%|█████▍    | 437/809 [08:41<07:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4326     0.1432        530        640:  54%|█████▍    | 438/809 [08:42<07:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4328     0.1431        228        640:  54%|█████▍    | 440/809 [08:44<07:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4325     0.1433        295        640:  55%|█████▍    | 442/809 [08:47<07:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4323     0.1434        297        640:  55%|█████▍    | 443/809 [08:48<07:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4323     0.1438        121        640:  55%|█████▌    | 445/809 [08:50<07:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4318      0.144        193        640:  55%|█████▌    | 447/809 [08:53<07:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4317      0.144        289        640:  55%|█████▌    | 448/809 [08:54<07:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4314     0.1441        206        640:  56%|█████▌    | 450/809 [08:56<07:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211      0.431     0.1443        393        640:  56%|█████▌    | 452/809 [08:59<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4311     0.1443        149        640:  56%|█████▌    | 454/809 [09:01<06:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4309     0.1444        384        640:  56%|█████▌    | 455/809 [09:02<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4304     0.1445        275        640:  56%|█████▋    | 457/809 [09:05<07:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214       0.43     0.1444        336        640:  57%|█████▋    | 459/809 [09:07<06:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4299     0.1444        259        640:  57%|█████▋    | 460/809 [09:08<06:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4297     0.1443        250        640:  57%|█████▋    | 462/809 [09:11<06:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1444        212        640:  57%|█████▋    | 464/809 [09:13<06:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1444        340        640:  57%|█████▋    | 465/809 [09:14<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4292     0.1444        359        640:  58%|█████▊    | 467/809 [09:17<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216      0.429     0.1446        370        640:  58%|█████▊    | 469/809 [09:19<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4293     0.1446        235        640:  58%|█████▊    | 470/809 [09:20<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4292     0.1446        374        640:  58%|█████▊    | 472/809 [09:23<06:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.217     0.4288     0.1447        255        640:  59%|█████▊    | 474/809 [09:25<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.218     0.4285     0.1449        345        640:  59%|█████▊    | 475/809 [09:26<06:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.219      0.428     0.1451        183        640:  59%|█████▉    | 477/809 [09:29<06:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22      0.428     0.1454        353        640:  59%|█████▉    | 479/809 [09:31<06:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4277     0.1454        299        640:  59%|█████▉    | 480/809 [09:32<06:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4274     0.1459        586        640:  60%|█████▉    | 482/809 [09:35<06:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222     0.4272      0.146        437        640:  60%|█████▉    | 484/809 [09:37<06:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4276     0.1459        116        640:  60%|█████▉    | 485/809 [09:38<06:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4278     0.1458        319        640:  60%|██████    | 487/809 [09:41<06:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4275     0.1458        373        640:  60%|██████    | 489/809 [09:43<06:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4276     0.1458        117        640:  61%|██████    | 490/809 [09:44<06:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4272     0.1459        193        640:  61%|██████    | 492/809 [09:47<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.223     0.4268      0.146        306        640:  61%|██████    | 494/809 [09:49<06:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.224     0.4264     0.1461        295        640:  61%|██████    | 495/809 [09:50<06:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.223     0.4268      0.146        203        640:  61%|██████▏   | 497/809 [09:53<06:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.223     0.4267     0.1459        287        640:  62%|██████▏   | 499/809 [09:55<06:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222      0.427     0.1459        210        640:  62%|██████▏   | 501/809 [09:57<06:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.222     0.4271     0.1458        219        640:  62%|██████▏   | 502/809 [09:58<05:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4273     0.1456        179        640:  62%|██████▏   | 504/809 [10:01<05:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4273     0.1455        307        640:  63%|██████▎   | 506/809 [10:03<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4274     0.1454        124        640:  63%|██████▎   | 507/809 [10:04<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4275     0.1452        253        640:  63%|██████▎   | 509/809 [10:07<05:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4275     0.1458        115        640:  63%|██████▎   | 511/809 [10:09<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4276     0.1457        180        640:  63%|██████▎   | 512/809 [10:10<05:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4276     0.1457        266        640:  64%|██████▎   | 514/809 [10:13<05:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4275     0.1457        335        640:  64%|██████▍   | 516/809 [10:15<05:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4276     0.1457        180        640:  64%|██████▍   | 518/809 [10:17<05:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.221     0.4275     0.1457        456        640:  64%|██████▍   | 519/809 [10:19<05:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.22     0.4278     0.1456        338        640:  64%|██████▍   | 521/809 [10:21<05:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.219     0.4281     0.1455        137        640:  65%|██████▍   | 523/809 [10:23<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.219     0.4282     0.1455        243        640:  65%|██████▍   | 524/809 [10:25<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.218     0.4284     0.1456        400        640:  65%|██████▌   | 526/809 [10:27<05:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.218     0.4286     0.1457        277        640:  65%|██████▌   | 528/809 [10:29<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.217      0.429     0.1455        114        640:  65%|██████▌   | 529/809 [10:30<05:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4294     0.1455        223        640:  66%|██████▌   | 531/809 [10:33<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4298     0.1453        105        640:  66%|██████▌   | 533/809 [10:35<05:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4298     0.1452        204        640:  66%|██████▌   | 535/809 [10:37<05:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214       0.43     0.1452        217        640:  66%|██████▋   | 536/809 [10:39<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4301      0.145        237        640:  67%|██████▋   | 538/809 [10:41<05:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4302      0.145        164        640:  67%|██████▋   | 539/809 [10:42<05:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4303     0.1449        238        640:  67%|██████▋   | 541/809 [10:45<05:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4303     0.1448        169        640:  67%|██████▋   | 543/809 [10:47<05:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4304     0.1448        184        640:  67%|██████▋   | 544/809 [10:48<05:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4302     0.1449        428        640:  67%|██████▋   | 546/809 [10:51<05:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4303     0.1448        239        640:  68%|██████▊   | 548/809 [10:53<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4306     0.1448        249        640:  68%|██████▊   | 550/809 [10:55<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4305     0.1448        312        640:  68%|██████▊   | 551/809 [10:57<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4304     0.1448        294        640:  68%|██████▊   | 553/809 [10:59<05:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4306     0.1446        137        640:  69%|██████▊   | 555/809 [11:01<04:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4308     0.1446        282        640:  69%|██████▊   | 556/809 [11:03<04:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4308     0.1446        376        640:  69%|██████▉   | 558/809 [11:05<04:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4311     0.1444        153        640:  69%|██████▉   | 560/809 [11:07<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4312     0.1444        282        640:  69%|██████▉   | 561/809 [11:08<04:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4314     0.1444        127        640:  70%|██████▉   | 563/809 [11:11<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21      0.431     0.1447        339        640:  70%|██████▉   | 565/809 [11:13<04:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4312     0.1446        153        640:  70%|███████   | 567/809 [11:15<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4314     0.1446        155        640:  70%|███████   | 568/809 [11:17<04:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4313     0.1446        317        640:  70%|███████   | 570/809 [11:19<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4313     0.1446        240        640:  71%|███████   | 572/809 [11:21<04:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4315     0.1445        267        640:  71%|███████   | 573/809 [11:23<04:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4314     0.1445        188        640:  71%|███████   | 575/809 [11:25<04:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4315     0.1444        159        640:  71%|███████▏  | 577/809 [11:27<04:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4315     0.1444        167        640:  71%|███████▏  | 578/809 [11:29<04:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4317     0.1442        373        640:  72%|███████▏  | 580/809 [11:31<04:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4319     0.1441         72        640:  72%|███████▏  | 582/809 [11:33<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4321      0.144        280        640:  72%|███████▏  | 583/809 [11:34<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206      0.432     0.1439        302        640:  72%|███████▏  | 585/809 [11:37<04:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4321     0.1439         95        640:  73%|███████▎  | 587/809 [11:39<04:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206      0.432     0.1439        304        640:  73%|███████▎  | 589/809 [11:42<04:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4318     0.1438        238        640:  73%|███████▎  | 590/809 [11:43<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4316     0.1437        290        640:  73%|███████▎  | 592/809 [11:45<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4318     0.1436        346        640:  73%|███████▎  | 594/809 [11:47<04:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205      0.432     0.1435        332        640:  74%|███████▎  | 596/809 [11:50<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4319     0.1439        258        640:  74%|███████▍  | 597/809 [11:51<04:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4317     0.1441        138        640:  74%|███████▍  | 599/809 [11:53<04:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4318     0.1439        233        640:  74%|███████▍  | 601/809 [11:56<04:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4319     0.1439        391        640:  75%|███████▍  | 603/809 [11:58<04:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4318     0.1439        263        640:  75%|███████▍  | 604/809 [11:59<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4316     0.1439        241        640:  75%|███████▍  | 606/809 [12:02<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4315     0.1439        382        640:  75%|███████▌  | 607/809 [12:03<04:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4311     0.1442        517        640:  75%|███████▌  | 609/809 [12:06<04:08,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208      0.431     0.1443        193        640:  76%|███████▌  | 611/809 [12:08<03:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4309     0.1444        350        640:  76%|███████▌  | 612/809 [12:09<03:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4309     0.1445        184        640:  76%|███████▌  | 614/809 [12:12<03:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4312     0.1444        104        640:  76%|███████▌  | 616/809 [12:14<03:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4311     0.1445        227        640:  76%|███████▋  | 617/809 [12:15<03:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4307     0.1446        315        640:  77%|███████▋  | 619/809 [12:18<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4307     0.1448        553        640:  77%|███████▋  | 621/809 [12:20<03:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4307     0.1448        219        640:  77%|███████▋  | 622/809 [12:21<03:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208      0.431     0.1447        132        640:  77%|███████▋  | 624/809 [12:24<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4312     0.1445        238        640:  77%|███████▋  | 626/809 [12:26<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4313     0.1444        150        640:  78%|███████▊  | 628/809 [12:28<03:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4313     0.1444        312        640:  78%|███████▊  | 629/809 [12:29<03:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4314     0.1445        134        640:  78%|███████▊  | 631/809 [12:32<03:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4315     0.1444        228        640:  78%|███████▊  | 632/809 [12:33<03:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4315     0.1445         39        640:  78%|███████▊  | 634/809 [12:36<03:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4315     0.1444        333        640:  79%|███████▊  | 636/809 [12:38<03:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4314     0.1445        186        640:  79%|███████▉  | 638/809 [12:40<03:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4312     0.1446        279        640:  79%|███████▉  | 639/809 [12:41<03:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4312     0.1445        318        640:  79%|███████▉  | 641/809 [12:44<03:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208      0.431     0.1446        328        640:  79%|███████▉  | 643/809 [12:46<03:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4308     0.1447        267        640:  80%|███████▉  | 644/809 [12:47<03:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4305     0.1446        274        640:  80%|███████▉  | 646/809 [12:50<03:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4304     0.1446        231        640:  80%|████████  | 648/809 [12:52<03:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4305     0.1445        143        640:  80%|████████  | 649/809 [12:53<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4299     0.1451        302        640:  80%|████████  | 651/809 [12:56<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4296     0.1451        200        640:  81%|████████  | 653/809 [12:58<03:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4293     0.1452        252        640:  81%|████████  | 654/809 [12:59<03:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4288     0.1452        247        640:  81%|████████  | 656/809 [13:02<03:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215      0.429     0.1451        163        640:  81%|████████▏ | 658/809 [13:04<03:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4289      0.145        352        640:  81%|████████▏ | 659/809 [13:05<02:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4288     0.1459        130        640:  82%|████████▏ | 661/809 [13:08<03:01,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4291     0.1457        113        640:  82%|████████▏ | 663/809 [13:10<02:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4291     0.1457        317        640:  82%|████████▏ | 664/809 [13:11<02:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4289     0.1457        226        640:  82%|████████▏ | 666/809 [13:14<02:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4289     0.1456        172        640:  83%|████████▎ | 668/809 [13:16<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4289     0.1456        219        640:  83%|████████▎ | 669/809 [13:17<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1455        287        640:  83%|████████▎ | 671/809 [13:20<02:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1455        282        640:  83%|████████▎ | 673/809 [13:22<02:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1454        288        640:  83%|████████▎ | 674/809 [13:23<02:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216      0.429     0.1454        569        640:  84%|████████▎ | 676/809 [13:26<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.216     0.4291     0.1452        135        640:  84%|████████▍ | 678/809 [13:28<02:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4293     0.1452        163        640:  84%|████████▍ | 679/809 [13:29<02:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4294      0.145        226        640:  84%|████████▍ | 681/809 [13:32<02:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4297     0.1449        330        640:  84%|████████▍ | 683/809 [13:34<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4297     0.1449        357        640:  85%|████████▍ | 685/809 [13:36<02:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.215     0.4297     0.1449        217        640:  85%|████████▍ | 686/809 [13:38<02:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.214     0.4299     0.1449        229        640:  85%|████████▌ | 688/809 [13:40<02:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4304     0.1448        118        640:  85%|████████▌ | 690/809 [13:42<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1447        216        640:  85%|████████▌ | 691/809 [13:43<02:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4303     0.1447        241        640:  86%|████████▌ | 693/809 [13:46<02:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1446        197        640:  86%|████████▌ | 695/809 [13:48<02:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1446        176        640:  86%|████████▌ | 696/809 [13:49<02:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1446        327        640:  86%|████████▋ | 698/809 [13:52<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4303     0.1445        249        640:  87%|████████▋ | 700/809 [13:54<02:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4305     0.1445         87        640:  87%|████████▋ | 701/809 [13:55<02:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4304     0.1445        272        640:  87%|████████▋ | 703/809 [13:58<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4306     0.1444        190        640:  87%|████████▋ | 705/809 [14:00<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4308     0.1443        130        640:  87%|████████▋ | 707/809 [14:02<01:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4306     0.1446        224        640:  88%|████████▊ | 708/809 [14:04<01:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4308     0.1446        135        640:  88%|████████▊ | 710/809 [14:06<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4307     0.1446        339        640:  88%|████████▊ | 712/809 [14:08<01:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4308     0.1446        242        640:  88%|████████▊ | 713/809 [14:10<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4308     0.1445        165        640:  88%|████████▊ | 715/809 [14:12<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4307     0.1446        425        640:  89%|████████▊ | 717/809 [14:14<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212      0.431     0.1445        258        640:  89%|████████▉ | 718/809 [14:17<01:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4311     0.1444        132        640:  89%|████████▉ | 720/809 [14:18<01:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4311     0.1445        241        640:  89%|████████▉ | 722/809 [14:20<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4309     0.1445        375        640:  89%|████████▉ | 724/809 [14:23<01:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4309     0.1444         70        640:  90%|████████▉ | 725/809 [14:24<01:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.213     0.4309     0.1447        124        640:  90%|████████▉ | 727/809 [14:26<01:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4312     0.1446        220        640:  90%|█████████ | 729/809 [14:28<01:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4312     0.1446        285        640:  90%|█████████ | 730/809 [14:30<01:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4312     0.1445        403        640:  90%|█████████ | 732/809 [14:32<01:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4312     0.1445        186        640:  91%|█████████ | 734/809 [14:34<01:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4312     0.1445        382        640:  91%|█████████ | 735/809 [14:36<01:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.212     0.4314     0.1445        176        640:  91%|█████████ | 737/809 [14:38<01:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4314     0.1444        181        640:  91%|█████████▏| 739/809 [14:40<01:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4317     0.1444        151        640:  91%|█████████▏| 740/809 [14:42<01:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.211     0.4318     0.1443        235        640:  92%|█████████▏| 742/809 [14:44<01:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G       1.21     0.4321     0.1442        103        640:  92%|█████████▏| 744/809 [14:46<01:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4323     0.1441        156        640:  92%|█████████▏| 745/809 [14:48<01:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4324     0.1442        365        640:  92%|█████████▏| 747/809 [14:50<01:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.209     0.4325     0.1443        109        640:  93%|█████████▎| 749/809 [14:52<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4327     0.1442        304        640:  93%|█████████▎| 751/809 [14:55<01:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4326     0.1442        497        640:  93%|█████████▎| 752/809 [14:56<01:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4326     0.1442        226        640:  93%|█████████▎| 754/809 [14:58<01:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4329      0.144        332        640:  93%|█████████▎| 756/809 [15:01<01:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207      0.433      0.144        282        640:  94%|█████████▎| 758/809 [15:03<00:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207      0.433      0.144        198        640:  94%|█████████▍| 759/809 [15:04<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4331     0.1439        182        640:  94%|█████████▍| 761/809 [15:06<00:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207      0.433     0.1438        170        640:  94%|█████████▍| 763/809 [15:09<00:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207      0.433     0.1438        332        640:  94%|█████████▍| 764/809 [15:10<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4328     0.1437        275        640:  95%|█████████▍| 766/809 [15:13<00:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1437        223        640:  95%|█████████▍| 768/809 [15:15<00:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1437        246        640:  95%|█████████▌| 769/809 [15:16<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4328     0.1441        335        640:  95%|█████████▌| 771/809 [15:19<00:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1441        292        640:  96%|█████████▌| 773/809 [15:21<00:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4328     0.1441        314        640:  96%|█████████▌| 774/809 [15:22<00:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4326     0.1443        280        640:  96%|█████████▌| 776/809 [15:25<00:40,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4327     0.1442        236        640:  96%|█████████▌| 778/809 [15:27<00:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1441        100        640:  96%|█████████▋| 779/809 [15:28<00:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4332      0.144        129        640:  97%|█████████▋| 781/809 [15:31<00:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4335     0.1439         58        640:  97%|█████████▋| 783/809 [15:33<00:32,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.205     0.4335     0.1439        216        640:  97%|█████████▋| 784/809 [15:34<00:30,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4335     0.1438        113        640:  97%|█████████▋| 786/809 [15:37<00:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4333     0.1438        205        640:  97%|█████████▋| 788/809 [15:39<00:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4333     0.1438        209        640:  98%|█████████▊| 790/809 [15:41<00:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4331     0.1438        454        640:  98%|█████████▊| 791/809 [15:43<00:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1439        264        640:  98%|█████████▊| 793/809 [15:45<00:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4328     0.1439        261        640:  98%|█████████▊| 795/809 [15:47<00:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207     0.4329     0.1439        213        640:  98%|█████████▊| 796/809 [15:49<00:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4331     0.1438        304        640:  99%|█████████▊| 798/809 [15:51<00:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4334     0.1436         94        640:  99%|█████████▉| 800/809 [15:53<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.206     0.4334     0.1436        163        640:  99%|█████████▉| 801/809 [15:54<00:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.207      0.433     0.1437        344        640:  99%|█████████▉| 803/809 [15:57<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4327     0.1438        497        640: 100%|█████████▉| 805/809 [15:59<00:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4327     0.1439        301        640: 100%|█████████▉| 806/809 [16:00<00:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     34/300      13.5G      1.208     0.4325     0.1439        145        640: 100%|█████████▉| 808/809 [16:03<00:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.04it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:09<00:08,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:11<00:06,  3.64it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:04,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300      13.9G      1.436     0.3708     0.1825        318        640:   0%|          | 2/809 [00:02<16:34,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300      13.9G      1.223     0.4322     0.1467        461        640:   0%|          | 4/809 [00:04<16:13,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300      13.9G      1.169     0.4355      0.134        207        640:   1%|          | 5/809 [00:06<16:03,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.215     0.4342     0.1702        181        640:   1%|          | 7/809 [00:08<16:23,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.185     0.4438     0.1592        292        640:   1%|          | 8/809 [00:09<16:45,  1.26s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.208     0.4396     0.1529        276        640:   1%|          | 10/809 [00:12<16:13,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4507     0.1489        410        640:   1%|▏         | 12/809 [00:14<16:03,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.188     0.4497     0.1494        348        640:   2%|▏         | 14/809 [00:16<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.153     0.4564     0.1447         87        640:   2%|▏         | 15/809 [00:18<15:56,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.204       0.44     0.1559        268        640:   2%|▏         | 17/809 [00:20<15:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4288     0.1533        396        640:   2%|▏         | 19/809 [00:23<15:55,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.243     0.4247     0.1521        187        640:   2%|▏         | 20/809 [00:24<16:19,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.22     0.4332     0.1474        184        640:   3%|▎         | 22/809 [00:26<15:39,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.215     0.4338     0.1454        296        640:   3%|▎         | 24/809 [00:29<15:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.213     0.4351      0.145        158        640:   3%|▎         | 25/809 [00:30<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.215     0.4348     0.1429        130        640:   3%|▎         | 27/809 [00:32<15:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.208     0.4347     0.1409        126        640:   4%|▎         | 29/809 [00:34<15:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.207     0.4343     0.1396        182        640:   4%|▎         | 30/809 [00:36<15:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.196     0.4383     0.1374        110        640:   4%|▍         | 32/809 [00:38<15:09,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4447     0.1346        110        640:   4%|▍         | 34/809 [00:40<14:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.171     0.4476     0.1326        199        640:   4%|▍         | 36/809 [00:42<14:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.169     0.4482     0.1328        305        640:   5%|▍         | 37/809 [00:44<15:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.162     0.4513     0.1351        399        640:   5%|▍         | 39/809 [00:46<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4463     0.1381        125        640:   5%|▌         | 41/809 [00:48<15:07,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.197     0.4415     0.1414        414        640:   5%|▌         | 42/809 [00:50<15:13,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.216     0.4367     0.1441        217        640:   5%|▌         | 44/809 [00:52<15:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4354     0.1462        158        640:   6%|▌         | 46/809 [00:54<15:20,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4342     0.1457        176        640:   6%|▌         | 47/809 [00:56<15:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.214     0.4387     0.1438        154        640:   6%|▌         | 49/809 [00:58<15:08,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.217     0.4373     0.1452        200        640:   6%|▋         | 51/809 [01:00<15:13,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.217     0.4377     0.1456        249        640:   6%|▋         | 52/809 [01:02<15:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.208       0.44     0.1437        196        640:   7%|▋         | 54/809 [01:04<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.199     0.4417     0.1418        228        640:   7%|▋         | 56/809 [01:06<14:39,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.198     0.4422     0.1419        224        640:   7%|▋         | 58/809 [01:09<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.191     0.4434     0.1409        287        640:   7%|▋         | 59/809 [01:10<14:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.195     0.4413     0.1401        181        640:   8%|▊         | 61/809 [01:12<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4447     0.1389         83        640:   8%|▊         | 63/809 [01:15<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.186     0.4439     0.1387        243        640:   8%|▊         | 64/809 [01:16<14:33,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.19     0.4426     0.1388        388        640:   8%|▊         | 66/809 [01:18<14:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4454      0.137        186        640:   8%|▊         | 68/809 [01:20<14:30,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.177     0.4465     0.1367        264        640:   9%|▊         | 69/809 [01:22<14:27,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172     0.4483     0.1366        116        640:   9%|▉         | 71/809 [01:24<14:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172      0.449     0.1362        203        640:   9%|▉         | 73/809 [01:26<14:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4475     0.1352        216        640:   9%|▉         | 75/809 [01:29<14:16,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.173     0.4479     0.1349        150        640:   9%|▉         | 76/809 [01:30<14:12,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.177     0.4462     0.1347        274        640:  10%|▉         | 78/809 [01:32<14:17,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.175     0.4465     0.1339         83        640:  10%|▉         | 80/809 [01:35<14:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.169     0.4479     0.1332        195        640:  10%|█         | 81/809 [01:36<14:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.166     0.4481      0.133        211        640:  10%|█         | 83/809 [01:38<14:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.161     0.4489     0.1322        120        640:  11%|█         | 85/809 [01:40<14:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.158     0.4491     0.1316        223        640:  11%|█         | 86/809 [01:42<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.159     0.4484     0.1314        398        640:  11%|█         | 88/809 [01:44<14:23,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.159      0.448     0.1309        365        640:  11%|█         | 90/809 [01:46<14:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.159     0.4485      0.131        294        640:  11%|█▏        | 92/809 [01:49<14:10,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.162      0.448     0.1314        388        640:  11%|█▏        | 93/809 [01:50<14:21,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.165     0.4465     0.1311        205        640:  12%|█▏        | 95/809 [01:52<14:23,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.167     0.4458     0.1305         74        640:  12%|█▏        | 97/809 [01:55<14:12,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.168     0.4452     0.1301        280        640:  12%|█▏        | 98/809 [01:56<14:18,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.17     0.4442     0.1302        218        640:  12%|█▏        | 100/809 [01:58<14:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.169     0.4441     0.1301        284        640:  13%|█▎        | 102/809 [02:01<14:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.166     0.4444     0.1294        252        640:  13%|█▎        | 104/809 [02:03<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.169     0.4439     0.1298        136        640:  13%|█▎        | 105/809 [02:04<13:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172      0.443     0.1297        221        640:  13%|█▎        | 107/809 [02:07<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172     0.4429     0.1291        224        640:  13%|█▎        | 109/809 [02:09<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.173      0.442      0.129        308        640:  14%|█▎        | 110/809 [02:10<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172     0.4421     0.1292        191        640:  14%|█▍        | 112/809 [02:13<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.166     0.4437     0.1285        119        640:  14%|█▍        | 114/809 [02:15<13:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.163     0.4447      0.128        260        640:  14%|█▍        | 116/809 [02:17<13:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.164     0.4445     0.1279        283        640:  14%|█▍        | 117/809 [02:18<13:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.164     0.4445     0.1281        344        640:  15%|█▍        | 119/809 [02:21<13:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.171     0.4421     0.1287        300        640:  15%|█▍        | 121/809 [02:23<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.172     0.4415     0.1288        255        640:  15%|█▌        | 122/809 [02:24<13:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4403     0.1286        182        640:  15%|█▌        | 124/809 [02:27<13:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.176     0.4396     0.1286         92        640:  16%|█▌        | 126/809 [02:29<13:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.182     0.4382     0.1312        383        640:  16%|█▌        | 127/809 [02:30<14:04,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183     0.4386     0.1315        274        640:  16%|█▌        | 129/809 [02:33<13:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.187     0.4377     0.1318        290        640:  16%|█▌        | 131/809 [02:35<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.187      0.438     0.1317        179        640:  16%|█▋        | 133/809 [02:37<13:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.186      0.438     0.1317        156        640:  17%|█▋        | 134/809 [02:38<13:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.187     0.4376     0.1316        257        640:  17%|█▋        | 136/809 [02:41<13:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4391     0.1313        113        640:  17%|█▋        | 138/809 [02:43<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183     0.4392     0.1312        206        640:  17%|█▋        | 139/809 [02:44<12:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179      0.441     0.1308        144        640:  17%|█▋        | 141/809 [02:47<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4403      0.131        249        640:  18%|█▊        | 143/809 [02:49<13:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4402     0.1308        241        640:  18%|█▊        | 145/809 [02:51<13:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4404     0.1307        209        640:  18%|█▊        | 146/809 [02:53<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4405     0.1308        207        640:  18%|█▊        | 148/809 [02:55<12:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4415     0.1306        138        640:  19%|█▊        | 150/809 [02:57<12:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4416     0.1307        321        640:  19%|█▊        | 151/809 [02:59<13:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.176     0.4418     0.1311        109        640:  19%|█▉        | 153/809 [03:01<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.173     0.4425     0.1309        199        640:  19%|█▉        | 155/809 [03:03<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174      0.442      0.131        224        640:  19%|█▉        | 156/809 [03:04<12:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.171     0.4428     0.1306        264        640:  20%|█▉        | 158/809 [03:07<12:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4419     0.1311        451        640:  20%|█▉        | 160/809 [03:09<13:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4415     0.1314        431        640:  20%|█▉        | 161/809 [03:10<13:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4413     0.1312        177        640:  20%|██        | 163/809 [03:13<12:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4396     0.1324        451        640:  20%|██        | 165/809 [03:15<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4393     0.1325        558        640:  21%|██        | 166/809 [03:16<12:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4397     0.1323        149        640:  21%|██        | 168/809 [03:19<12:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178       0.44      0.132        248        640:  21%|██        | 170/809 [03:21<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.177     0.4403     0.1318        205        640:  21%|██        | 171/809 [03:22<12:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.177       0.44     0.1318        350        640:  21%|██▏       | 173/809 [03:25<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.174     0.4409     0.1314        172        640:  22%|██▏       | 175/809 [03:27<12:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.176     0.4401     0.1313        288        640:  22%|██▏       | 177/809 [03:30<12:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4396     0.1317        349        640:  22%|██▏       | 178/809 [03:31<12:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18      0.439     0.1322        262        640:  22%|██▏       | 180/809 [03:33<12:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4388     0.1322        256        640:  22%|██▏       | 182/809 [03:35<12:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4382     0.1321        304        640:  23%|██▎       | 183/809 [03:37<12:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181      0.438     0.1321        285        640:  23%|██▎       | 185/809 [03:39<12:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4384      0.132        309        640:  23%|██▎       | 187/809 [03:41<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4386     0.1317        103        640:  23%|██▎       | 188/809 [03:43<12:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4379     0.1319        357        640:  23%|██▎       | 190/809 [03:45<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.177     0.4391     0.1315        132        640:  24%|██▎       | 192/809 [03:47<12:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4386     0.1325        262        640:  24%|██▍       | 193/809 [03:48<12:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4384     0.1324        350        640:  24%|██▍       | 195/809 [03:51<12:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4381     0.1325        335        640:  24%|██▍       | 197/809 [03:53<12:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4381     0.1325        235        640:  24%|██▍       | 198/809 [03:54<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4383     0.1326        233        640:  25%|██▍       | 200/809 [03:57<12:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181      0.438     0.1327        362        640:  25%|██▍       | 202/809 [03:59<11:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4388     0.1323        192        640:  25%|██▌       | 204/809 [04:01<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4387     0.1323        291        640:  25%|██▌       | 205/809 [04:03<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4388     0.1321        198        640:  26%|██▌       | 207/809 [04:05<12:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183      0.438     0.1338        211        640:  26%|██▌       | 209/809 [04:08<12:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4378     0.1342        511        640:  26%|██▌       | 210/809 [04:09<12:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4381      0.134        308        640:  26%|██▌       | 212/809 [04:11<11:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4379     0.1341        384        640:  26%|██▋       | 214/809 [04:14<11:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4379     0.1341        139        640:  27%|██▋       | 215/809 [04:15<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183     0.4384     0.1337        225        640:  27%|██▋       | 217/809 [04:17<11:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.182     0.4386     0.1338        263        640:  27%|██▋       | 219/809 [04:19<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181      0.439     0.1338        184        640:  27%|██▋       | 220/809 [04:21<11:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.178     0.4398     0.1334        161        640:  27%|██▋       | 222/809 [04:23<11:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.179     0.4395     0.1344        212        640:  28%|██▊       | 224/809 [04:25<11:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4391     0.1346        239        640:  28%|██▊       | 225/809 [04:27<11:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181      0.439     0.1346        276        640:  28%|██▊       | 227/809 [04:29<11:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4388     0.1347        340        640:  28%|██▊       | 229/809 [04:31<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.181     0.4385     0.1345        267        640:  29%|██▊       | 231/809 [04:34<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.18     0.4392     0.1342        184        640:  29%|██▊       | 232/809 [04:35<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.182     0.4388     0.1348        234        640:  29%|██▉       | 234/809 [04:37<11:45,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183     0.4385     0.1351        380        640:  29%|██▉       | 235/809 [04:39<11:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.183     0.4384     0.1351        381        640:  29%|██▉       | 237/809 [04:41<11:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4383     0.1354        212        640:  30%|██▉       | 239/809 [04:43<11:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.184     0.4384     0.1353        303        640:  30%|██▉       | 241/809 [04:46<11:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.185      0.438     0.1356        350        640:  30%|██▉       | 242/809 [04:47<11:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.187     0.4378     0.1358        415        640:  30%|███       | 244/809 [04:49<11:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.185     0.4383     0.1359        358        640:  30%|███       | 246/809 [04:52<11:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.185     0.4384     0.1361        165        640:  31%|███       | 248/809 [04:54<11:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.185     0.4383      0.136        211        640:  31%|███       | 249/809 [04:55<11:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.189     0.4369     0.1367        442        640:  31%|███       | 251/809 [04:58<11:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.191     0.4363     0.1371        225        640:  31%|███▏      | 253/809 [05:00<10:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.192      0.436     0.1375        203        640:  31%|███▏      | 254/809 [05:01<10:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.194     0.4358     0.1381        274        640:  32%|███▏      | 256/809 [05:04<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.195     0.4358     0.1385        527        640:  32%|███▏      | 258/809 [05:06<11:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.193     0.4363     0.1383        152        640:  32%|███▏      | 259/809 [05:07<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.193     0.4366     0.1383        172        640:  32%|███▏      | 261/809 [05:10<10:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.195      0.436     0.1388        350        640:  33%|███▎      | 263/809 [05:12<11:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.198     0.4352     0.1391        676        640:  33%|███▎      | 264/809 [05:13<11:10,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.203     0.4337     0.1396        232        640:  33%|███▎      | 266/809 [05:16<10:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.208     0.4322       0.14        478        640:  33%|███▎      | 268/809 [05:18<10:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.21     0.4316     0.1402        283        640:  33%|███▎      | 269/809 [05:19<10:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.215       0.43     0.1408        376        640:  33%|███▎      | 271/809 [05:22<11:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.219     0.4289     0.1409        274        640:  34%|███▎      | 273/809 [05:24<10:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.218     0.4293     0.1408         95        640:  34%|███▍      | 274/809 [05:25<10:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.219     0.4291      0.141        304        640:  34%|███▍      | 276/809 [05:28<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.22     0.4289      0.141        296        640:  34%|███▍      | 278/809 [05:30<10:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.22      0.429     0.1409        217        640:  34%|███▍      | 279/809 [05:31<10:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.221     0.4287     0.1411        260        640:  35%|███▍      | 281/809 [05:34<10:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4279     0.1414        261        640:  35%|███▍      | 283/809 [05:36<10:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4278     0.1415        258        640:  35%|███▌      | 284/809 [05:37<10:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4272     0.1421        365        640:  35%|███▌      | 286/809 [05:40<10:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228      0.427     0.1424        250        640:  36%|███▌      | 288/809 [05:42<10:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4267     0.1427        409        640:  36%|███▌      | 289/809 [05:43<10:43,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4261     0.1433        362        640:  36%|███▌      | 291/809 [05:46<10:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4261     0.1434        278        640:  36%|███▌      | 293/809 [05:48<10:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4257     0.1436        590        640:  36%|███▋      | 294/809 [05:50<10:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4254     0.1437        328        640:  37%|███▋      | 296/809 [05:52<10:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4251     0.1438        464        640:  37%|███▋      | 297/809 [05:53<10:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.237     0.4246     0.1451        180        640:  37%|███▋      | 299/809 [05:56<10:25,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.239     0.4241     0.1453        334        640:  37%|███▋      | 301/809 [05:58<10:34,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.239      0.424     0.1453        406        640:  37%|███▋      | 302/809 [05:59<10:26,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4239     0.1456        262        640:  38%|███▊      | 304/809 [06:02<10:25,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.239     0.4243     0.1454        399        640:  38%|███▊      | 306/809 [06:04<10:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241      0.424     0.1459        328        640:  38%|███▊      | 307/809 [06:06<10:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.242      0.424     0.1458        172        640:  38%|███▊      | 309/809 [06:08<10:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4245     0.1457        198        640:  38%|███▊      | 311/809 [06:10<09:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241     0.4244     0.1459        385        640:  39%|███▊      | 312/809 [06:12<09:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.243     0.4238     0.1459        218        640:  39%|███▉      | 314/809 [06:14<09:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241     0.4245     0.1456        265        640:  39%|███▉      | 316/809 [06:16<09:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241     0.4245     0.1455        179        640:  39%|███▉      | 317/809 [06:17<09:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4244     0.1453        240        640:  39%|███▉      | 319/809 [06:20<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4247     0.1451        211        640:  40%|███▉      | 321/809 [06:22<09:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4247      0.145        440        640:  40%|███▉      | 322/809 [06:23<09:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.239     0.4251     0.1452        403        640:  40%|████      | 324/809 [06:26<09:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.238     0.4255      0.145        161        640:  40%|████      | 326/809 [06:28<09:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.238     0.4255     0.1451        370        640:  40%|████      | 327/809 [06:29<09:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.238     0.4256     0.1451        440        640:  41%|████      | 329/809 [06:32<09:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.237     0.4258     0.1448        199        640:  41%|████      | 331/809 [06:34<09:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.237     0.4258     0.1445        219        640:  41%|████      | 333/809 [06:36<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.236     0.4258     0.1445        356        640:  41%|████▏     | 334/809 [06:38<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.237     0.4254     0.1452        340        640:  42%|████▏     | 336/809 [06:40<09:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.238     0.4253     0.1451        252        640:  42%|████▏     | 337/809 [06:41<09:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.239     0.4248     0.1455        221        640:  42%|████▏     | 339/809 [06:44<09:36,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24     0.4244     0.1455        295        640:  42%|████▏     | 341/809 [06:46<09:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241     0.4241     0.1456        485        640:  42%|████▏     | 342/809 [06:48<09:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.242     0.4238     0.1463        556        640:  43%|████▎     | 344/809 [06:50<09:38,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241      0.424      0.146        174        640:  43%|████▎     | 346/809 [06:52<09:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.241     0.4238      0.146        304        640:  43%|████▎     | 347/809 [06:54<09:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.24      0.424     0.1457        215        640:  43%|████▎     | 349/809 [06:56<09:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.238     0.4245     0.1454        163        640:  43%|████▎     | 351/809 [06:58<09:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.236     0.4251     0.1451        197        640:  44%|████▎     | 353/809 [07:01<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.236     0.4252      0.145        186        640:  44%|████▍     | 354/809 [07:02<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4255      0.145        249        640:  44%|████▍     | 356/809 [07:04<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4262     0.1447        115        640:  44%|████▍     | 358/809 [07:06<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4261     0.1451        204        640:  44%|████▍     | 359/809 [07:08<08:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1449        175        640:  45%|████▍     | 361/809 [07:10<08:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4263     0.1449        222        640:  45%|████▍     | 363/809 [07:12<08:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1448        145        640:  45%|████▍     | 364/809 [07:13<08:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4269     0.1446        181        640:  45%|████▌     | 366/809 [07:16<08:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1447        236        640:  45%|████▌     | 368/809 [07:18<08:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1446        353        640:  46%|████▌     | 369/809 [07:20<08:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4275     0.1444        198        640:  46%|████▌     | 371/809 [07:22<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4273     0.1446        380        640:  46%|████▌     | 373/809 [07:24<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4273     0.1445        214        640:  46%|████▌     | 374/809 [07:25<08:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227      0.428     0.1442        139        640:  46%|████▋     | 376/809 [07:28<08:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4281     0.1441        137        640:  47%|████▋     | 378/809 [07:30<08:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1439        288        640:  47%|████▋     | 380/809 [07:32<08:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4282     0.1438        145        640:  47%|████▋     | 381/809 [07:34<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4282     0.1441        228        640:  47%|████▋     | 383/809 [07:36<08:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4275     0.1442        320        640:  48%|████▊     | 385/809 [07:38<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4279     0.1441         60        640:  48%|████▊     | 386/809 [07:40<08:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4281     0.1442        127        640:  48%|████▊     | 388/809 [07:42<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4282      0.144        327        640:  48%|████▊     | 390/809 [07:44<08:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4283     0.1439        192        640:  48%|████▊     | 391/809 [07:46<08:38,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4276      0.144        280        640:  49%|████▊     | 393/809 [07:48<08:30,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4274     0.1439        172        640:  49%|████▉     | 395/809 [07:50<08:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4273     0.1438        266        640:  49%|████▉     | 396/809 [07:52<08:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4266     0.1439        182        640:  49%|████▉     | 398/809 [07:54<08:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4266     0.1437        162        640:  49%|████▉     | 400/809 [07:56<08:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1438        131        640:  50%|████▉     | 402/809 [07:59<08:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4271     0.1437         78        640:  50%|████▉     | 403/809 [08:00<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23      0.427     0.1436        177        640:  50%|█████     | 405/809 [08:02<07:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4269     0.1435        186        640:  50%|█████     | 407/809 [08:05<08:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4274     0.1432        146        640:  51%|█████     | 409/809 [08:07<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4272     0.1435        537        640:  51%|█████     | 410/809 [08:08<07:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4269     0.1437        286        640:  51%|█████     | 412/809 [08:11<07:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4265     0.1442        302        640:  51%|█████     | 414/809 [08:13<07:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266      0.144        154        640:  51%|█████▏    | 415/809 [08:14<07:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1441        339        640:  52%|█████▏    | 417/809 [08:17<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4262     0.1451        323        640:  52%|█████▏    | 419/809 [08:19<07:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4261      0.145        274        640:  52%|█████▏    | 420/809 [08:20<07:56,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4263     0.1449        225        640:  52%|█████▏    | 422/809 [08:23<07:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4261     0.1453        606        640:  52%|█████▏    | 423/809 [08:24<08:11,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4261     0.1454        278        640:  53%|█████▎    | 425/809 [08:27<07:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265     0.1451        212        640:  53%|█████▎    | 427/809 [08:29<07:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4264      0.145        245        640:  53%|█████▎    | 429/809 [08:31<07:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232      0.426     0.1451        442        640:  53%|█████▎    | 430/809 [08:33<07:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4265     0.1448        214        640:  53%|█████▎    | 432/809 [08:35<07:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4263     0.1451        588        640:  54%|█████▎    | 433/809 [08:36<07:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4261     0.1451        328        640:  54%|█████▍    | 435/809 [08:39<07:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4259     0.1454        319        640:  54%|█████▍    | 437/809 [08:41<07:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232      0.426     0.1455        366        640:  54%|█████▍    | 438/809 [08:42<07:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4264     0.1452        122        640:  54%|█████▍    | 440/809 [08:45<07:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4267     0.1452        145        640:  55%|█████▍    | 442/809 [08:47<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229      0.427      0.145         85        640:  55%|█████▍    | 444/809 [08:49<07:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229      0.427     0.1449        254        640:  55%|█████▌    | 445/809 [08:50<07:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228      0.427     0.1448        394        640:  55%|█████▌    | 447/809 [08:53<07:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4274     0.1446        248        640:  56%|█████▌    | 449/809 [08:55<07:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4274     0.1445        267        640:  56%|█████▌    | 450/809 [08:56<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4275     0.1443        422        640:  56%|█████▌    | 452/809 [08:59<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4273     0.1445        170        640:  56%|█████▌    | 454/809 [09:01<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4274     0.1444         95        640:  56%|█████▌    | 455/809 [09:02<07:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4278     0.1442        300        640:  56%|█████▋    | 457/809 [09:05<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4276     0.1441        361        640:  57%|█████▋    | 459/809 [09:07<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4275     0.1441        438        640:  57%|█████▋    | 460/809 [09:08<06:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4276     0.1439        120        640:  57%|█████▋    | 462/809 [09:11<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4281     0.1437        174        640:  57%|█████▋    | 464/809 [09:13<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4286     0.1435        112        640:  58%|█████▊    | 466/809 [09:15<06:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4288     0.1434        196        640:  58%|█████▊    | 467/809 [09:17<06:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223     0.4287     0.1438        260        640:  58%|█████▊    | 469/809 [09:19<06:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223     0.4285     0.1439        200        640:  58%|█████▊    | 471/809 [09:21<06:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4283     0.1439        293        640:  58%|█████▊    | 472/809 [09:23<06:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4284     0.1439        223        640:  59%|█████▊    | 474/809 [09:25<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4284     0.1438        318        640:  59%|█████▉    | 476/809 [09:27<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4281     0.1443        637        640:  59%|█████▉    | 477/809 [09:29<06:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4282     0.1442        191        640:  59%|█████▉    | 479/809 [09:31<06:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4278     0.1444        425        640:  59%|█████▉    | 481/809 [09:33<06:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4279     0.1443        197        640:  60%|█████▉    | 482/809 [09:35<06:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4277     0.1448        360        640:  60%|█████▉    | 484/809 [09:37<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4276     0.1447        242        640:  60%|██████    | 486/809 [09:39<06:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4275     0.1451        189        640:  60%|██████    | 487/809 [09:41<06:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4278     0.1449        294        640:  60%|██████    | 489/809 [09:43<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4282     0.1447        248        640:  61%|██████    | 491/809 [09:45<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4285     0.1447        166        640:  61%|██████    | 493/809 [09:48<06:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4282     0.1452        482        640:  61%|██████    | 494/809 [09:49<06:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4284     0.1451        245        640:  61%|██████▏   | 496/809 [09:51<06:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4278     0.1453        364        640:  62%|██████▏   | 498/809 [09:54<06:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4275     0.1453        307        640:  62%|██████▏   | 499/809 [09:55<06:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4274     0.1452        295        640:  62%|██████▏   | 501/809 [09:57<06:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23      0.427     0.1452        178        640:  62%|██████▏   | 503/809 [10:00<06:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231      0.427     0.1453        269        640:  62%|██████▏   | 504/809 [10:01<06:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4272     0.1451        177        640:  63%|██████▎   | 506/809 [10:03<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4272     0.1452        289        640:  63%|██████▎   | 508/809 [10:05<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4273     0.1452        247        640:  63%|██████▎   | 509/809 [10:07<05:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4278      0.145        250        640:  63%|██████▎   | 511/809 [10:09<05:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228      0.428      0.145        247        640:  63%|██████▎   | 513/809 [10:11<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4281     0.1451        104        640:  64%|██████▎   | 515/809 [10:14<05:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4279     0.1452        461        640:  64%|██████▍   | 516/809 [10:15<05:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4279      0.145        388        640:  64%|██████▍   | 518/809 [10:17<05:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4281     0.1448        399        640:  64%|██████▍   | 520/809 [10:20<05:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4282     0.1447        237        640:  64%|██████▍   | 521/809 [10:21<05:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4281     0.1449        427        640:  65%|██████▍   | 523/809 [10:23<05:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4285     0.1447        298        640:  65%|██████▍   | 525/809 [10:26<05:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4285     0.1447        249        640:  65%|██████▌   | 526/809 [10:27<05:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4287     0.1446        294        640:  65%|██████▌   | 528/809 [10:29<05:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4287     0.1447        355        640:  66%|██████▌   | 530/809 [10:32<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4289     0.1446        115        640:  66%|██████▌   | 531/809 [10:33<05:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224      0.429     0.1446        348        640:  66%|██████▌   | 533/809 [10:35<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4291     0.1445        308        640:  66%|██████▌   | 535/809 [10:37<05:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4291     0.1444        292        640:  66%|██████▋   | 536/809 [10:39<05:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4288     0.1444        225        640:  67%|██████▋   | 538/809 [10:41<05:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223     0.4292     0.1442        124        640:  67%|██████▋   | 540/809 [10:43<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223     0.4292     0.1441        296        640:  67%|██████▋   | 542/809 [10:46<05:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4293      0.144        189        640:  67%|██████▋   | 543/809 [10:47<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4294      0.144        153        640:  67%|██████▋   | 545/809 [10:49<05:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4292      0.144        274        640:  68%|██████▊   | 547/809 [10:52<05:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4291      0.144        432        640:  68%|██████▊   | 549/809 [10:54<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223      0.429      0.144        307        640:  68%|██████▊   | 550/809 [10:55<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4288     0.1442        638        640:  68%|██████▊   | 552/809 [10:58<05:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4287     0.1442        122        640:  68%|██████▊   | 554/809 [11:00<05:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223      0.429     0.1441        113        640:  69%|██████▊   | 555/809 [11:01<05:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4292      0.144        183        640:  69%|██████▉   | 557/809 [11:04<04:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4294     0.1439        217        640:  69%|██████▉   | 559/809 [11:06<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.222     0.4293     0.1438        260        640:  69%|██████▉   | 560/809 [11:07<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223      0.429     0.1441        598        640:  69%|██████▉   | 562/809 [11:10<05:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.223     0.4289     0.1442        449        640:  70%|██████▉   | 564/809 [11:12<04:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4287     0.1443        577        640:  70%|██████▉   | 565/809 [11:13<04:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4288     0.1443        262        640:  70%|███████   | 567/809 [11:16<04:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4286     0.1446        419        640:  70%|███████   | 569/809 [11:18<04:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.225     0.4287     0.1446        221        640:  70%|███████   | 570/809 [11:19<04:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224     0.4288     0.1445        381        640:  71%|███████   | 572/809 [11:22<04:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224      0.429     0.1444        202        640:  71%|███████   | 574/809 [11:24<04:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.224      0.429     0.1444        260        640:  71%|███████   | 575/809 [11:25<04:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4284     0.1448        335        640:  71%|███████▏  | 577/809 [11:28<04:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227     0.4281     0.1447        225        640:  72%|███████▏  | 579/809 [11:30<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1446        210        640:  72%|███████▏  | 581/809 [11:32<04:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1446        239        640:  72%|███████▏  | 582/809 [11:34<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4282     0.1445        276        640:  72%|███████▏  | 584/809 [11:36<04:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4281     0.1444        132        640:  72%|███████▏  | 586/809 [11:38<04:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.227      0.428     0.1444        233        640:  73%|███████▎  | 587/809 [11:40<04:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4281     0.1443        133        640:  73%|███████▎  | 589/809 [11:42<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1441        188        640:  73%|███████▎  | 591/809 [11:44<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1441        212        640:  73%|███████▎  | 592/809 [11:45<04:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283      0.144         97        640:  73%|███████▎  | 594/809 [11:48<04:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4284      0.144        207        640:  74%|███████▎  | 596/809 [11:50<04:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4284     0.1441        290        640:  74%|███████▍  | 597/809 [11:51<04:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.226     0.4283     0.1441        187        640:  74%|███████▍  | 599/809 [11:54<04:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4279     0.1442        463        640:  74%|███████▍  | 601/809 [11:56<04:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4275     0.1449        350        640:  74%|███████▍  | 602/809 [11:58<04:16,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4274     0.1449        129        640:  75%|███████▍  | 604/809 [12:00<04:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4272     0.1449        102        640:  75%|███████▍  | 606/809 [12:02<04:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4271     0.1449        234        640:  75%|███████▌  | 607/809 [12:04<04:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4273     0.1448        237        640:  75%|███████▌  | 609/809 [12:06<03:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4273     0.1448        349        640:  76%|███████▌  | 611/809 [12:08<03:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4272     0.1448        445        640:  76%|███████▌  | 612/809 [12:09<03:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4274     0.1447         80        640:  76%|███████▌  | 614/809 [12:12<03:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4274     0.1447        175        640:  76%|███████▌  | 616/809 [12:14<03:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4273     0.1448        379        640:  76%|███████▋  | 617/809 [12:16<03:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4274     0.1447        347        640:  77%|███████▋  | 619/809 [12:18<03:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4274     0.1446        154        640:  77%|███████▋  | 621/809 [12:20<03:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4273     0.1445        246        640:  77%|███████▋  | 622/809 [12:22<03:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4271     0.1455        219        640:  77%|███████▋  | 624/809 [12:24<03:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1457        191        640:  77%|███████▋  | 626/809 [12:26<03:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1458        234        640:  78%|███████▊  | 627/809 [12:28<03:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1458        289        640:  78%|███████▊  | 629/809 [12:30<03:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266     0.1464        307        640:  78%|███████▊  | 631/809 [12:32<03:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1463         61        640:  78%|███████▊  | 632/809 [12:34<03:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1463        240        640:  78%|███████▊  | 634/809 [12:36<03:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1461        206        640:  79%|███████▊  | 636/809 [12:38<03:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1461        228        640:  79%|███████▊  | 637/809 [12:40<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1461        199        640:  79%|███████▉  | 639/809 [12:42<03:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1462        410        640:  79%|███████▉  | 641/809 [12:44<03:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1462        381        640:  79%|███████▉  | 642/809 [12:46<03:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267      0.146        276        640:  80%|███████▉  | 644/809 [12:48<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1463        301        640:  80%|███████▉  | 646/809 [12:50<03:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1462        254        640:  80%|███████▉  | 647/809 [12:52<03:20,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4263     0.1461        249        640:  80%|████████  | 649/809 [12:54<03:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1461        187        640:  80%|████████  | 651/809 [12:56<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4263     0.1461        289        640:  81%|████████  | 652/809 [12:58<03:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4263      0.146        246        640:  81%|████████  | 654/809 [13:00<03:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4264     0.1459        387        640:  81%|████████  | 656/809 [13:02<03:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265     0.1459        164        640:  81%|████████  | 657/809 [13:04<03:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23      0.427     0.1457         83        640:  81%|████████▏ | 659/809 [13:06<02:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23      0.427     0.1457        171        640:  82%|████████▏ | 661/809 [13:08<02:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4269     0.1457        207        640:  82%|████████▏ | 662/809 [13:09<02:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4268     0.1457        362        640:  82%|████████▏ | 664/809 [13:12<02:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1456        170        640:  82%|████████▏ | 666/809 [13:14<02:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4271     0.1455        241        640:  83%|████████▎ | 668/809 [13:17<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4269     0.1457        418        640:  83%|████████▎ | 669/809 [13:18<02:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1458        156        640:  83%|████████▎ | 671/809 [13:20<02:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1457        150        640:  83%|████████▎ | 672/809 [13:22<02:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4266     0.1457        119        640:  83%|████████▎ | 674/809 [13:24<02:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1455        147        640:  84%|████████▎ | 676/809 [13:26<02:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4266     0.1456        174        640:  84%|████████▍ | 678/809 [13:29<02:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4265     0.1456        269        640:  84%|████████▍ | 679/809 [13:30<02:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265     0.1457        287        640:  84%|████████▍ | 681/809 [13:32<02:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266     0.1458        311        640:  84%|████████▍ | 682/809 [13:33<02:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265      0.146        345        640:  85%|████████▍ | 684/809 [13:36<02:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266     0.1459        275        640:  85%|████████▍ | 686/809 [13:38<02:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266     0.1461        267        640:  85%|████████▍ | 687/809 [13:40<02:30,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4267     0.1461        157        640:  85%|████████▌ | 689/809 [13:42<02:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266      0.146        201        640:  85%|████████▌ | 691/809 [13:44<02:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265     0.1459         93        640:  86%|████████▌ | 693/809 [13:47<02:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4265     0.1459        358        640:  86%|████████▌ | 694/809 [13:48<02:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4262      0.146        241        640:  86%|████████▌ | 696/809 [13:50<02:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4261     0.1461        234        640:  86%|████████▋ | 698/809 [13:53<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4262      0.146        164        640:  86%|████████▋ | 699/809 [13:54<02:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4264     0.1459        206        640:  87%|████████▋ | 701/809 [13:56<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4262     0.1459        144        640:  87%|████████▋ | 703/809 [13:59<02:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4262     0.1459        391        640:  87%|████████▋ | 704/809 [14:00<02:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4263     0.1458        273        640:  87%|████████▋ | 706/809 [14:02<02:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4267     0.1457        159        640:  88%|████████▊ | 708/809 [14:04<01:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4266     0.1457        168        640:  88%|████████▊ | 710/809 [14:07<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4266     0.1457        239        640:  88%|████████▊ | 712/809 [14:09<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4266     0.1457        362        640:  88%|████████▊ | 713/809 [14:10<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4267     0.1456        196        640:  88%|████████▊ | 715/809 [14:13<01:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4267     0.1455        325        640:  89%|████████▊ | 717/809 [14:15<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4267     0.1455        171        640:  89%|████████▉ | 718/809 [14:16<01:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.228     0.4268     0.1454        268        640:  89%|████████▉ | 720/809 [14:19<01:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4267     0.1454        186        640:  89%|████████▉ | 722/809 [14:21<01:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4268     0.1455        176        640:  89%|████████▉ | 724/809 [14:23<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4268     0.1455        259        640:  90%|████████▉ | 725/809 [14:24<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4267     0.1455        187        640:  90%|████████▉ | 727/809 [14:27<01:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4266     0.1456        194        640:  90%|█████████ | 729/809 [14:29<01:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4267     0.1455        113        640:  90%|█████████ | 730/809 [14:30<01:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4266     0.1458        399        640:  90%|█████████ | 732/809 [14:33<01:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4268     0.1456        287        640:  91%|█████████ | 734/809 [14:35<01:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4268     0.1456        234        640:  91%|█████████ | 735/809 [14:36<01:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4268     0.1455        299        640:  91%|█████████ | 737/809 [14:39<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4265     0.1456        274        640:  91%|█████████▏| 739/809 [14:41<01:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4265     0.1456        249        640:  91%|█████████▏| 740/809 [14:42<01:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4264     0.1455        294        640:  92%|█████████▏| 742/809 [14:45<01:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4263     0.1456        385        640:  92%|█████████▏| 744/809 [14:47<01:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4263     0.1455        141        640:  92%|█████████▏| 745/809 [14:48<01:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4261     0.1455        317        640:  92%|█████████▏| 747/809 [14:51<01:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4263     0.1453        193        640:  93%|█████████▎| 749/809 [14:53<01:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.229     0.4264     0.1452        227        640:  93%|█████████▎| 750/809 [14:54<01:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4263     0.1456        195        640:  93%|█████████▎| 752/809 [14:57<01:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4261     0.1455        237        640:  93%|█████████▎| 754/809 [14:59<01:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4261     0.1456        346        640:  93%|█████████▎| 755/809 [15:00<01:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4262     0.1455        294        640:  94%|█████████▎| 757/809 [15:03<01:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4258     0.1459        643        640:  94%|█████████▍| 759/809 [15:05<01:02,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4258     0.1459        383        640:  94%|█████████▍| 760/809 [15:07<01:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4256     0.1464        180        640:  94%|█████████▍| 762/809 [15:09<00:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4256     0.1463        117        640:  94%|█████████▍| 764/809 [15:11<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4254     0.1465        438        640:  95%|█████████▍| 765/809 [15:13<00:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4253     0.1464        296        640:  95%|█████████▍| 767/809 [15:15<00:52,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4252     0.1464        377        640:  95%|█████████▌| 769/809 [15:17<00:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4253     0.1463        215        640:  95%|█████████▌| 770/809 [15:19<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4252     0.1463        434        640:  95%|█████████▌| 772/809 [15:21<00:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4249     0.1466        672        640:  96%|█████████▌| 773/809 [15:22<00:44,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235      0.425     0.1465        215        640:  96%|█████████▌| 775/809 [15:25<00:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4248     0.1466        297        640:  96%|█████████▌| 777/809 [15:27<00:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4248     0.1466        197        640:  96%|█████████▌| 778/809 [15:28<00:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234      0.425     0.1464        287        640:  96%|█████████▋| 780/809 [15:31<00:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.235     0.4249     0.1465        224        640:  97%|█████████▋| 782/809 [15:33<00:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4252     0.1464         86        640:  97%|█████████▋| 784/809 [15:35<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4254     0.1463        162        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4253     0.1462        386        640:  97%|█████████▋| 787/809 [15:39<00:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.234     0.4253     0.1462        354        640:  98%|█████████▊| 789/809 [15:41<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4254     0.1461        125        640:  98%|█████████▊| 790/809 [15:42<00:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4254     0.1462        455        640:  98%|█████████▊| 792/809 [15:45<00:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4255     0.1461        133        640:  98%|█████████▊| 794/809 [15:47<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.233     0.4256      0.146        213        640:  98%|█████████▊| 795/809 [15:48<00:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232     0.4258     0.1459        182        640:  99%|█████████▊| 797/809 [15:51<00:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.232      0.426     0.1458        376        640:  99%|█████████▉| 799/809 [15:53<00:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4263     0.1458        463        640:  99%|█████████▉| 801/809 [15:56<00:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4264     0.1457        246        640:  99%|█████████▉| 802/809 [15:57<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G      1.231     0.4266     0.1456        255        640:  99%|█████████▉| 804/809 [15:59<00:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1456        323        640: 100%|█████████▉| 806/809 [16:01<00:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     35/300        14G       1.23     0.4268     0.1457        347        640: 100%|█████████▉| 807/809 [16:03<00:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   3%|▎         | 2/69 [00:00<00:17,  3.85it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  38%|███▊      | 26/69 [00:06<00:11,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  49%|████▉     | 34/69 [00:08<00:08,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  61%|██████    | 42/69 [00:10<00:06,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:06,  3.31it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  3.80it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G     0.8593     0.5692    0.08357        193        640:   0%|          | 1/809 [00:01<15:10,  1.13s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G     0.8336      0.529     0.1006         87        640:   0%|          | 3/809 [00:03<15:34,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G     0.9119     0.5198     0.1033        221        640:   1%|          | 5/809 [00:05<15:38,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G     0.9956     0.4976     0.1283        430        640:   1%|          | 6/809 [00:07<16:15,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.055     0.4757     0.1426        473        640:   1%|          | 8/809 [00:09<16:03,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.062     0.4718     0.1406        129        640:   1%|          | 10/809 [00:11<15:43,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.036     0.4711     0.1314        107        640:   1%|▏         | 12/809 [00:14<15:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.042     0.4721     0.1301        184        640:   2%|▏         | 13/809 [00:15<15:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.074     0.4635     0.1326        171        640:   2%|▏         | 15/809 [00:17<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.065     0.4664     0.1281        196        640:   2%|▏         | 17/809 [00:20<15:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.071      0.467     0.1272        155        640:   2%|▏         | 18/809 [00:21<15:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.085     0.4601     0.1275        225        640:   2%|▏         | 20/809 [00:23<15:53,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.082     0.4602     0.1276        112        640:   3%|▎         | 22/809 [00:26<15:57,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.106     0.4523     0.1309        578        640:   3%|▎         | 23/809 [00:27<15:52,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.118     0.4481     0.1354        219        640:   3%|▎         | 25/809 [00:29<15:46,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.123     0.4488     0.1395        523        640:   3%|▎         | 27/809 [00:32<15:45,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.144     0.4435     0.1458        418        640:   3%|▎         | 28/809 [00:33<15:50,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.148     0.4456     0.1455        515        640:   4%|▎         | 30/809 [00:35<15:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.145     0.4481     0.1445        179        640:   4%|▍         | 31/809 [00:37<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.147     0.4479     0.1445        264        640:   4%|▍         | 33/809 [00:39<15:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.157      0.445     0.1483        547        640:   4%|▍         | 35/809 [00:41<15:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.164     0.4418     0.1509        404        640:   4%|▍         | 36/809 [00:43<15:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.164     0.4417     0.1498        235        640:   5%|▍         | 38/809 [00:45<15:52,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.154      0.444     0.1468        210        640:   5%|▍         | 40/809 [00:47<15:22,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.16      0.442     0.1479        292        640:   5%|▌         | 41/809 [00:49<15:30,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.174     0.4375     0.1514        602        640:   5%|▌         | 43/809 [00:51<15:49,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.163     0.4402      0.149        361        640:   6%|▌         | 45/809 [00:54<15:21,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.161     0.4407     0.1482        219        640:   6%|▌         | 46/809 [00:55<15:27,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.171     0.4373     0.1484        347        640:   6%|▌         | 48/809 [00:57<15:24,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.173     0.4359     0.1468        189        640:   6%|▌         | 50/809 [01:00<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.176      0.435     0.1466        296        640:   6%|▋         | 52/809 [01:02<14:54,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.172     0.4367     0.1447        135        640:   7%|▋         | 54/809 [01:04<15:01,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.165     0.4382     0.1441        135        640:   7%|▋         | 55/809 [01:05<14:47,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.18     0.4358     0.1541        172        640:   7%|▋         | 57/809 [01:08<15:04,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.172     0.4388     0.1522        228        640:   7%|▋         | 59/809 [01:10<14:46,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.167     0.4405      0.151        265        640:   7%|▋         | 60/809 [01:11<14:40,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.167     0.4423     0.1502        456        640:   8%|▊         | 62/809 [01:14<14:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.179      0.439     0.1502        327        640:   8%|▊         | 64/809 [01:16<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.18     0.4388     0.1498        163        640:   8%|▊         | 65/809 [01:17<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.181     0.4393     0.1496        186        640:   8%|▊         | 67/809 [01:20<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.193     0.4356      0.152        331        640:   9%|▊         | 69/809 [01:22<14:45,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.196      0.434     0.1512        185        640:   9%|▊         | 70/809 [01:23<14:52,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.199     0.4332     0.1507        317        640:   9%|▉         | 72/809 [01:26<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G        1.2     0.4329     0.1499        276        640:   9%|▉         | 74/809 [01:28<14:35,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.201     0.4321     0.1495        154        640:   9%|▉         | 75/809 [01:29<14:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.197     0.4347     0.1494         88        640:  10%|▉         | 77/809 [01:32<14:27,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.198     0.4345     0.1486        284        640:  10%|▉         | 79/809 [01:34<14:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.202     0.4336     0.1487         74        640:  10%|█         | 81/809 [01:36<14:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.204     0.4326     0.1493        130        640:  10%|█         | 82/809 [01:38<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.213     0.4291     0.1493        280        640:  10%|█         | 84/809 [01:40<14:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.219     0.4272     0.1491        286        640:  11%|█         | 86/809 [01:42<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.217     0.4283     0.1484        128        640:  11%|█         | 87/809 [01:43<14:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.217     0.4282     0.1475        132        640:  11%|█         | 89/809 [01:46<14:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.213     0.4302      0.147        146        640:  11%|█         | 91/809 [01:48<14:34,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.21     0.4305     0.1469        123        640:  11%|█▏        | 92/809 [01:49<14:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.212       0.43     0.1464        201        640:  12%|█▏        | 94/809 [01:52<14:21,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.215     0.4291     0.1461        242        640:  12%|█▏        | 96/809 [01:54<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.213     0.4304     0.1457        139        640:  12%|█▏        | 97/809 [01:55<13:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.22     0.4286     0.1463        226        640:  12%|█▏        | 99/809 [01:58<14:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4271     0.1484        667        640:  12%|█▏        | 101/809 [02:00<14:43,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4271     0.1483        216        640:  13%|█▎        | 102/809 [02:02<14:38,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4266      0.148        238        640:  13%|█▎        | 104/809 [02:04<14:29,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4279     0.1473        231        640:  13%|█▎        | 106/809 [02:06<14:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4274     0.1478        430        640:  13%|█▎        | 107/809 [02:08<14:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4282     0.1473        277        640:  13%|█▎        | 109/809 [02:10<13:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226      0.428     0.1466        275        640:  14%|█▎        | 111/809 [02:12<13:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1464        160        640:  14%|█▍        | 112/809 [02:14<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4273     0.1461        249        640:  14%|█▍        | 114/809 [02:16<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238      0.425     0.1473        165        640:  14%|█▍        | 116/809 [02:18<13:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4243     0.1479        191        640:  14%|█▍        | 117/809 [02:19<13:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241      0.425     0.1472         72        640:  15%|█▍        | 119/809 [02:22<13:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4254     0.1464        218        640:  15%|█▍        | 121/809 [02:24<13:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4258      0.146        422        640:  15%|█▌        | 122/809 [02:25<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4265     0.1457        296        640:  15%|█▌        | 124/809 [02:28<13:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232      0.427     0.1454        137        640:  16%|█▌        | 126/809 [02:30<13:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277      0.145         49        640:  16%|█▌        | 128/809 [02:33<13:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.233      0.427     0.1462        314        640:  16%|█▌        | 129/809 [02:34<13:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4264      0.147        180        640:  16%|█▌        | 131/809 [02:36<13:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4273     0.1467        108        640:  16%|█▋        | 133/809 [02:38<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4263     0.1476        498        640:  17%|█▋        | 134/809 [02:40<13:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232      0.428     0.1474        224        640:  17%|█▋        | 136/809 [02:42<13:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4292     0.1465        239        640:  17%|█▋        | 138/809 [02:44<13:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23      0.429     0.1475        369        640:  17%|█▋        | 139/809 [02:46<13:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4302     0.1473         83        640:  17%|█▋        | 141/809 [02:48<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4309     0.1467        277        640:  18%|█▊        | 143/809 [02:50<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4306     0.1466        331        640:  18%|█▊        | 144/809 [02:52<13:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228       0.43     0.1463        234        640:  18%|█▊        | 146/809 [02:54<13:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4297     0.1494        298        640:  18%|█▊        | 148/809 [02:56<13:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4298     0.1491        134        640:  18%|█▊        | 149/809 [02:57<13:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4298     0.1489        155        640:  19%|█▊        | 151/809 [03:00<13:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4289     0.1485        401        640:  19%|█▉        | 153/809 [03:02<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4285      0.148        170        640:  19%|█▉        | 155/809 [03:05<12:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4276     0.1481        334        640:  19%|█▉        | 156/809 [03:06<12:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4277     0.1475        266        640:  20%|█▉        | 158/809 [03:08<12:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4279     0.1471        251        640:  20%|█▉        | 160/809 [03:11<13:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4276     0.1473        433        640:  20%|█▉        | 161/809 [03:12<12:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4278     0.1468        404        640:  20%|██        | 163/809 [03:14<12:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4288     0.1463        268        640:  20%|██        | 165/809 [03:16<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.233     0.4291     0.1459        295        640:  21%|██        | 166/809 [03:18<12:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4289     0.1456        265        640:  21%|██        | 168/809 [03:20<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232     0.4292     0.1453        498        640:  21%|██        | 170/809 [03:22<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231      0.429     0.1453        221        640:  21%|██        | 171/809 [03:24<12:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232     0.4286     0.1453        211        640:  21%|██▏       | 173/809 [03:26<12:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4276     0.1454        419        640:  22%|██▏       | 175/809 [03:28<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4265     0.1451        218        640:  22%|██▏       | 177/809 [03:31<12:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4263     0.1449        152        640:  22%|██▏       | 178/809 [03:32<12:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4247     0.1456        383        640:  22%|██▏       | 180/809 [03:34<12:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.244     0.4245     0.1457        180        640:  22%|██▏       | 181/809 [03:35<12:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.249     0.4234     0.1473        551        640:  23%|██▎       | 183/809 [03:38<12:55,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.249     0.4235     0.1474        465        640:  23%|██▎       | 185/809 [03:40<12:46,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.246     0.4246     0.1471        148        640:  23%|██▎       | 186/809 [03:42<12:58,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.246     0.4251     0.1483        388        640:  23%|██▎       | 188/809 [03:44<12:47,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.248     0.4246     0.1486        465        640:  23%|██▎       | 190/809 [03:47<12:55,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.247     0.4252     0.1491        380        640:  24%|██▎       | 191/809 [03:48<12:45,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.249     0.4249     0.1513        522        640:  24%|██▍       | 193/809 [03:50<12:36,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4249     0.1526        253        640:  24%|██▍       | 195/809 [03:53<12:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25      0.425     0.1525        184        640:  24%|██▍       | 196/809 [03:54<12:47,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.247      0.426     0.1524        181        640:  24%|██▍       | 198/809 [03:57<12:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4255     0.1535        603        640:  25%|██▍       | 199/809 [03:58<12:44,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4251     0.1537        262        640:  25%|██▍       | 201/809 [04:00<12:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.248     0.4256     0.1535        243        640:  25%|██▌       | 203/809 [04:03<12:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25      0.425     0.1539        367        640:  25%|██▌       | 204/809 [04:04<12:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4249     0.1541        380        640:  25%|██▌       | 206/809 [04:06<12:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.254     0.4245     0.1561        212        640:  26%|██▌       | 208/809 [04:09<12:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.256     0.4241     0.1568        415        640:  26%|██▌       | 209/809 [04:10<12:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.255     0.4242      0.157        236        640:  26%|██▌       | 211/809 [04:12<11:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.253     0.4248     0.1568        411        640:  26%|██▋       | 213/809 [04:15<11:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.254     0.4245      0.157        373        640:  26%|██▋       | 214/809 [04:16<12:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.255      0.424     0.1574        209        640:  27%|██▋       | 216/809 [04:18<11:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.259     0.4231     0.1597        417        640:  27%|██▋       | 218/809 [04:21<11:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.259     0.4232     0.1596        289        640:  27%|██▋       | 219/809 [04:22<11:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.258     0.4233     0.1595         70        640:  27%|██▋       | 221/809 [04:24<11:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.257     0.4234     0.1592        155        640:  28%|██▊       | 223/809 [04:27<11:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.255     0.4238     0.1588        267        640:  28%|██▊       | 225/809 [04:29<11:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.255     0.4239     0.1586        165        640:  28%|██▊       | 226/809 [04:30<11:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.253     0.4243     0.1582        241        640:  28%|██▊       | 228/809 [04:32<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4248     0.1576        165        640:  28%|██▊       | 230/809 [04:35<11:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.251     0.4247      0.158        351        640:  29%|██▊       | 232/809 [04:37<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.251     0.4245     0.1581        344        640:  29%|██▉       | 233/809 [04:38<11:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.249     0.4249     0.1578        209        640:  29%|██▉       | 235/809 [04:41<11:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.254     0.4238     0.1598        634        640:  29%|██▉       | 237/809 [04:43<11:55,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.253     0.4241     0.1597         99        640:  29%|██▉       | 238/809 [04:45<11:56,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.25     0.4249     0.1594        177        640:  30%|██▉       | 240/809 [04:47<11:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.247     0.4256     0.1589        265        640:  30%|██▉       | 242/809 [04:49<11:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.246     0.4258     0.1587        320        640:  30%|███       | 243/809 [04:51<11:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4264     0.1582        164        640:  30%|███       | 245/809 [04:53<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271     0.1581        288        640:  31%|███       | 247/809 [04:55<11:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4276     0.1578        174        640:  31%|███       | 248/809 [04:57<11:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4283     0.1574        194        640:  31%|███       | 250/809 [04:59<10:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4292     0.1569         99        640:  31%|███       | 252/809 [05:01<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.233     0.4295     0.1567        252        640:  31%|███▏      | 253/809 [05:02<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4297     0.1564        171        640:  32%|███▏      | 255/809 [05:05<11:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232     0.4293     0.1563        312        640:  32%|███▏      | 257/809 [05:07<11:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4303     0.1558        233        640:  32%|███▏      | 259/809 [05:10<10:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4306     0.1556        191        640:  32%|███▏      | 260/809 [05:11<10:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4301     0.1554        313        640:  32%|███▏      | 262/809 [05:13<11:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4299     0.1556        203        640:  33%|███▎      | 264/809 [05:16<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4291     0.1564        390        640:  33%|███▎      | 265/809 [05:17<10:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4286      0.157        463        640:  33%|███▎      | 267/809 [05:19<10:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4286      0.157        197        640:  33%|███▎      | 269/809 [05:21<10:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4283     0.1569        216        640:  33%|███▎      | 270/809 [05:23<10:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4282     0.1569         89        640:  34%|███▎      | 272/809 [05:25<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4274     0.1572        458        640:  34%|███▍      | 274/809 [05:27<10:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4278      0.157        129        640:  34%|███▍      | 275/809 [05:29<10:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239      0.428     0.1567         89        640:  34%|███▍      | 277/809 [05:31<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4283     0.1564        200        640:  34%|███▍      | 279/809 [05:33<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4279     0.1562        311        640:  35%|███▍      | 280/809 [05:36<10:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4275     0.1563        253        640:  35%|███▍      | 282/809 [05:37<10:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4273     0.1562        307        640:  35%|███▌      | 284/809 [05:39<10:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4271     0.1562        371        640:  35%|███▌      | 286/809 [05:42<10:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4274     0.1561        107        640:  35%|███▌      | 287/809 [05:43<10:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237      0.428     0.1558        182        640:  36%|███▌      | 289/809 [05:45<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4279     0.1558        386        640:  36%|███▌      | 291/809 [05:48<10:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4277     0.1556        360        640:  36%|███▌      | 292/809 [05:49<10:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4276     0.1555        297        640:  36%|███▋      | 294/809 [05:51<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4273     0.1554        200        640:  37%|███▋      | 296/809 [05:53<09:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4274     0.1552         89        640:  37%|███▋      | 297/809 [05:55<09:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4274      0.155        149        640:  37%|███▋      | 299/809 [05:57<09:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4275     0.1557        239        640:  37%|███▋      | 301/809 [05:59<09:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4274     0.1554        207        640:  37%|███▋      | 303/809 [06:02<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4272     0.1553        312        640:  38%|███▊      | 304/809 [06:03<09:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4268     0.1552        288        640:  38%|███▊      | 306/809 [06:05<10:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4265     0.1549        154        640:  38%|███▊      | 308/809 [06:08<09:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4265     0.1548        107        640:  38%|███▊      | 309/809 [06:09<09:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4263     0.1546        147        640:  38%|███▊      | 311/809 [06:11<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4265     0.1546        146        640:  39%|███▊      | 313/809 [06:14<09:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4266     0.1545        307        640:  39%|███▉      | 314/809 [06:15<09:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4266     0.1544        341        640:  39%|███▉      | 316/809 [06:17<09:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271      0.154        331        640:  39%|███▉      | 318/809 [06:20<09:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271     0.1539        156        640:  39%|███▉      | 319/809 [06:21<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4266     0.1541        150        640:  40%|███▉      | 321/809 [06:23<09:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.244     0.4264     0.1542        259        640:  40%|███▉      | 323/809 [06:25<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.244     0.4262     0.1543        321        640:  40%|████      | 324/809 [06:27<09:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4266      0.154        108        640:  40%|████      | 326/809 [06:29<09:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242      0.427     0.1537        246        640:  41%|████      | 328/809 [06:31<09:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4273     0.1535        350        640:  41%|████      | 330/809 [06:34<09:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4272     0.1533        307        640:  41%|████      | 331/809 [06:35<09:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4279     0.1531        114        640:  41%|████      | 333/809 [06:37<09:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4279     0.1528        153        640:  41%|████▏     | 335/809 [06:40<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4278     0.1527        172        640:  42%|████▏     | 336/809 [06:41<09:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4279     0.1525        199        640:  42%|████▏     | 338/809 [06:43<09:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4284     0.1522        203        640:  42%|████▏     | 340/809 [06:45<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4282     0.1522        334        640:  42%|████▏     | 341/809 [06:47<09:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4283     0.1519        212        640:  42%|████▏     | 343/809 [06:49<09:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4288     0.1515        209        640:  43%|████▎     | 345/809 [06:51<09:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4288     0.1512        206        640:  43%|████▎     | 347/809 [06:54<08:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4289     0.1512        278        640:  43%|████▎     | 348/809 [06:55<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4291     0.1511        440        640:  43%|████▎     | 350/809 [06:57<09:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236      0.429     0.1518        593        640:  44%|████▎     | 352/809 [07:00<09:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236      0.429     0.1518        391        640:  44%|████▎     | 353/809 [07:01<09:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4288     0.1519        317        640:  44%|████▍     | 355/809 [07:03<09:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4285     0.1523        624        640:  44%|████▍     | 356/809 [07:05<09:19,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4283     0.1522        394        640:  44%|████▍     | 358/809 [07:07<09:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4283     0.1521        325        640:  44%|████▍     | 360/809 [07:10<09:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24      0.428     0.1525        440        640:  45%|████▍     | 361/809 [07:11<09:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4273      0.153        495        640:  45%|████▍     | 363/809 [07:13<09:15,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4273     0.1533        616        640:  45%|████▌     | 365/809 [07:16<09:11,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.243     0.4272     0.1532        208        640:  45%|████▌     | 366/809 [07:17<09:12,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.242     0.4275     0.1532        137        640:  45%|████▌     | 368/809 [07:20<09:03,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4275      0.153        185        640:  46%|████▌     | 370/809 [07:22<08:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4279     0.1528        131        640:  46%|████▌     | 371/809 [07:23<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4278     0.1529        335        640:  46%|████▌     | 373/809 [07:25<08:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4281     0.1527        182        640:  46%|████▋     | 375/809 [07:28<08:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4278     0.1533        259        640:  47%|████▋     | 377/809 [07:30<08:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4279     0.1532        337        640:  47%|████▋     | 378/809 [07:31<08:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4275     0.1533        299        640:  47%|████▋     | 380/809 [07:34<08:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4275     0.1532        222        640:  47%|████▋     | 382/809 [07:36<08:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4279      0.153         92        640:  47%|████▋     | 384/809 [07:38<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4279     0.1529        267        640:  48%|████▊     | 385/809 [07:40<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4273     0.1531        291        640:  48%|████▊     | 387/809 [07:42<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271     0.1532        199        640:  48%|████▊     | 389/809 [07:44<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271     0.1531        235        640:  48%|████▊     | 390/809 [07:46<08:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4274     0.1529        182        640:  48%|████▊     | 392/809 [07:48<08:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4272     0.1529        298        640:  49%|████▊     | 394/809 [07:50<08:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4272      0.153        343        640:  49%|████▉     | 395/809 [07:52<08:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.241     0.4271      0.153        245        640:  49%|████▉     | 397/809 [07:54<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.239     0.4274     0.1528        145        640:  49%|████▉     | 399/809 [07:56<08:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.24     0.4273     0.1529        341        640:  49%|████▉     | 400/809 [07:58<08:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4276     0.1527        342        640:  50%|████▉     | 402/809 [08:00<08:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.238     0.4278     0.1528        409        640:  50%|████▉     | 404/809 [08:02<08:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.237     0.4281     0.1527        165        640:  50%|█████     | 405/809 [08:04<08:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4285     0.1526        206        640:  50%|█████     | 407/809 [08:06<08:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4286     0.1525        153        640:  51%|█████     | 409/809 [08:08<07:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.236     0.4285     0.1525        300        640:  51%|█████     | 410/809 [08:10<07:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4287     0.1522        136        640:  51%|█████     | 412/809 [08:12<07:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4286     0.1527        372        640:  51%|█████     | 414/809 [08:14<08:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4288     0.1525        268        640:  51%|█████▏    | 415/809 [08:16<08:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4288     0.1525        183        640:  52%|█████▏    | 417/809 [08:18<07:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4291     0.1524        361        640:  52%|█████▏    | 419/809 [08:20<07:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4291     0.1524        245        640:  52%|█████▏    | 420/809 [08:22<07:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234      0.429     0.1523        434        640:  52%|█████▏    | 422/809 [08:24<08:01,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234      0.429     0.1522        239        640:  52%|█████▏    | 423/809 [08:25<07:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234      0.429     0.1529        103        640:  53%|█████▎    | 425/809 [08:28<07:54,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234     0.4291     0.1529        270        640:  53%|█████▎    | 427/809 [08:30<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.234      0.429     0.1529        322        640:  53%|█████▎    | 428/809 [08:31<07:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.235     0.4288     0.1528        135        640:  53%|█████▎    | 430/809 [08:34<07:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.233     0.4295     0.1525        235        640:  53%|█████▎    | 432/809 [08:36<07:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.233     0.4296     0.1524        409        640:  54%|█████▎    | 434/809 [08:39<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.232     0.4301     0.1523         75        640:  54%|█████▍    | 435/809 [08:40<07:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4302     0.1522        131        640:  54%|█████▍    | 437/809 [08:42<07:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4305     0.1522        153        640:  54%|█████▍    | 439/809 [08:44<07:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4303     0.1523        269        640:  54%|█████▍    | 440/809 [08:46<07:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4303     0.1523        253        640:  55%|█████▍    | 442/809 [08:48<07:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4305     0.1522        353        640:  55%|█████▍    | 444/809 [08:50<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4308     0.1521        170        640:  55%|█████▌    | 445/809 [08:51<07:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4303     0.1521        289        640:  55%|█████▌    | 447/809 [08:54<07:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4305      0.152        166        640:  56%|█████▌    | 449/809 [08:56<07:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4304      0.152        370        640:  56%|█████▌    | 450/809 [08:58<07:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4304     0.1518        258        640:  56%|█████▌    | 452/809 [09:00<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4306     0.1516        129        640:  56%|█████▌    | 454/809 [09:02<07:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4305     0.1516        328        640:  56%|█████▋    | 456/809 [09:05<07:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4305     0.1515        217        640:  56%|█████▋    | 457/809 [09:06<06:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4304     0.1518         88        640:  57%|█████▋    | 459/809 [09:08<06:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4301     0.1519        139        640:  57%|█████▋    | 461/809 [09:11<06:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4299     0.1519        317        640:  57%|█████▋    | 462/809 [09:12<07:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4295     0.1519        369        640:  57%|█████▋    | 464/809 [09:14<06:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4291      0.152        377        640:  58%|█████▊    | 466/809 [09:17<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4289      0.152        118        640:  58%|█████▊    | 467/809 [09:18<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4292     0.1519        333        640:  58%|█████▊    | 469/809 [09:20<06:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4294     0.1517        180        640:  58%|█████▊    | 471/809 [09:22<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4293     0.1516        205        640:  58%|█████▊    | 473/809 [09:25<06:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4291     0.1515        311        640:  59%|█████▊    | 474/809 [09:26<06:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4288     0.1516        288        640:  59%|█████▉    | 476/809 [09:28<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4292     0.1513         73        640:  59%|█████▉    | 477/809 [09:31<06:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4293     0.1512        181        640:  59%|█████▉    | 479/809 [09:32<06:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4291      0.151        136        640:  59%|█████▉    | 481/809 [09:34<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4291     0.1509        365        640:  60%|█████▉    | 483/809 [09:37<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4294     0.1508        212        640:  60%|█████▉    | 484/809 [09:38<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4294     0.1507        197        640:  60%|██████    | 486/809 [09:40<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4297     0.1505        183        640:  60%|██████    | 488/809 [09:43<06:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4298     0.1504        169        640:  60%|██████    | 489/809 [09:45<06:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4299     0.1503        141        640:  61%|██████    | 491/809 [09:46<06:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225       0.43     0.1501        265        640:  61%|██████    | 493/809 [09:48<06:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4303       0.15        286        640:  61%|██████    | 495/809 [09:51<06:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4302     0.1499        194        640:  61%|██████▏   | 496/809 [09:52<06:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4298       0.15        260        640:  62%|██████▏   | 498/809 [09:54<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4297     0.1498        179        640:  62%|██████▏   | 500/809 [09:57<05:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4294     0.1498        337        640:  62%|██████▏   | 501/809 [09:58<06:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4292     0.1498        273        640:  62%|██████▏   | 503/809 [10:00<06:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4291     0.1496        131        640:  62%|██████▏   | 505/809 [10:03<05:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4289     0.1496        313        640:  63%|██████▎   | 507/809 [10:05<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226      0.429     0.1495        149        640:  63%|██████▎   | 508/809 [10:06<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4289     0.1495        264        640:  63%|██████▎   | 510/809 [10:09<06:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4286     0.1493        194        640:  63%|██████▎   | 512/809 [10:11<05:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4287     0.1492        168        640:  63%|██████▎   | 513/809 [10:12<05:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4286     0.1491        222        640:  64%|██████▎   | 515/809 [10:14<05:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4284      0.149        192        640:  64%|██████▍   | 517/809 [10:17<05:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1489        104        640:  64%|██████▍   | 519/809 [10:19<05:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1489        402        640:  64%|██████▍   | 520/809 [10:20<05:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4285     0.1489        313        640:  65%|██████▍   | 522/809 [10:23<05:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4288     0.1487        118        640:  65%|██████▍   | 524/809 [10:25<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4287     0.1486        202        640:  65%|██████▍   | 525/809 [10:26<05:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4285     0.1487        421        640:  65%|██████▌   | 527/809 [10:29<05:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4284     0.1487        422        640:  65%|██████▌   | 528/809 [10:30<05:45,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4283     0.1489        178        640:  66%|██████▌   | 530/809 [10:32<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4283     0.1488        343        640:  66%|██████▌   | 532/809 [10:35<05:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4282     0.1488        325        640:  66%|██████▌   | 533/809 [10:36<05:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4284     0.1487        334        640:  66%|██████▋   | 536/809 [10:39<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4285     0.1486        260        640:  66%|██████▋   | 537/809 [10:41<05:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4286     0.1485        333        640:  67%|██████▋   | 539/809 [10:43<05:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4284     0.1489        457        640:  67%|██████▋   | 541/809 [10:45<05:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4285     0.1489        294        640:  67%|██████▋   | 542/809 [10:47<05:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1488        153        640:  67%|██████▋   | 544/809 [10:49<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4289     0.1487        156        640:  67%|██████▋   | 546/809 [10:51<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4292     0.1485        235        640:  68%|██████▊   | 548/809 [10:54<05:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4292     0.1484        282        640:  68%|██████▊   | 549/809 [10:55<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4289     0.1487        460        640:  68%|██████▊   | 551/809 [10:57<05:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4291     0.1486        204        640:  68%|██████▊   | 552/809 [10:58<05:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228      0.429     0.1485        264        640:  68%|██████▊   | 554/809 [11:01<05:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4288     0.1484        251        640:  69%|██████▊   | 556/809 [11:03<05:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1485        472        640:  69%|██████▉   | 557/809 [11:05<05:11,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4287     0.1485        164        640:  69%|██████▉   | 559/809 [11:07<05:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4288     0.1485        313        640:  69%|██████▉   | 561/809 [11:09<04:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4288     0.1484        211        640:  69%|██████▉   | 562/809 [11:11<04:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1487        278        640:  70%|██████▉   | 564/809 [11:13<04:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4286     0.1487        248        640:  70%|██████▉   | 566/809 [11:15<04:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4287     0.1486        195        640:  70%|███████   | 567/809 [11:17<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228      0.429     0.1485        299        640:  70%|███████   | 569/809 [11:19<04:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4289     0.1484        321        640:  71%|███████   | 571/809 [11:21<04:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228      0.429     0.1483        234        640:  71%|███████   | 573/809 [11:24<04:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4288     0.1483        338        640:  71%|███████   | 574/809 [11:25<04:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4286     0.1484        282        640:  71%|███████   | 576/809 [11:27<04:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4287     0.1483        185        640:  71%|███████▏  | 578/809 [11:29<04:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4288     0.1482        169        640:  72%|███████▏  | 579/809 [11:31<04:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227      0.429     0.1482        111        640:  72%|███████▏  | 581/809 [11:33<04:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4291     0.1482        397        640:  72%|███████▏  | 583/809 [11:35<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4289     0.1481        244        640:  72%|███████▏  | 584/809 [11:37<04:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4287     0.1482        422        640:  72%|███████▏  | 586/809 [11:39<04:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4285     0.1481        227        640:  73%|███████▎  | 588/809 [11:42<04:32,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4284      0.148        273        640:  73%|███████▎  | 589/809 [11:43<04:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4283     0.1481        160        640:  73%|███████▎  | 591/809 [11:45<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4283     0.1481        231        640:  73%|███████▎  | 593/809 [11:48<04:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4282     0.1485        356        640:  73%|███████▎  | 594/809 [11:49<04:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4286     0.1483        129        640:  74%|███████▎  | 596/809 [11:51<04:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4286     0.1482        180        640:  74%|███████▍  | 598/809 [11:54<04:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4285     0.1484        356        640:  74%|███████▍  | 599/809 [11:55<04:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4286     0.1483        365        640:  74%|███████▍  | 601/809 [11:57<04:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4288     0.1482        207        640:  75%|███████▍  | 603/809 [11:59<04:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289      0.148        122        640:  75%|███████▍  | 605/809 [12:02<03:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4292     0.1479        146        640:  75%|███████▍  | 606/809 [12:03<03:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289     0.1479        168        640:  75%|███████▌  | 608/809 [12:05<03:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289     0.1478        160        640:  75%|███████▌  | 610/809 [12:08<03:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289     0.1479        307        640:  76%|███████▌  | 611/809 [12:09<03:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4286     0.1483        145        640:  76%|███████▌  | 613/809 [12:11<03:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4286     0.1484        276        640:  76%|███████▌  | 615/809 [12:14<03:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1482        137        640:  76%|███████▌  | 616/809 [12:15<03:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4286     0.1482        257        640:  76%|███████▋  | 618/809 [12:17<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.223     0.4289     0.1481        153        640:  77%|███████▋  | 620/809 [12:20<03:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4287     0.1481        322        640:  77%|███████▋  | 621/809 [12:21<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4285     0.1482        535        640:  77%|███████▋  | 623/809 [12:23<03:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4282     0.1484        352        640:  77%|███████▋  | 625/809 [12:26<03:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4283     0.1482        113        640:  77%|███████▋  | 626/809 [12:27<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4282     0.1482        409        640:  78%|███████▊  | 628/809 [12:29<03:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4283     0.1481        317        640:  78%|███████▊  | 630/809 [12:32<03:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1481        318        640:  78%|███████▊  | 631/809 [12:33<03:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4278     0.1482        372        640:  78%|███████▊  | 633/809 [12:35<03:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4279      0.148        128        640:  78%|███████▊  | 635/809 [12:38<03:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4279      0.148        186        640:  79%|███████▊  | 636/809 [12:39<03:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4276      0.148        392        640:  79%|███████▉  | 638/809 [12:41<03:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1479         97        640:  79%|███████▉  | 640/809 [12:43<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1478        249        640:  79%|███████▉  | 642/809 [12:46<03:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4281     0.1477        157        640:  79%|███████▉  | 643/809 [12:47<03:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4282     0.1476        281        640:  80%|███████▉  | 645/809 [12:49<03:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4281     0.1479        543        640:  80%|███████▉  | 647/809 [12:52<03:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1479        268        640:  80%|████████  | 648/809 [12:53<03:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4279     0.1478        144        640:  80%|████████  | 650/809 [12:55<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4281     0.1477        312        640:  81%|████████  | 652/809 [12:58<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225      0.428     0.1477        452        640:  81%|████████  | 653/809 [12:59<03:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4282     0.1477        257        640:  81%|████████  | 655/809 [13:01<03:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4285     0.1475        244        640:  81%|████████  | 657/809 [13:04<02:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4283     0.1476        463        640:  81%|████████▏ | 658/809 [13:05<03:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4283     0.1476        398        640:  82%|████████▏ | 660/809 [13:07<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4284     0.1475        207        640:  82%|████████▏ | 662/809 [13:10<02:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4284     0.1475        249        640:  82%|████████▏ | 663/809 [13:11<02:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4282     0.1475        262        640:  82%|████████▏ | 665/809 [13:13<02:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226      0.428     0.1478        197        640:  82%|████████▏ | 667/809 [13:16<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4278     0.1481        366        640:  83%|████████▎ | 668/809 [13:17<02:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4278     0.1483        303        640:  83%|████████▎ | 670/809 [13:19<02:51,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4278     0.1483        276        640:  83%|████████▎ | 672/809 [13:22<02:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4277     0.1482        183        640:  83%|████████▎ | 673/809 [13:23<02:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4277     0.1482        171        640:  83%|████████▎ | 675/809 [13:25<02:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1481        152        640:  84%|████████▎ | 677/809 [13:28<02:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1481        254        640:  84%|████████▍ | 678/809 [13:29<02:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4275     0.1482        359        640:  84%|████████▍ | 680/809 [13:31<02:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4276     0.1481        129        640:  84%|████████▍ | 682/809 [13:34<02:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277      0.148        334        640:  85%|████████▍ | 684/809 [13:36<02:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4278     0.1481        464        640:  85%|████████▍ | 686/809 [13:38<02:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4279      0.148         87        640:  85%|████████▍ | 687/809 [13:40<02:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4278     0.1479        147        640:  85%|████████▌ | 689/809 [13:42<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1479        242        640:  85%|████████▌ | 691/809 [13:44<02:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4274     0.1479        178        640:  86%|████████▌ | 692/809 [13:45<02:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4274     0.1478        211        640:  86%|████████▌ | 694/809 [13:48<02:27,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4273     0.1478        179        640:  86%|████████▌ | 695/809 [13:49<02:21,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4273     0.1477        220        640:  86%|████████▌ | 697/809 [13:52<02:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4274     0.1476        243        640:  86%|████████▋ | 699/809 [13:54<02:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4272     0.1475        222        640:  87%|████████▋ | 701/809 [13:56<02:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4272     0.1477        237        640:  87%|████████▋ | 702/809 [13:58<02:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4274     0.1476         90        640:  87%|████████▋ | 704/809 [14:00<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4275     0.1475        224        640:  87%|████████▋ | 706/809 [14:02<02:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4273     0.1475        366        640:  87%|████████▋ | 707/809 [14:04<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.231     0.4273     0.1474        125        640:  88%|████████▊ | 709/809 [14:06<01:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4275     0.1473        220        640:  88%|████████▊ | 711/809 [14:08<01:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4277     0.1472        206        640:  88%|████████▊ | 712/809 [14:09<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4276     0.1473        150        640:  88%|████████▊ | 714/809 [14:12<01:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4275     0.1472        164        640:  89%|████████▊ | 716/809 [14:14<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4276     0.1472        247        640:  89%|████████▊ | 717/809 [14:16<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G       1.23     0.4277     0.1471        263        640:  89%|████████▉ | 719/809 [14:18<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4278     0.1472         77        640:  89%|████████▉ | 721/809 [14:20<01:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4278     0.1471        413        640:  89%|████████▉ | 723/809 [14:23<01:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4279     0.1471        276        640:  89%|████████▉ | 724/809 [14:24<01:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229      0.428     0.1475        103        640:  90%|████████▉ | 726/809 [14:26<01:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229      0.428     0.1476        158        640:  90%|████████▉ | 728/809 [14:29<01:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4279     0.1477        210        640:  90%|█████████ | 729/809 [14:30<01:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4282     0.1476        351        640:  90%|█████████ | 731/809 [14:32<01:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4282     0.1477        415        640:  91%|█████████ | 733/809 [14:35<01:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4282     0.1477        370        640:  91%|█████████ | 734/809 [14:36<01:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229      0.428     0.1479        324        640:  91%|█████████ | 736/809 [14:38<01:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4284     0.1478        138        640:  91%|█████████ | 738/809 [14:41<01:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4284     0.1479        291        640:  91%|█████████▏| 739/809 [14:42<01:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4285     0.1478        259        640:  92%|█████████▏| 741/809 [14:44<01:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4285     0.1478         76        640:  92%|█████████▏| 743/809 [14:47<01:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4285     0.1479        356        640:  92%|█████████▏| 744/809 [14:48<01:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4284     0.1479        280        640:  92%|█████████▏| 746/809 [14:50<01:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4286     0.1478        260        640:  92%|█████████▏| 748/809 [14:53<01:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4284     0.1479        779        640:  93%|█████████▎| 749/809 [14:54<01:16,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4283     0.1481        361        640:  93%|█████████▎| 751/809 [14:57<01:12,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4281     0.1483        487        640:  93%|█████████▎| 752/809 [14:58<01:11,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4281     0.1484        160        640:  93%|█████████▎| 754/809 [15:00<01:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4281     0.1484        226        640:  93%|████████��▎| 756/809 [15:03<01:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4279     0.1485        449        640:  94%|█████████▎| 757/809 [15:04<01:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1487        272        640:  94%|█████████▍| 759/809 [15:06<01:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1487        142        640:  94%|█████████▍| 761/809 [15:09<00:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4277     0.1488        392        640:  94%|█████████▍| 762/809 [15:10<00:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.229     0.4278     0.1488        273        640:  94%|█████████▍| 764/809 [15:12<00:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228      0.428     0.1488        192        640:  95%|█████████▍| 766/809 [15:15<00:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228      0.428     0.1488        219        640:  95%|█████████▍| 767/809 [15:16<00:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4281     0.1487        276        640:  95%|█████████▌| 769/809 [15:18<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4281     0.1487        252        640:  95%|█████████▌| 771/809 [15:21<00:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4281     0.1487        330        640:  95%|█████████▌| 772/809 [15:22<00:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227      0.428      0.149        268        640:  96%|█████████▌| 774/809 [15:24<00:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.228     0.4279      0.149        367        640:  96%|█████████▌| 776/809 [15:27<00:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227      0.428      0.149        181        640:  96%|█████████▌| 777/809 [15:28<00:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4284     0.1489        134        640:  96%|█████████▋| 779/809 [15:30<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4285     0.1488        256        640:  97%|█████████▋| 781/809 [15:32<00:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1487        107        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1487        260        640:  97%|█████████▋| 784/809 [15:36<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289     0.1486        181        640:  97%|█████████▋| 786/809 [15:38<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4289     0.1486        112        640:  97%|█████████▋| 788/809 [15:41<00:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224      0.429     0.1485        136        640:  98%|█████████▊| 789/809 [15:42<00:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.224     0.4289     0.1485        227        640:  98%|█████████▊| 791/809 [15:44<00:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1486        282        640:  98%|█████████▊| 793/809 [15:47<00:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1486        206        640:  98%|█████████▊| 794/809 [15:48<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1486        219        640:  98%|█████████▊| 796/809 [15:50<00:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4287     0.1486        152        640:  99%|█████████▊| 798/809 [15:53<00:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4288     0.1485        256        640:  99%|█████████▉| 800/809 [15:55<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.225     0.4288     0.1486        386        640:  99%|█████████▉| 801/809 [15:56<00:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4285     0.1486        295        640:  99%|█████████▉| 803/809 [15:58<00:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4285     0.1486        283        640: 100%|█████████▉| 805/809 [16:01<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.226     0.4284     0.1487        302        640: 100%|█████████▉| 806/809 [16:02<00:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     36/300      13.5G      1.227     0.4283     0.1487        460        640: 100%|█████████▉| 808/809 [16:04<00:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.04it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:09<00:08,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:11<00:06,  3.78it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:04,  3.64it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.85it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.366     0.4007     0.1664        172        640:   0%|          | 2/809 [00:02<15:48,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.365     0.3972     0.1476        193        640:   0%|          | 3/809 [00:03<15:36,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.399     0.4023     0.1605        180        640:   1%|          | 5/809 [00:05<15:54,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.324     0.4261     0.1522        168        640:   1%|          | 7/809 [00:08<15:39,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.314     0.4243     0.1489        154        640:   1%|          | 8/809 [00:09<15:39,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.255     0.4539     0.1391        185        640:   1%|          | 10/809 [00:11<15:52,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4513     0.1344        283        640:   1%|▏         | 12/809 [00:14<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4517     0.1322        218        640:   2%|▏         | 13/809 [00:15<15:54,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228     0.4543     0.1326        158        640:   2%|▏         | 15/809 [00:17<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.253     0.4487     0.1436        163        640:   2%|▏         | 17/809 [00:20<16:10,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.251     0.4454     0.1424        212        640:   2%|▏         | 18/809 [00:21<15:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.244     0.4496     0.1394        216        640:   2%|▏         | 20/809 [00:23<15:53,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.264     0.4417     0.1383        257        640:   3%|▎         | 22/809 [00:26<15:53,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.31     0.4273     0.1419        131        640:   3%|▎         | 24/809 [00:28<15:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.333     0.4187      0.143        185        640:   3%|▎         | 26/809 [00:31<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.341     0.4158     0.1434        114        640:   3%|▎         | 27/809 [00:32<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.359     0.4099      0.146        146        640:   4%|▎         | 29/809 [00:34<15:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.348     0.4123     0.1458        114        640:   4%|▍         | 31/809 [00:36<15:09,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.349     0.4121     0.1472        192        640:   4%|▍         | 32/809 [00:38<15:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.353     0.4105     0.1476        394        640:   4%|▍         | 34/809 [00:40<15:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.36     0.4094     0.1545        459        640:   4%|▍         | 36/809 [00:42<15:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.357     0.4103     0.1536        337        640:   5%|▍         | 37/809 [00:44<15:39,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.348     0.4119      0.152         94        640:   5%|▍         | 39/809 [00:46<15:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.353     0.4089     0.1539        406        640:   5%|▌         | 41/809 [00:48<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.354     0.4089     0.1548        193        640:   5%|▌         | 42/809 [00:49<15:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.358     0.4073     0.1545        322        640:   5%|▌         | 44/809 [00:52<15:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.357     0.4089     0.1579        402        640:   6%|▌         | 46/809 [00:54<15:26,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.365     0.4065      0.161        268        640:   6%|▌         | 47/809 [00:56<15:30,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.363     0.4057     0.1598        263        640:   6%|▌         | 49/809 [00:58<15:09,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.359     0.4061     0.1588         94        640:   6%|▋         | 51/809 [01:00<14:56,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.37     0.4036     0.1643        440        640:   6%|▋         | 52/809 [01:02<15:19,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.374     0.4019     0.1645        426        640:   7%|▋         | 54/809 [01:04<15:14,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.382        0.4     0.1654        326        640:   7%|▋         | 56/809 [01:06<15:12,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.38     0.4007      0.165        348        640:   7%|▋         | 57/809 [01:08<15:08,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.387     0.3985     0.1666        387        640:   7%|▋         | 59/809 [01:10<15:13,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.383     0.3998     0.1653        344        640:   8%|▊         | 61/809 [01:12<15:15,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.387     0.3978     0.1673        557        640:   8%|▊         | 62/809 [01:14<15:19,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.379      0.401     0.1668        194        640:   8%|▊         | 64/809 [01:16<14:54,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.371     0.4036     0.1659        372        640:   8%|▊         | 66/809 [01:18<14:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.365      0.405     0.1652        239        640:   8%|▊         | 67/809 [01:20<14:39,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.364     0.4059     0.1651        427        640:   9%|▊         | 69/809 [01:22<14:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.359     0.4067     0.1639        179        640:   9%|▉         | 71/809 [01:24<14:29,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.362     0.4057     0.1639        396        640:   9%|▉         | 72/809 [01:26<14:30,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.364     0.4039      0.164        206        640:   9%|▉         | 74/809 [01:28<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.364     0.4031     0.1631        336        640:   9%|▉         | 76/809 [01:30<14:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.367     0.4018     0.1635        319        640:  10%|▉         | 77/809 [01:32<15:12,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.366     0.4021     0.1627        316        640:  10%|▉         | 79/809 [01:34<14:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.36     0.4031     0.1614        231        640:  10%|█         | 81/809 [01:36<14:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.362     0.4019     0.1616        240        640:  10%|█         | 83/809 [01:39<14:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.357     0.4028     0.1607        127        640:  10%|█         | 84/809 [01:40<14:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.358     0.4023     0.1618        281        640:  11%|█         | 86/809 [01:42<14:35,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.358     0.4022     0.1615        177        640:  11%|█         | 87/809 [01:44<14:27,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.352     0.4028     0.1605        212        640:  11%|█         | 89/809 [01:46<14:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.346     0.4046     0.1598        329        640:  11%|█         | 91/809 [01:48<14:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.341     0.4061     0.1592        241        640:  11%|█▏        | 93/809 [01:51<14:21,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.342     0.4059     0.1594        253        640:  12%|█▏        | 94/809 [01:52<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.333     0.4089     0.1582        179        640:  12%|█▏        | 96/809 [01:54<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.323     0.4118     0.1571        186        640:  12%|█▏        | 98/809 [01:57<13:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.331     0.4096      0.162        594        640:  12%|█▏        | 99/809 [01:58<14:30,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.321     0.4127     0.1606        210        640:  12%|█▏        | 101/809 [02:00<14:38,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.315     0.4144     0.1595         96        640:  13%|█▎        | 102/809 [02:03<14:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.313     0.4155     0.1592        198        640:  13%|█▎        | 104/809 [02:04<14:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.318     0.4142     0.1601        382        640:  13%|█▎        | 106/809 [02:06<14:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.315     0.4151     0.1596        168        640:  13%|█▎        | 108/809 [02:09<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.316     0.4145     0.1594        445        640:  13%|█▎        | 109/809 [02:10<14:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.313     0.4152     0.1586        240        640:  14%|█▎        | 111/809 [02:12<13:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.306     0.4177     0.1576        216        640:  14%|█▍        | 113/809 [02:15<13:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.304      0.418     0.1574        333        640:  14%|█▍        | 114/809 [02:16<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.303     0.4188      0.157        249        640:  14%|��▍        | 116/809 [02:18<13:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.302     0.4187     0.1564        145        640:  15%|█▍        | 118/809 [02:21<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.305     0.4177     0.1579        510        640:  15%|█▍        | 119/809 [02:22<13:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.306     0.4174     0.1583        177        640:  15%|█▍        | 121/809 [02:24<13:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.309     0.4163     0.1589        357        640:  15%|█▌        | 123/809 [02:27<13:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.307     0.4165     0.1589        155        640:  15%|█▌        | 124/809 [02:28<13:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.306     0.4165     0.1586        307        640:  16%|█▌        | 126/809 [02:30<13:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.306     0.4164     0.1584        139        640:  16%|█▌        | 128/809 [02:33<13:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.305     0.4168     0.1581        259        640:  16%|█▌        | 129/809 [02:34<13:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.309     0.4153     0.1598        520        640:  16%|█▌        | 131/809 [02:36<13:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.309     0.4154     0.1597        726        640:  16%|█▋        | 133/809 [02:39<14:25,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.306     0.4163     0.1591        168        640:  17%|█▋        | 134/809 [02:40<13:55,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.305     0.4165     0.1589        249        640:  17%|█▋        | 136/809 [02:42<13:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.302      0.417     0.1586        174        640:  17%|█▋        | 138/809 [02:45<13:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.301     0.4169     0.1582        363        640:  17%|█▋        | 139/809 [02:46<13:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4176     0.1574        354        640:  17%|█▋        | 141/809 [02:48<13:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4177     0.1598        474        640:  18%|█▊        | 143/809 [02:51<13:38,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.299     0.4174     0.1596        373        640:  18%|█▊        | 144/809 [02:52<13:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4172     0.1611        139        640:  18%|█▊        | 146/809 [02:54<13:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4173     0.1607        135        640:  18%|█▊        | 148/809 [02:57<13:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.301     0.4169     0.1608        340        640:  18%|█▊        | 149/809 [02:58<13:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4168     0.1602        317        640:  19%|█▊        | 151/809 [03:00<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.302     0.4158     0.1599        370        640:  19%|█▉        | 153/809 [03:03<13:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.299     0.4172     0.1594        178        640:  19%|█▉        | 154/809 [03:04<13:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4176     0.1588        212        640:  19%|█▉        | 156/809 [03:06<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4172     0.1584        352        640:  20%|█▉        | 158/809 [03:09<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4165     0.1584        413        640:  20%|█▉        | 159/809 [03:10<12:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4158     0.1579        258        640:  20%|█▉        | 161/809 [03:12<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3      0.416     0.1574        130        640:  20%|██        | 163/809 [03:15<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G        1.3     0.4158     0.1572        337        640:  20%|██        | 164/809 [03:16<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4169     0.1567        229        640:  21%|██        | 166/809 [03:18<12:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.297     0.4172     0.1563        218        640:  21%|██        | 168/809 [03:21<12:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.298     0.4169     0.1564        289        640:  21%|██        | 170/809 [03:23<12:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.293     0.4184     0.1559         71        640:  21%|██        | 171/809 [03:24<12:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.295     0.4179      0.157        317        640:  21%|██▏       | 173/809 [03:27<13:08,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.295     0.4181     0.1569        271        640:  22%|██▏       | 175/809 [03:29<12:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.294     0.4185     0.1567        103        640:  22%|██▏       | 176/809 [03:30<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.291     0.4191     0.1561        300        640:  22%|██▏       | 178/809 [03:33<12:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.289     0.4194     0.1554        241        640:  22%|██▏       | 180/809 [03:35<12:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.292     0.4188     0.1562        333        640:  22%|██▏       | 181/809 [03:37<12:57,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.289     0.4197     0.1556        207        640:  23%|██▎       | 183/809 [03:39<12:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.289     0.4196     0.1553        324        640:  23%|██▎       | 185/809 [03:41<12:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.288     0.4199     0.1549        240        640:  23%|██▎       | 187/809 [03:44<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.289     0.4197     0.1558        636        640:  23%|██▎       | 188/809 [03:45<12:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.292     0.4191     0.1572        473        640:  23%|██▎       | 190/809 [03:47<12:47,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.294     0.4185     0.1579        443        640:  24%|██▎       | 191/809 [03:49<12:49,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.29     0.4198     0.1571        191        640:  24%|██▍       | 193/809 [03:51<12:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.288     0.4201     0.1567        247        640:  24%|██▍       | 195/809 [03:53<12:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.287     0.4204     0.1564        183        640:  24%|██▍       | 196/809 [03:55<12:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.287     0.4201     0.1563        258        640:  24%|██▍       | 198/809 [03:57<12:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.285     0.4206     0.1558        139        640:  25%|██▍       | 200/809 [03:59<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.284     0.4208     0.1555        185        640:  25%|██▍       | 201/809 [04:01<11:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.286     0.4202     0.1566        712        640:  25%|██▌       | 203/809 [04:03<12:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.285     0.4205     0.1562        260        640:  25%|██▌       | 205/809 [04:05<12:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.285     0.4203     0.1561        126        640:  25%|██▌       | 206/809 [04:07<12:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.287     0.4194     0.1558        155        640:  26%|██▌       | 208/809 [04:09<12:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.288     0.4187     0.1554        265        640:  26%|██▌       | 210/809 [04:11<11:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.288     0.4183     0.1552        161        640:  26%|██▌       | 211/809 [04:13<11:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.287     0.4179      0.155        139        640:  26%|██▋       | 213/809 [04:15<11:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.284     0.4187     0.1544        237        640:  27%|██▋       | 215/809 [04:17<11:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.281     0.4194     0.1539        260        640:  27%|██▋       | 217/809 [04:20<11:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.278       0.42     0.1535        215        640:  27%|██▋       | 218/809 [04:21<11:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.275      0.421     0.1531        202        640:  27%|██▋       | 220/809 [04:23<11:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.273     0.4216     0.1531         96        640:  27%|██▋       | 222/809 [04:25<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.272      0.422     0.1529        144        640:  28%|██▊       | 223/809 [04:27<11:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.269      0.423     0.1523        232        640:  28%|██▊       | 225/809 [04:29<11:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.268     0.4234     0.1519        239        640:  28%|██▊       | 227/809 [04:31<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.268     0.4234     0.1519        209        640:  28%|██▊       | 228/809 [04:33<11:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.266     0.4236     0.1515        207        640:  28%|██▊       | 230/809 [04:35<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.263     0.4248     0.1511        138        640:  29%|██▊       | 232/809 [04:37<11:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.26     0.4253     0.1506        166        640:  29%|██▉       | 234/809 [04:40<11:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.259     0.4258     0.1503        100        640:  29%|██▉       | 235/809 [04:41<11:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.257     0.4263     0.1499        217        640:  29%|██▉       | 237/809 [04:43<11:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.253     0.4283     0.1493         80        640:  30%|██▉       | 239/809 [04:46<11:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.251     0.4288      0.149        244        640:  30%|██▉       | 240/809 [04:47<11:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.248     0.4294     0.1486         93        640:  30%|██▉       | 242/809 [04:49<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.247       0.43     0.1485        384        640:  30%|███       | 244/809 [04:52<11:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.245     0.4304     0.1482        163        640:  30%|███       | 245/809 [04:53<11:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4306     0.1479        138        640:  31%|███       | 247/809 [04:55<11:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242     0.4309     0.1475        183        640:  31%|███       | 249/809 [04:57<10:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.241     0.4309     0.1475        198        640:  31%|███       | 251/809 [05:00<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242     0.4306     0.1476        465        640:  31%|███       | 252/809 [05:01<10:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242     0.4304     0.1477        375        640:  31%|███▏      | 254/809 [05:03<11:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.241     0.4304     0.1475        149        640:  32%|███▏      | 255/809 [05:05<10:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242     0.4301      0.148        176        640:  32%|███▏      | 257/809 [05:07<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4298     0.1482        192        640:  32%|███▏      | 259/809 [05:09<10:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.244     0.4294     0.1483        271        640:  32%|███▏      | 260/809 [05:11<11:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.245      0.429     0.1487        445        640:  32%|███▏      | 262/809 [05:13<11:31,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.244     0.4291     0.1487        449        640:  33%|███▎      | 264/809 [05:16<11:14,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4293     0.1485        166        640:  33%|███▎      | 265/809 [05:17<11:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4291     0.1488        344        640:  33%|███▎      | 267/809 [05:19<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4289     0.1485        351        640:  33%|███▎      | 269/809 [05:22<10:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.244     0.4286     0.1486        248        640:  33%|███▎      | 270/809 [05:23<10:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4286     0.1483         92        640:  34%|███▎      | 272/809 [05:25<10:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4285     0.1483        178        640:  34%|███▍      | 274/809 [05:28<10:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242     0.4288     0.1481        132        640:  34%|███▍      | 276/809 [05:30<10:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.242      0.429      0.148        186        640:  34%|███▍      | 277/809 [05:31<10:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4287     0.1485        216        640:  34%|███▍      | 279/809 [05:34<10:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4287     0.1486        363        640:  35%|███▍      | 280/809 [05:35<10:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.243     0.4287     0.1496        425        640:  35%|███▍      | 282/809 [05:37<10:57,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.245     0.4281     0.1518        364        640:  35%|███▌      | 284/809 [05:40<11:00,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.244     0.4282     0.1516        247        640:  35%|███▌      | 285/809 [05:41<10:57,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.241     0.4294     0.1514         97        640:  35%|███▌      | 287/809 [05:43<10:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.24     0.4299     0.1512        176        640:  36%|███▌      | 289/809 [05:46<10:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4304      0.151        141        640:  36%|███▌      | 290/809 [05:47<10:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4304     0.1508        225        640:  36%|███▌      | 292/809 [05:49<10:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4302     0.1512        221        640:  36%|███▋      | 294/809 [05:52<10:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239       0.43     0.1514        491        640:  36%|███▋      | 295/809 [05:53<10:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4309      0.151        153        640:  37%|███▋      | 297/809 [05:55<10:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4304     0.1516        316        640:  37%|███▋      | 299/809 [05:58<10:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4307     0.1514        186        640:  37%|███▋      | 300/809 [05:59<10:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4314      0.151        291        640:  37%|███▋      | 302/809 [06:01<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4317     0.1507        249        640:  38%|███▊      | 304/809 [06:04<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4315     0.1506        187        640:  38%|███▊      | 305/809 [06:05<10:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4313     0.1506        254        640:  38%|███▊      | 307/809 [06:07<09:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4315     0.1504        430        640:  38%|███▊      | 309/809 [06:10<10:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4313     0.1503        240        640:  38%|███▊      | 310/809 [06:11<09:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4317     0.1499        240        640:  39%|███▊      | 312/809 [06:13<09:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4314       0.15        200        640:  39%|███▉      | 314/809 [06:16<09:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4313       0.15        344        640:  39%|███▉      | 316/809 [06:18<09:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4311     0.1498        232        640:  39%|███▉      | 318/809 [06:20<09:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4309     0.1497        139        640:  39%|███▉      | 319/809 [06:22<09:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4305     0.1496        125        640:  40%|███▉      | 321/809 [06:24<09:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237       0.43     0.1497        159        640:  40%|███▉      | 323/809 [06:26<09:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4301     0.1495        110        640:  40%|████      | 324/809 [06:27<09:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4303     0.1494        208        640:  40%|████      | 326/809 [06:30<09:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4304       0.15        498        640:  41%|████      | 328/809 [06:32<09:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4305     0.1498        171        640:  41%|████      | 329/809 [06:33<09:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4312     0.1495         99        640:  41%|████      | 331/809 [06:36<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4313     0.1493        303        640:  41%|████      | 333/809 [06:38<09:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4313     0.1492        304        640:  41%|████▏     | 334/809 [06:39<09:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4312     0.1491        438        640:  42%|████▏     | 336/809 [06:42<09:40,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4307     0.1499        672        640:  42%|████▏     | 338/809 [06:44<09:47,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4306     0.1497        213        640:  42%|████▏     | 339/809 [06:46<09:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4309     0.1495        237        640:  42%|████▏     | 341/809 [06:48<09:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4312     0.1494        515        640:  42%|████▏     | 343/809 [06:50<09:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4314     0.1491        146        640:  43%|████▎     | 344/809 [06:52<09:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4316     0.1489        174        640:  43%|████▎     | 346/809 [06:54<09:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4316     0.1488        231        640:  43%|████▎     | 348/809 [06:56<09:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4315     0.1489        433        640:  43%|████▎     | 349/809 [06:58<09:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4315      0.149        402        640:  43%|████▎     | 351/809 [07:00<09:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4315     0.1489        197        640:  44%|████▎     | 353/809 [07:02<09:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229      0.432     0.1485        151        640:  44%|████▍     | 355/809 [07:05<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228      0.432     0.1484        169        640:  44%|████▍     | 356/809 [07:06<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4316     0.1485        108        640:  44%|████▍     | 358/809 [07:08<08:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4311     0.1489        119        640:  44%|████▍     | 360/809 [07:11<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4307      0.149        263        640:  45%|████▍     | 361/809 [07:12<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4307     0.1489        205        640:  45%|████▍     | 363/809 [07:14<08:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4303     0.1488        253        640:  45%|████▌     | 365/809 [07:17<08:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4302     0.1487        312        640:  45%|████▌     | 366/809 [07:18<08:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4298     0.1487        414        640:  45%|████▌     | 368/809 [07:20<08:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4294       0.15        455        640:  46%|████▌     | 370/809 [07:23<09:05,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234      0.429     0.1502        346        640:  46%|████▌     | 371/809 [07:24<09:04,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4293     0.1501        153        640:  46%|████▌     | 373/809 [07:26<09:01,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4295     0.1499        154        640:  46%|████▌     | 374/809 [07:28<08:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4293     0.1499        223        640:  46%|████▋     | 376/809 [07:30<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4292     0.1498        148        640:  47%|████▋     | 378/809 [07:32<08:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4288     0.1498        193        640:  47%|████▋     | 380/809 [07:35<08:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4288     0.1498         95        640:  47%|████▋     | 381/809 [07:36<08:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4286     0.1501        272        640:  47%|████▋     | 383/809 [07:38<08:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4285     0.1502        216        640:  48%|████▊     | 385/809 [07:41<08:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4287     0.1502        323        640:  48%|████▊     | 386/809 [07:42<08:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232      0.429       0.15        230        640:  48%|████▊     | 388/809 [07:44<08:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4288     0.1501        280        640:  48%|████▊     | 390/809 [07:47<08:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4285     0.1502        339        640:  48%|████▊     | 391/809 [07:48<08:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4283     0.1502        282        640:  49%|████▊     | 393/809 [07:50<08:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234      0.428     0.1507        485        640:  49%|████▉     | 395/809 [07:53<08:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234      0.428     0.1506        172        640:  49%|████▉     | 396/809 [07:54<08:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4282     0.1504        172        640:  49%|████▉     | 398/809 [07:56<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4284     0.1502        150        640:  49%|████▉     | 400/809 [07:58<07:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4286     0.1501        275        640:  50%|████▉     | 402/809 [08:01<07:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4285       0.15        177        640:  50%|████▉     | 403/809 [08:02<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4284     0.1503        322        640:  50%|█████     | 405/809 [08:04<08:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4285     0.1502        227        640:  50%|█████     | 407/809 [08:07<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4285       0.15        200        640:  50%|█████     | 408/809 [08:08<07:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4281     0.1513        394        640:  51%|█████     | 410/809 [08:10<08:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4281     0.1513        270        640:  51%|█████     | 412/809 [08:13<07:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4279     0.1517        459        640:  51%|█████     | 413/809 [08:14<08:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234      0.428     0.1516        218        640:  51%|█████▏    | 415/809 [08:16<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4282     0.1513        164        640:  52%|█████▏    | 417/809 [08:19<07:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4282     0.1513        189        640:  52%|█████▏    | 418/809 [08:20<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4276     0.1515        285        640:  52%|█████▏    | 420/809 [08:22<07:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4279     0.1513        170        640:  52%|█████▏    | 422/809 [08:25<07:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4279     0.1511        186        640:  52%|█████▏    | 423/809 [08:26<07:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4275     0.1514        339        640:  53%|█████▎    | 425/809 [08:28<07:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4274     0.1519        408        640:  53%|█████▎    | 427/809 [08:31<07:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4273     0.1518        263        640:  53%|█████▎    | 428/809 [08:32<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4272     0.1516        167        640:  53%|█████▎    | 430/809 [08:34<07:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4272     0.1519        251        640:  53%|█████▎    | 432/809 [08:37<07:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4271     0.1519        230        640:  54%|█████▎    | 433/809 [08:38<07:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4271      0.152        401        640:  54%|█████▍    | 435/809 [08:40<07:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4273     0.1519        251        640:  54%|█████▍    | 437/809 [08:43<07:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4273     0.1519        206        640:  54%|█████▍    | 438/809 [08:44<07:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4269     0.1525        540        640:  54%|█████▍    | 440/809 [08:46<07:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4273     0.1523        127        640:  55%|█████▍    | 442/809 [08:49<07:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4274     0.1521        165        640:  55%|█████▍    | 443/809 [08:50<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234      0.428     0.1518        171        640:  55%|█████▌    | 446/809 [08:53<07:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4278     0.1522        358        640:  55%|█████▌    | 447/809 [08:55<07:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4278     0.1522        243        640:  56%|█████▌    | 449/809 [08:57<07:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4278      0.152        183        640:  56%|█████▌    | 451/809 [08:59<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4278      0.152        339        640:  56%|█████▌    | 452/809 [09:01<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4276     0.1521        392        640:  56%|█████▌    | 454/809 [09:03<07:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4276     0.1519        188        640:  56%|█████▋    | 456/809 [09:05<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4275     0.1518        124        640:  56%|█████▋    | 457/809 [09:07<07:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4271      0.152        149        640:  57%|█████▋    | 459/809 [09:09<07:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4269     0.1519        111        640:  57%|█████▋    | 461/809 [09:11<06:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4268      0.152        281        640:  57%|█████▋    | 462/809 [09:13<06:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4271     0.1518        292        640:  57%|█████▋    | 464/809 [09:15<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4272     0.1516        189        640:  58%|█████▊    | 466/809 [09:17<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4273     0.1515        280        640:  58%|█████▊    | 467/809 [09:18<06:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4272     0.1514        273        640:  58%|█████▊    | 469/809 [09:21<06:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4269     0.1523        438        640:  58%|█████▊    | 471/809 [09:23<06:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.239     0.4269     0.1523        291        640:  58%|█████▊    | 472/809 [09:25<06:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.238     0.4271     0.1522        239        640:  59%|█████▊    | 474/809 [09:27<06:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4274     0.1519        193        640:  59%|█████▉    | 476/809 [09:29<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236      0.428     0.1518         42        640:  59%|█████▉    | 477/809 [09:30<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235      0.428     0.1517        251        640:  59%|█████▉    | 479/809 [09:33<06:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4279     0.1519        153        640:  59%|█████▉    | 481/809 [09:35<06:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4278     0.1521        314        640:  60%|█████▉    | 482/809 [09:36<06:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4283     0.1519        274        640:  60%|█████▉    | 484/809 [09:39<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4284     0.1519         68        640:  60%|██████    | 486/809 [09:41<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4287     0.1517        298        640:  60%|██████    | 488/809 [09:44<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4287     0.1517        203        640:  60%|██████    | 489/809 [09:45<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233      0.429     0.1516        225        640:  61%|██████    | 491/809 [09:47<06:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232      0.429     0.1514        253        640:  61%|██████    | 493/809 [09:49<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232      0.429     0.1513        273        640:  61%|██████    | 494/809 [09:51<06:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4293     0.1512        167        640:  61%|██████▏   | 496/809 [09:53<06:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4294     0.1511        224        640:  62%|██████▏   | 498/809 [09:55<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4295      0.151        320        640:  62%|██████▏   | 499/809 [09:58<06:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4295     0.1509        136        640:  62%|██████▏   | 501/809 [09:59<06:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4295     0.1509        123        640:  62%|██████▏   | 503/809 [10:01<06:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4299     0.1508        108        640:  62%|██████▏   | 505/809 [10:04<05:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4302     0.1506        133        640:  63%|██████▎   | 506/809 [10:05<05:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4301     0.1506        201        640:  63%|██████▎   | 508/809 [10:07<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4303     0.1504        199        640:  63%|██████▎   | 510/809 [10:09<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4302     0.1505        281        640:  63%|██████▎   | 511/809 [10:11<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4296     0.1507        571        640:  63%|██████▎   | 513/809 [10:13<06:04,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4296     0.1506        143        640:  64%|██████▎   | 515/809 [10:15<05:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4297     0.1505        158        640:  64%|██████▍   | 516/809 [10:17<05:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4296     0.1506        439        640:  64%|██████▍   | 518/809 [10:19<05:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4296     0.1506        198        640:  64%|██████▍   | 520/809 [10:22<05:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4295     0.1506        188        640:  64%|██████▍   | 521/809 [10:23<05:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4293      0.151        251        640:  65%|██████▍   | 523/809 [10:25<05:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4292     0.1511        246        640:  65%|██████▍   | 525/809 [10:28<05:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4291     0.1513        347        640:  65%|██████▌   | 526/809 [10:29<05:50,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4291     0.1515        446        640:  65%|██████▌   | 528/809 [10:31<05:47,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4289     0.1515        279        640:  65%|██████▌   | 529/809 [10:33<05:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4292     0.1515        210        640:  66%|██████▌   | 531/809 [10:35<05:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4293     0.1514        128        640:  66%|██████▌   | 533/809 [10:37<05:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4295     0.1513        126        640:  66%|██████▌   | 534/809 [10:39<05:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4298     0.1511        293        640:  66%|██████▋   | 536/809 [10:41<05:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4297     0.1511        323        640:  67%|██████▋   | 538/809 [10:44<05:39,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4296      0.151        272        640:  67%|██████▋   | 539/809 [10:45<05:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229       0.43     0.1509        178        640:  67%|██████▋   | 541/809 [10:47<05:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4299      0.151        379        640:  67%|██████▋   | 543/809 [10:49<05:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4297     0.1512        260        640:  67%|██████▋   | 544/809 [10:51<05:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4295     0.1512        313        640:  67%|██████▋   | 546/809 [10:53<05:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4294     0.1512        307        640:  68%|██████▊   | 548/809 [10:55<05:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4293     0.1513        309        640:  68%|██████▊   | 549/809 [10:57<05:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4293     0.1513        197        640:  68%|██████▊   | 551/809 [10:59<05:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4291     0.1512        347        640:  68%|██████▊   | 553/809 [11:02<05:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4293     0.1511        172        640:  68%|██████▊   | 554/809 [11:03<05:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4297     0.1509        192        640:  69%|██████▊   | 556/809 [11:05<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4298     0.1516        302        640:  69%|██████▉   | 558/809 [11:08<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228       0.43     0.1514        193        640:  69%|██████▉   | 559/809 [11:09<05:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228     0.4299     0.1514        406        640:  69%|██████▉   | 561/809 [11:11<04:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4301     0.1513        139        640:  70%|██████▉   | 563/809 [11:13<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227       0.43     0.1513        240        640:  70%|██████▉   | 564/809 [11:15<04:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4301     0.1512        237        640:  70%|██████▉   | 566/809 [11:17<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4301     0.1511        377        640:  70%|███████   | 568/809 [11:19<04:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4302     0.1514        370        640:  70%|███████   | 570/809 [11:22<04:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4302     0.1515        366        640:  71%|███████   | 571/809 [11:23<04:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4303     0.1513        293        640:  71%|███████   | 573/809 [11:25<04:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4303     0.1513        264        640:  71%|███████   | 575/809 [11:28<04:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4302     0.1514        346        640:  71%|███████   | 576/809 [11:29<04:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227       0.43     0.1514        273        640:  71%|███████▏  | 578/809 [11:31<04:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4299     0.1513        294        640:  72%|███████▏  | 580/809 [11:34<04:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228     0.4297     0.1514        378        640:  72%|███████▏  | 581/809 [11:35<04:42,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4295     0.1519        629        640:  72%|███████▏  | 583/809 [11:38<04:42,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4292     0.1522        445        640:  72%|███████▏  | 584/809 [11:39<04:40,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231      0.429     0.1524        379        640:  72%|███████▏  | 586/809 [11:41<04:36,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4288     0.1526        290        640:  73%|███████▎  | 588/809 [11:44<04:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4287      0.153        391        640:  73%|███████▎  | 589/809 [11:45<04:37,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4287     0.1528        295        640:  73%|███████▎  | 591/809 [11:47<04:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4285      0.153        454        640:  73%|███████▎  | 593/809 [11:50<04:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4283      0.153        450        640:  73%|███████▎  | 594/809 [11:51<04:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4283     0.1528        176        640:  74%|███████▎  | 596/809 [11:53<04:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4279     0.1531        464        640:  74%|███████▍  | 598/809 [11:56<04:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4281     0.1531        193        640:  74%|███████▍  | 599/809 [11:57<04:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4283      0.153        324        640:  74%|███████▍  | 601/809 [11:59<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4282      0.153        256        640:  75%|███████▍  | 603/809 [12:02<04:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235      0.428      0.153        336        640:  75%|███████▍  | 605/809 [12:04<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4279      0.153        256        640:  75%|███████▍  | 606/809 [12:05<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235      0.428     0.1529        266        640:  75%|███████▌  | 608/809 [12:08<03:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4282     0.1528        126        640:  75%|███████▌  | 610/809 [12:10<03:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4279     0.1527        277        640:  76%|███████▌  | 612/809 [12:13<03:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4278     0.1527        259        640:  76%|███████▌  | 613/809 [12:14<03:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4281     0.1525        158        640:  76%|███████▌  | 615/809 [12:16<03:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4282     0.1524        322        640:  76%|███████▋  | 617/809 [12:18<03:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4281     0.1524        441        640:  76%|███████▋  | 618/809 [12:20<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4281     0.1522        453        640:  77%|███████▋  | 620/809 [12:22<03:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4285     0.1521        220        640:  77%|███████▋  | 622/809 [12:24<03:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4283      0.152        351        640:  77%|███████▋  | 623/809 [12:26<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4283      0.152        127        640:  77%|███████▋  | 625/809 [12:28<03:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4283     0.1521        339        640:  78%|███████▊  | 627/809 [12:30<03:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4281     0.1521        258        640:  78%|███████▊  | 628/809 [12:31<03:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4279      0.152        298        640:  78%|███████▊  | 630/809 [12:34<03:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4283     0.1519        153        640:  78%|███████▊  | 632/809 [12:36<03:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4283     0.1519        346        640:  78%|███████▊  | 633/809 [12:38<03:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4284     0.1521        457        640:  78%|███████▊  | 635/809 [12:40<03:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4286      0.152        268        640:  79%|███████▊  | 637/809 [12:42<03:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4286      0.152        335        640:  79%|███████▉  | 638/809 [12:44<03:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4285     0.1524        134        640:  79%|███████▉  | 640/809 [12:46<03:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4284     0.1524        250        640:  79%|███████▉  | 641/809 [12:47<03:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4283     0.1524        340        640:  79%|███████▉  | 643/809 [12:50<03:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4283     0.1523         75        640:  80%|███████▉  | 645/809 [12:52<03:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.237     0.4283     0.1522        263        640:  80%|███████▉  | 647/809 [12:54<03:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4284     0.1521        157        640:  80%|████████  | 648/809 [12:56<03:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.236     0.4284     0.1522        115        640:  80%|████████  | 650/809 [12:58<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4287      0.152        155        640:  81%|████████  | 652/809 [13:00<03:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4289     0.1519        179        640:  81%|████████  | 653/809 [13:02<03:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4287     0.1521        505        640:  81%|████████  | 655/809 [13:04<03:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4287     0.1522        514        640:  81%|████████  | 657/809 [13:07<03:09,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.235     0.4288     0.1521        161        640:  81%|████████▏ | 658/809 [13:08<03:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.234     0.4291     0.1519        122        640:  82%|████████▏ | 660/809 [13:10<02:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.233     0.4294     0.1517        173        640:  82%|████████▏ | 662/809 [13:12<02:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4296     0.1517        176        640:  82%|████████▏ | 663/809 [13:14<02:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232       0.43     0.1515        165        640:  82%|████████▏ | 665/809 [13:16<02:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4302     0.1514        152        640:  82%|████████▏ | 667/809 [13:18<02:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4303     0.1512        298        640:  83%|████████▎ | 669/809 [13:21<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4303     0.1512        250        640:  83%|████████▎ | 670/809 [13:22<02:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4303     0.1511        246        640:  83%|████████▎ | 672/809 [13:24<02:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4301     0.1511        289        640:  83%|████████▎ | 674/809 [13:27<02:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232       0.43     0.1511        226        640:  83%|████████▎ | 675/809 [13:28<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232       0.43      0.151        223        640:  84%|████████▎ | 677/809 [13:30<02:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4302     0.1509        185        640:  84%|████████▍ | 679/809 [13:32<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4302     0.1509        137        640:  84%|████████▍ | 680/809 [13:34<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4301     0.1513        279        640:  84%|████████▍ | 682/809 [13:36<02:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232       0.43     0.1514        532        640:  85%|████████▍ | 684/809 [13:39<02:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.232     0.4301     0.1514        123        640:  85%|████████▍ | 685/809 [13:40<02:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4303     0.1512         89        640:  85%|████████▍ | 687/809 [13:42<02:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4303     0.1511        183        640:  85%|████████▌ | 689/809 [13:44<02:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4303      0.151        251        640:  85%|████████▌ | 690/809 [13:46<02:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.231     0.4302     0.1509        248        640:  86%|████████▌ | 692/809 [13:48<02:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4302     0.1508        204        640:  86%|████████▌ | 694/809 [13:50<02:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.23     0.4304     0.1507        127        640:  86%|████████▌ | 696/809 [13:53<02:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4304     0.1506        199        640:  86%|████████▌ | 697/809 [13:54<02:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4305     0.1505         91        640:  86%|████████▋ | 699/809 [13:56<02:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.229     0.4305     0.1504        131        640:  87%|████████▋ | 701/809 [13:59<02:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228     0.4306     0.1504        115        640:  87%|████████▋ | 702/809 [14:00<02:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.228     0.4308     0.1502        158        640:  87%|████████▋ | 704/809 [14:02<02:10,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227      0.431     0.1501        213        640:  87%|████████▋ | 705/809 [14:04<02:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.227     0.4312     0.1501        189        640:  87%|████████▋ | 707/809 [14:06<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4314       0.15        386        640:  88%|████████▊ | 709/809 [14:08<02:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.226     0.4316     0.1499        126        640:  88%|████████▊ | 711/809 [14:11<01:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.225     0.4317     0.1499        299        640:  88%|████████▊ | 712/809 [14:12<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.225     0.4319     0.1497        160        640:  88%|████████▊ | 714/809 [14:14<01:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.225     0.4318     0.1501        487        640:  89%|████████▊ | 716/809 [14:17<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.225     0.4318       0.15        266        640:  89%|████████▊ | 717/809 [14:18<01:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.224      0.432     0.1499        206        640:  89%|████████▉ | 719/809 [14:20<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.224     0.4321     0.1497        320        640:  89%|████████▉ | 721/809 [14:23<01:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.224     0.4322     0.1497        262        640:  89%|████████▉ | 722/809 [14:24<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4325     0.1496        167        640:  89%|████████▉ | 724/809 [14:26<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1494         96        640:  90%|████████▉ | 726/809 [14:29<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1494        238        640:  90%|████████▉ | 727/809 [14:30<01:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.221     0.4329     0.1497        127        640:  90%|█████████ | 729/809 [14:32<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.221     0.4332     0.1496        185        640:  90%|█████████ | 731/809 [14:35<01:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4334     0.1495        117        640:  90%|█████████ | 732/809 [14:36<01:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4337     0.1493        155        640:  91%|█████████ | 734/809 [14:38<01:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4336     0.1493        202        640:  91%|█████████ | 736/809 [14:40<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4338     0.1491        288        640:  91%|█████████ | 738/809 [14:43<01:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4336     0.1494        449        640:  91%|█████████▏| 739/809 [14:44<01:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.218      0.434     0.1492         94        640:  92%|█████████▏| 741/809 [14:47<01:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4338     0.1494        350        640:  92%|█████████▏| 742/809 [14:48<01:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4334     0.1495        444        640:  92%|█████████▏| 744/809 [14:50<01:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.219     0.4336     0.1494        230        640:  92%|█████████▏| 746/809 [14:52<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4335     0.1493        206        640:  92%|█████████▏| 748/809 [14:55<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4335     0.1493        205        640:  93%|█████████▎| 749/809 [14:56<01:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4335     0.1493        334        640:  93%|█████████▎| 751/809 [14:58<01:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4334     0.1495        364        640:  93%|█████████▎| 753/809 [15:01<01:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4333     0.1494        162        640:  93%|█████████▎| 755/809 [15:03<01:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G       1.22     0.4332     0.1494        184        640:  93%|█████████▎| 756/809 [15:04<01:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4328     0.1495        424        640:  94%|█████████▎| 758/809 [15:07<01:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4327     0.1495        447        640:  94%|█████████▍| 760/809 [15:09<00:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.221     0.4329     0.1495         85        640:  94%|█████████▍| 761/809 [15:10<00:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4326     0.1495        339        640:  94%|█████████▍| 763/809 [15:13<00:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4326     0.1494        246        640:  95%|█████████▍| 765/809 [15:15<00:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4325     0.1495        320        640:  95%|█████████▍| 766/809 [15:17<00:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4326     0.1495        171        640:  95%|█████████▍| 768/809 [15:19<00:50,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4325     0.1497        733        640:  95%|█████████▌| 769/809 [15:20<00:50,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4326     0.1495        316        640:  95%|█████████▌| 771/809 [15:23<00:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4324     0.1497        420        640:  96%|█████████▌| 773/809 [15:25<00:44,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4325     0.1497        283        640:  96%|█████████▌| 774/809 [15:27<00:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4326     0.1496        253        640:  96%|█████████▌| 776/809 [15:29<00:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4328     0.1494        412        640:  96%|█████████▌| 778/809 [15:31<00:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1493        194        640:  96%|█████████▋| 779/809 [15:34<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1494        225        640:  97%|█████████▋| 781/809 [15:35<00:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222      0.433     0.1492        176        640:  97%|█████████▋| 783/809 [15:37<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1493        380        640:  97%|█████████▋| 785/809 [15:39<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222      0.433     0.1492        294        640:  97%|█████████▋| 786/809 [15:41<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222      0.433     0.1491        345        640:  97%|█████████▋| 788/809 [15:43<00:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1491        323        640:  98%|█████████▊| 790/809 [15:45<00:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329     0.1491        237        640:  98%|█████████▊| 791/809 [15:47<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4329      0.149        138        640:  98%|█████████▊| 793/809 [15:49<00:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4328     0.1489        263        640:  98%|█████████▊| 795/809 [15:51<00:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4326     0.1489        392        640:  98%|█████████▊| 796/809 [15:52<00:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4326     0.1489        425        640:  99%|█████████▊| 798/809 [15:55<00:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4327     0.1488        304        640:  99%|█████████▉| 800/809 [15:57<00:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4326     0.1488        412        640:  99%|█████████▉| 801/809 [15:59<00:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4326     0.1489        202        640:  99%|█████████▉| 803/809 [16:01<00:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.222     0.4327     0.1489        230        640: 100%|█████████▉| 805/809 [16:03<00:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4326     0.1488        427        640: 100%|█████████▉| 807/809 [16:06<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     37/300      13.8G      1.223     0.4325     0.1489        306        640: 100%|█████████▉| 808/809 [16:07<00:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.03it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  54%|█████▎    | 37/69 [00:09<00:08,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  65%|██████▌   | 45/69 [00:11<00:06,  3.67it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:05,  3.55it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward passinit__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward passinit__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.7G      1.237     0.4433     0.1352         90        640:   0%|          | 2/809 [00:02<15:37,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.288     0.4044     0.1285        280        640:   0%|          | 4/809 [00:04<16:00,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4423      0.116        271        640:   1%|          | 6/809 [00:07<15:41,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4459     0.1166        185        640:   1%|          | 7/809 [00:08<15:48,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.158     0.4531       0.11        196        640:   1%|          | 9/809 [00:10<15:34,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.126     0.4591     0.1068        177        640:   1%|▏         | 11/809 [00:12<15:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.14      0.456     0.1092        324        640:   1%|▏         | 12/809 [00:14<16:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4401     0.1147        479        640:   2%|▏         | 14/809 [00:16<15:55,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.213     0.4338     0.1175        291        640:   2%|▏         | 16/809 [00:18<15:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4405     0.1153         68        640:   2%|▏         | 17/809 [00:20<15:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.222     0.4346     0.1189        209        640:   2%|▏         | 19/809 [00:22<15:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.231     0.4307     0.1202        200        640:   3%|▎         | 21/809 [00:25<15:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.21     0.4358     0.1188        152        640:   3%|▎         | 22/809 [00:26<15:46,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4404     0.1163        222        640:   3%|▎         | 24/809 [00:28<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4434     0.1143        220        640:   3%|▎         | 26/809 [00:30<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.176     0.4462     0.1131        245        640:   3%|▎         | 27/809 [00:31<15:08,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4466     0.1115        183        640:   4%|▎         | 29/809 [00:34<15:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.169     0.4485     0.1111        228        640:   4%|▍         | 31/809 [00:36<15:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.154     0.4539     0.1095        113        640:   4%|▍         | 33/809 [00:39<15:03,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.148     0.4557     0.1086        188        640:   4%|▍         | 34/809 [00:40<15:01,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.148     0.4548     0.1075        270        640:   4%|▍         | 36/809 [00:42<15:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.144     0.4551     0.1085        228        640:   5%|▍         | 38/809 [00:45<15:35,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.149     0.4522     0.1092        377        640:   5%|▍         | 39/809 [00:46<15:29,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.153     0.4505     0.1092        342        640:   5%|▌         | 41/809 [00:48<15:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.16     0.4484     0.1104        546        640:   5%|▌         | 43/809 [00:51<15:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.155     0.4492     0.1096        255        640:   5%|▌         | 44/809 [00:52<15:23,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.173     0.4446     0.1163        305        640:   6%|▌         | 46/809 [00:54<15:13,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.167     0.4465      0.115        249        640:   6%|▌         | 48/809 [00:57<15:12,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.173     0.4455     0.1171        421        640:   6%|▌         | 49/809 [00:58<15:25,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4474     0.1164         71        640:   6%|▋         | 51/809 [01:00<15:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4484     0.1158        222        640:   7%|▋         | 53/809 [01:02<15:00,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.163     0.4494     0.1153        268        640:   7%|▋         | 54/809 [01:04<14:50,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4484     0.1153        216        640:   7%|▋         | 56/809 [01:06<14:49,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4487     0.1148        162        640:   7%|▋         | 58/809 [01:08<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4485     0.1145        211        640:   7%|▋         | 59/809 [01:10<14:49,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.169     0.4473     0.1153        337        640:   8%|▊         | 61/809 [01:12<14:59,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.167     0.4483     0.1158        162        640:   8%|▊         | 63/809 [01:14<14:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.161       0.45     0.1152        234        640:   8%|▊         | 65/809 [01:17<14:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4487     0.1159        408        640:   8%|▊         | 66/809 [01:18<14:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.167     0.4482     0.1157        239        640:   8%|▊         | 68/809 [01:20<14:47,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4474      0.116        172        640:   9%|▊         | 70/809 [01:23<14:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4485     0.1157        218        640:   9%|▉         | 71/809 [01:24<14:30,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.162     0.4505     0.1148        192        640:   9%|▉         | 73/809 [01:26<14:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4501     0.1172        179        640:   9%|▉         | 75/809 [01:29<14:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4506     0.1177        291        640:   9%|▉         | 76/809 [01:30<14:50,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.169     0.4488     0.1201        320        640:  10%|▉         | 78/809 [01:32<14:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179      0.446     0.1222        209        640:  10%|▉         | 80/809 [01:35<14:44,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4458     0.1228        297        640:  10%|█         | 81/809 [01:36<14:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.187     0.4433     0.1233        146        640:  10%|█         | 83/809 [01:38<14:31,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4446     0.1229        293        640:  11%|█         | 85/809 [01:41<14:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4447     0.1229        177        640:  11%|█         | 86/809 [01:42<14:40,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.175     0.4469     0.1221        222        640:  11%|█         | 88/809 [01:44<14:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172      0.448     0.1215        197        640:  11%|█         | 90/809 [01:47<14:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4473     0.1214        331        640:  11%|█▏        | 92/809 [01:49<14:34,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4475     0.1211        246        640:  11%|█▏        | 93/809 [01:50<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4478     0.1206        164        640:  12%|█▏        | 95/809 [01:53<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4475     0.1243        247        640:  12%|█▏        | 97/809 [01:55<14:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4474     0.1241        265        640:  12%|█▏        | 99/809 [01:57<13:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.175     0.4467     0.1243        376        640:  12%|█▏        | 100/809 [01:59<14:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.173     0.4472     0.1241        485        640:  13%|█▎        | 102/809 [02:01<14:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4484     0.1237        145        640:  13%|█▎        | 103/809 [02:02<14:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.169     0.4492     0.1236        294        640:  13%|█▎        | 105/809 [02:05<14:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4497     0.1256        386        640:  13%|█▎        | 107/809 [02:07<14:21,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.173     0.4494     0.1276        406        640:  13%|█▎        | 108/809 [02:08<14:31,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4508     0.1274        251        640:  14%|█▎        | 110/809 [02:11<13:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.175     0.4499     0.1279        238        640:  14%|█▍        | 112/809 [02:13<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4487      0.129        549        640:  14%|█▍        | 113/809 [02:14<14:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.177     0.4509     0.1288         80        640:  14%|█▍        | 115/809 [02:17<13:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.175     0.4515     0.1288        329        640:  14%|█▍        | 117/809 [02:19<13:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.174     0.4518     0.1287        151        640:  15%|█▍        | 118/809 [02:20<13:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4543     0.1281        172        640:  15%|█▍        | 120/809 [02:23<13:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4549     0.1277        251        640:  15%|█▌        | 122/809 [02:25<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4551     0.1275        263        640:  15%|█▌        | 123/809 [02:26<13:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4548     0.1272        254        640:  15%|█▌        | 125/809 [02:29<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.167     0.4542     0.1284        437        640:  16%|█▌        | 127/809 [02:31<13:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4538     0.1283        249        640:  16%|█▌        | 128/809 [02:32<13:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171      0.453     0.1287        271        640:  16%|█▌        | 130/809 [02:35<13:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.172     0.4529     0.1296        190        640:  16%|█▋        | 132/809 [02:37<13:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.17     0.4538     0.1296        336        640:  17%|█▋        | 134/809 [02:39<13:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.17     0.4544     0.1296        198        640:  17%|█▋        | 135/809 [02:41<13:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4547     0.1293        249        640:  17%|█▋        | 137/809 [02:43<13:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4551     0.1293        134        640:  17%|█▋        | 139/809 [02:45<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166      0.455     0.1295        401        640:  17%|█▋        | 140/809 [02:47<13:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4551     0.1294        365        640:  18%|█▊        | 142/809 [02:49<13:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4553     0.1304        369        640:  18%|█▊        | 144/809 [02:51<13:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4556       0.13        158        640:  18%|█▊        | 145/809 [02:52<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166     0.4555     0.1304        154        640:  18%|█▊        | 147/809 [02:55<13:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168      0.455     0.1304        195        640:  18%|█▊        | 149/809 [02:57<13:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.165     0.4555       0.13        160        640:  19%|█▊        | 150/809 [02:58<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4555     0.1295        113        640:  19%|█▉        | 152/809 [03:01<12:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4557     0.1295        232        640:  19%|█▉        | 154/809 [03:03<12:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4556     0.1293        334        640:  19%|█▉        | 155/809 [03:04<12:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4545     0.1301        172        640:  19%|█▉        | 157/809 [03:07<13:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4533     0.1303        353        640:  20%|█▉        | 159/809 [03:09<12:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.17     0.4529     0.1302         84        640:  20%|█▉        | 161/809 [03:11<12:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4524     0.1302        181        640:  20%|██        | 162/809 [03:13<12:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.173     0.4512       0.13        277        640:  20%|██        | 164/809 [03:15<12:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.171     0.4517     0.1297        268        640:  21%|██        | 166/809 [03:17<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4522     0.1293         96        640:  21%|██        | 167/809 [03:19<12:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4524     0.1291        357        640:  21%|██        | 169/809 [03:21<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.163     0.4533     0.1286        161        640:  21%|██        | 171/809 [03:23<12:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.161     0.4539     0.1285        147        640:  21%|██▏       | 172/809 [03:24<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.16     0.4539     0.1285        294        640:  22%|██▏       | 174/809 [03:27<12:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.159      0.454     0.1284        370        640:  22%|██▏       | 176/809 [03:29<12:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.159     0.4538     0.1285        298        640:  22%|██▏       | 177/809 [03:31<12:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.156     0.4547     0.1282        148        640:  22%|██▏       | 179/809 [03:33<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.155     0.4552     0.1279        259        640:  22%|██▏       | 181/809 [03:36<13:13,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.153     0.4559     0.1277         70        640:  22%|██▏       | 182/809 [03:37<13:04,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.155     0.4554     0.1276        214        640:  23%|██▎       | 184/809 [03:39<12:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.157     0.4546     0.1276        267        640:  23%|██▎       | 186/809 [03:41<12:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.157     0.4549     0.1275        118        640:  23%|██▎       | 187/809 [03:43<12:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.159     0.4546      0.128        356        640:  23%|██▎       | 189/809 [03:45<12:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4533     0.1291        305        640:  24%|██▎       | 191/809 [03:47<12:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.164     0.4533     0.1291        133        640:  24%|██▎       | 192/809 [03:49<12:40,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166      0.453       0.13        475        640:  24%|██▍       | 194/809 [03:51<12:46,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.166      0.453       0.13        344        640:  24%|██▍       | 195/809 [03:52<12:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.169     0.4519     0.1304        243        640:  24%|██▍       | 197/809 [03:55<12:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.167     0.4526     0.1299        137        640:  25%|██▍       | 199/809 [03:57<11:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168      0.452     0.1301        152        640:  25%|██▍       | 201/809 [03:59<11:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.168     0.4519       0.13        237        640:  25%|██▍       | 202/809 [04:01<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.17     0.4512       0.13        398        640:  25%|██▌       | 204/809 [04:03<12:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.174     0.4501     0.1306        417        640:  25%|██▌       | 206/809 [04:05<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.174     0.4498     0.1308        285        640:  26%|██▌       | 207/809 [04:07<11:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.176     0.4491     0.1319        324        640:  26%|██▌       | 209/809 [04:09<12:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.177     0.4485      0.132        218        640:  26%|██▌       | 211/809 [04:11<11:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.177     0.4485     0.1321        315        640:  26%|██▌       | 212/809 [04:13<12:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.178     0.4484     0.1321        238        640:  26%|██▋       | 214/809 [04:15<11:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4475     0.1324        311        640:  27%|██▋       | 216/809 [04:17<11:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4481     0.1321         89        640:  27%|██▋       | 217/809 [04:19<11:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4472     0.1339        268        640:  27%|██▋       | 219/809 [04:21<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4479     0.1335        167        640:  27%|██▋       | 221/809 [04:24<11:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4484     0.1333        118        640:  27%|██▋       | 222/809 [04:25<11:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4478     0.1337        251        640:  28%|██▊       | 224/809 [04:27<11:56,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4478     0.1334        477        640:  28%|██▊       | 226/809 [04:30<11:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4473     0.1343        454        640:  28%|██▊       | 227/809 [04:31<11:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4475     0.1341        250        640:  28%|██▊       | 229/809 [04:33<11:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4482     0.1339        198        640:  29%|██▊       | 231/809 [04:36<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181      0.448     0.1339        292        640:  29%|██▊       | 232/809 [04:37<11:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.177     0.4488     0.1335        104        640:  29%|██▉       | 234/809 [04:39<11:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.178     0.4486     0.1336        295        640:  29%|██▉       | 236/809 [04:42<11:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4482      0.134        241        640:  29%|██▉       | 238/809 [04:44<11:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18      0.448      0.134        270        640:  30%|██▉       | 239/809 [04:45<11:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4475     0.1342        445        640:  30%|██▉       | 241/809 [04:48<11:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4476     0.1341        183        640:  30%|███       | 243/809 [04:50<11:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4479     0.1339        268        640:  30%|███       | 244/809 [04:51<11:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4482     0.1336        194        640:  30%|███       | 246/809 [04:54<11:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18      0.448     0.1338         96        640:  31%|███       | 248/809 [04:56<11:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4482     0.1336        208        640:  31%|███       | 250/809 [04:58<10:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4479     0.1336        366        640:  31%|███       | 251/809 [04:59<10:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4481     0.1333        240        640:  31%|███▏      | 253/809 [05:02<10:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4484     0.1334        219        640:  32%|███▏      | 255/809 [05:04<11:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.179     0.4488     0.1333        106        640:  32%|███▏      | 256/809 [05:05<10:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.177     0.4492     0.1332         90        640:  32%|███▏      | 258/809 [05:08<10:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.178     0.4489     0.1338        202        640:  32%|███▏      | 260/809 [05:10<11:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4485     0.1347        411        640:  32%|███▏      | 261/809 [05:12<11:24,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4476      0.135        271        640:  33%|███▎      | 263/809 [05:14<11:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.18     0.4483     0.1347        100        640:  33%|███▎      | 264/809 [05:15<10:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4478      0.135        295        640:  33%|███▎      | 266/809 [05:18<11:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4479     0.1353        246        640:  33%|███▎      | 268/809 [05:20<11:10,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4477     0.1356        443        640:  33%|███▎      | 269/809 [05:21<11:03,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4475     0.1357        466        640:  33%|███▎      | 271/809 [05:24<10:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4467     0.1363        390        640:  34%|███▎      | 273/809 [05:26<10:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4467     0.1362        302        640:  34%|███▍      | 274/809 [05:27<10:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4463     0.1365        154        640:  34%|███▍      | 276/809 [05:30<10:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4472     0.1362        148        640:  34%|███▍      | 278/809 [05:32<10:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4472     0.1361        256        640:  34%|███▍      | 279/809 [05:33<10:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4473      0.136        159        640:  35%|███▍      | 281/809 [05:36<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4472     0.1359        244        640:  35%|███▍      | 283/809 [05:38<10:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4473     0.1363        341        640:  35%|███���      | 285/809 [05:40<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4471     0.1363        397        640:  35%|███▌      | 286/809 [05:42<10:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4473     0.1363        370        640:  36%|███▌      | 288/809 [05:44<10:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4477     0.1361        228        640:  36%|███▌      | 290/809 [05:46<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.181     0.4476     0.1361        330        640:  36%|███▌      | 291/809 [05:48<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4473     0.1361        139        640:  36%|███▌      | 293/809 [05:50<10:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4465     0.1367        266        640:  36%|███▋      | 295/809 [05:52<10:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4468     0.1366        108        640:  37%|███▋      | 296/809 [05:54<10:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4465     0.1365        153        640:  37%|███▋      | 298/809 [05:56<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4461     0.1364        226        640:  37%|███▋      | 300/809 [05:58<10:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182     0.4462     0.1363        110        640:  37%|███▋      | 301/809 [05:59<10:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.182      0.446     0.1362        242        640:  37%|███▋      | 303/809 [06:02<09:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4456     0.1362        272        640:  38%|███▊      | 305/809 [06:04<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4455      0.136        238        640:  38%|███▊      | 307/809 [06:07<09:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.183     0.4454      0.136        256        640:  38%|███▊      | 308/809 [06:08<09:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4452     0.1366        221        640:  38%|███▊      | 310/809 [06:10<10:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4447     0.1372        573        640:  38%|███▊      | 311/809 [06:12<10:15,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4449     0.1371        308        640:  39%|███▊      | 313/809 [06:14<10:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184      0.445      0.137        129        640:  39%|███▉      | 315/809 [06:16<09:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4452     0.1369        213        640:  39%|███▉      | 316/809 [06:18<09:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184     0.4453     0.1369        177        640:  39%|███▉      | 318/809 [06:20<09:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185      0.445     0.1371        148        640:  40%|███▉      | 320/809 [06:22<09:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.184      0.445      0.137        143        640:  40%|███▉      | 321/809 [06:24<09:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4445      0.137        357        640:  40%|███▉      | 323/809 [06:26<09:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.185     0.4444      0.137        194        640:  40%|████      | 325/809 [06:28<09:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.187     0.4439     0.1372        324        640:  40%|████      | 326/809 [06:30<09:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.189      0.443     0.1373        315        640:  41%|████      | 328/809 [06:32<09:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4426     0.1372        157        640:  41%|████      | 330/809 [06:34<09:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4427     0.1371        115        640:  41%|████      | 331/809 [06:35<09:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4426      0.137        143        640:  41%|████      | 333/809 [06:38<09:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.189     0.4429     0.1367        126        640:  41%|████▏     | 335/809 [06:40<09:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4423     0.1369        300        640:  42%|████▏     | 337/809 [06:43<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4421     0.1369        415        640:  42%|████▏     | 338/809 [06:44<09:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4424     0.1367        219        640:  42%|████▏     | 340/809 [06:46<09:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4426     0.1369        347        640:  42%|████▏     | 342/809 [06:49<09:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.189     0.4429     0.1368         76        640:  42%|████▏     | 343/809 [06:50<09:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4426     0.1378        389        640:  43%|████▎     | 345/809 [06:52<09:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192     0.4422     0.1379        325        640:  43%|████▎     | 347/809 [06:55<09:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4418     0.1379        274        640:  43%|████▎     | 348/809 [06:56<09:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193      0.442      0.138        288        640:  43%|████▎     | 350/809 [06:58<09:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4418     0.1379        183        640:  44%|████▎     | 352/809 [07:01<08:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4417     0.1379        222        640:  44%|████▎     | 353/809 [07:02<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.194     0.4414     0.1378        353        640:  44%|████▍     | 355/809 [07:04<08:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.194     0.4412     0.1379        136        640:  44%|████▍     | 357/809 [07:06<09:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.194     0.4412     0.1377        204        640:  44%|████▍     | 358/809 [07:08<08:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192     0.4416     0.1375        138        640:  44%|████▍     | 360/809 [07:10<08:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.194     0.4413     0.1383        560        640:  45%|████▍     | 361/809 [07:11<08:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4412     0.1383        304        640:  45%|████▍     | 363/809 [07:14<09:38,  1.30s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4417     0.1382        156        640:  45%|████▌     | 365/809 [07:16<09:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192     0.4421     0.1381        134        640:  45%|████▌     | 366/809 [07:18<08:59,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192     0.4421     0.1379        211        640:  45%|████▌     | 368/809 [07:20<08:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192     0.4421     0.1378        361        640:  46%|████▌     | 370/809 [07:22<08:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4427     0.1375        143        640:  46%|████▌     | 372/809 [07:25<08:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4428     0.1375        165        640:  46%|████▌     | 373/809 [07:26<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4425     0.1373        253        640:  46%|████▋     | 375/809 [07:28<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4424     0.1373        197        640:  47%|████▋     | 377/809 [07:31<08:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4419     0.1375        239        640:  47%|████▋     | 379/809 [07:33<08:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4418     0.1375        359        640:  47%|████▋     | 380/809 [07:34<08:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4419     0.1374        208        640:  47%|████▋     | 382/809 [07:37<08:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4419     0.1374        209        640:  47%|████▋     | 384/809 [07:39<08:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191      0.442     0.1382        299        640:  48%|████▊     | 386/809 [07:41<08:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4421     0.1382        153        640:  48%|████▊     | 387/809 [07:43<08:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4419     0.1381        269        640:  48%|████▊     | 389/809 [07:45<08:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19      0.442      0.138        210        640:  48%|████▊     | 391/809 [07:47<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4419     0.1379        237        640:  48%|████▊     | 392/809 [07:48<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4417     0.1382        321        640:  49%|████▊     | 394/809 [07:51<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4417     0.1382        245        640:  49%|████▉     | 396/809 [07:53<08:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4417     0.1382        329        640:  49%|████▉     | 397/809 [07:55<08:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4418     0.1381        292        640:  49%|████▉     | 399/809 [07:57<08:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4414     0.1384        136        640:  50%|████▉     | 401/809 [07:59<08:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G       1.19     0.4415     0.1383        156        640:  50%|████▉     | 402/809 [08:00<07:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.191     0.4413     0.1386        227        640:  50%|████▉     | 404/809 [08:03<08:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.192      0.441     0.1387        170        640:  50%|█████     | 406/809 [08:05<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4407     0.1388        186        640:  50%|█████     | 407/809 [08:06<07:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.193     0.4405     0.1391        191        640:  51%|█████     | 409/809 [08:09<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.194     0.4402      0.139        352        640:  51%|█████     | 411/809 [08:11<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4399     0.1392        372        640:  51%|█████     | 412/809 [08:12<08:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4397     0.1392        243        640:  51%|█████     | 414/809 [08:15<07:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4395     0.1392        345        640:  51%|█████▏    | 416/809 [08:17<07:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4396      0.139        223        640:  52%|█████▏    | 418/809 [08:19<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4393     0.1392        417        640:  52%|█████▏    | 419/809 [08:21<07:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4389     0.1393        186        640:  52%|█████▏    | 421/809 [08:23<07:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4386     0.1398        529        640:  52%|█████▏    | 422/809 [08:24<08:01,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4385     0.1401        505        640:  52%|█████▏    | 424/809 [08:27<07:59,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4382     0.1401        251        640:  53%|█████▎    | 426/809 [08:29<07:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4381     0.1401        323        640:  53%|█████▎    | 427/809 [08:30<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1401        233        640:  53%|█████▎    | 429/809 [08:33<07:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4378       0.14        243        640:  53%|█████▎    | 431/809 [08:35<07:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4376     0.1399        368        640:  53%|█████▎    | 432/809 [08:36<07:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4379     0.1398        350        640:  54%|█████▎    | 434/809 [08:39<07:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4375     0.1398        300        640:  54%|█████▍    | 436/809 [08:41<07:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4373     0.1398        216        640:  54%|█████▍    | 437/809 [08:42<07:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4366     0.1399        230        640:  54%|█████▍    | 439/809 [08:45<07:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4364     0.1399        134        640:  55%|█████▍    | 441/809 [08:47<07:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4364     0.1398        174        640:  55%|█████▍    | 443/809 [08:49<07:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4362     0.1398        328        640:  55%|█████▍    | 444/809 [08:51<07:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4361     0.1397        234        640:  55%|█████▌    | 446/809 [08:53<07:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4357     0.1402        368        640:  55%|█████▌    | 447/809 [08:54<07:32,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4358     0.1403        271        640:  56%|█████▌    | 449/809 [08:57<07:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.206     0.4355     0.1404        511        640:  56%|█████▌    | 451/809 [08:59<07:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4356     0.1402        168        640:  56%|█████▌    | 452/809 [09:00<07:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.206     0.4353     0.1405        278        640:  56%|█████▌    | 454/809 [09:03<07:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.207     0.4354     0.1405        236        640:  56%|█████▋    | 456/809 [09:05<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.208     0.4351      0.141        128        640:  57%|█████▋    | 458/809 [09:08<06:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.207     0.4354     0.1409        103        640:  57%|█████▋    | 459/809 [09:09<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.206     0.4355     0.1408        207        640:  57%|█████▋    | 461/809 [09:11<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4359     0.1406        226        640:  57%|█████▋    | 463/809 [09:13<06:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4359     0.1406        224        640:  57%|█████▋    | 464/809 [09:15<06:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.207     0.4355     0.1409        424        640:  58%|█████▊    | 466/809 [09:17<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4358     0.1407        226        640:  58%|█████▊    | 468/809 [09:19<06:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.206     0.4358      0.141        286        640:  58%|█████▊    | 469/809 [09:21<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4361     0.1409        120        640:  58%|█████▊    | 471/809 [09:23<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4359     0.1407        251        640:  58%|█████▊    | 473/809 [09:25<06:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4358     0.1407        306        640:  59%|█████▊    | 474/809 [09:27<06:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4359     0.1405        232        640:  59%|█████▉    | 476/809 [09:29<06:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.205     0.4357     0.1405        137        640:  59%|█████▉    | 478/809 [09:31<06:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.206     0.4357     0.1406        298        640:  59%|█████▉    | 479/809 [09:33<06:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.204      0.436     0.1404        173        640:  59%|█████▉    | 481/809 [09:35<06:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.204     0.4358     0.1403        217        640:  60%|█████▉    | 483/809 [09:37<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203      0.436     0.1401        186        640:  60%|█████▉    | 485/809 [09:40<06:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4362       0.14        179        640:  60%|██████    | 486/809 [09:41<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.204      0.436     0.1402        421        640:  60%|██████    | 488/809 [09:43<06:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.204     0.4359     0.1402        232        640:  61%|██████    | 490/809 [09:45<06:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.204      0.436     0.1401        226        640:  61%|██████    | 491/809 [09:47<06:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4367     0.1399        126        640:  61%|██████    | 493/809 [09:49<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4368     0.1399        273        640:  61%|██████    | 495/809 [09:51<06:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4369     0.1398        232        640:  61%|██████▏   | 496/809 [09:53<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4369     0.1398        320        640:  62%|██████▏   | 498/809 [09:55<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201      0.437     0.1397        373        640:  62%|██████▏   | 500/809 [09:57<06:21,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4372     0.1396        125        640:  62%|██████▏   | 501/809 [09:59<06:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4371     0.1398        199        640:  62%|██████▏   | 503/809 [10:01<06:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4369     0.1399        426        640:  62%|██████▏   | 505/809 [10:03<06:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4369     0.1399        157        640:  63%|██████▎   | 506/809 [10:05<06:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437     0.1397        168        640:  63%|██████▎   | 508/809 [10:07<06:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4371     0.1396        300        640:  63%|██████▎   | 510/809 [10:09<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4372     0.1395        173        640:  63%|██████▎   | 512/809 [10:12<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4371     0.1394        152        640:  63%|██████▎   | 513/809 [10:13<05:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4371     0.1392        158        640:  64%|██████▎   | 515/809 [10:15<05:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1391        296        640:  64%|██████▍   | 517/809 [10:18<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4378      0.139         60        640:  64%|██████▍   | 518/809 [10:19<05:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199      0.438      0.139        408        640:  64%|██████▍   | 520/809 [10:21<05:54,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4384     0.1389        227        640:  65%|██████▍   | 522/809 [10:24<05:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4382     0.1393        499        640:  65%|██████▍   | 523/809 [10:25<05:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4381     0.1393        471        640:  65%|██████▍   | 525/809 [10:27<05:49,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4383     0.1392        229        640:  65%|██████▌   | 527/809 [10:30<05:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4382     0.1391        316        640:  65%|██████▌   | 528/809 [10:31<05:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4384      0.139        359        640:  66%|██████▌   | 530/809 [10:33<05:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4387     0.1388        414        640:  66%|██████▌   | 532/809 [10:36<05:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4385     0.1395        405        640:  66%|██████▌   | 533/809 [10:37<05:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1398        347        640:  66%|██████▌   | 535/809 [10:40<05:43,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4381     0.1401        264        640:  66%|██████▋   | 537/809 [10:42<05:37,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4382     0.1401        387        640:  67%|██████▋   | 538/809 [10:43<05:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4381     0.1403        353        640:  67%|██████▋   | 540/809 [10:46<05:32,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201      0.438     0.1403        439        640:  67%|██████▋   | 542/809 [10:48<05:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4378     0.1407        411        640:  67%|██████▋   | 543/809 [10:49<05:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4379     0.1405        295        640:  67%|██████▋   | 545/809 [10:52<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4379     0.1406        375        640:  68%|██████▊   | 547/809 [10:54<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4377     0.1411        462        640:  68%|██████▊   | 548/809 [10:55<05:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.203     0.4376     0.1411        266        640:  68%|██████▊   | 550/809 [10:58<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4378      0.141        193        640:  68%|██████▊   | 552/809 [11:00<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201      0.438     0.1409        321        640:  68%|██████▊   | 553/809 [11:01<05:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1408        305        640:  69%|██████▊   | 555/809 [11:04<05:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1408        284        640:  69%|██████▉   | 557/809 [11:06<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4387     0.1406        228        640:  69%|██████▉   | 559/809 [11:08<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4385      0.141        328        640:  69%|██████▉   | 560/809 [11:10<04:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4389     0.1409        237        640:  69%|██████▉   | 562/809 [11:12<04:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199      0.439     0.1409         99        640:  70%|██████▉   | 564/809 [11:14<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4391     0.1409        206        640:  70%|██████▉   | 565/809 [11:16<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4392     0.1408        170        640:  70%|███████   | 567/809 [11:18<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4393     0.1407        425        640:  70%|███████   | 569/809 [11:20<04:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4393     0.1407        310        640:  70%|███████   | 570/809 [11:21<04:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4396     0.1406        114        640:  71%|███████   | 572/809 [11:24<04:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4394     0.1407        222        640:  71%|███████   | 574/809 [11:26<04:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4392     0.1407        282        640:  71%|███████   | 575/809 [11:27<04:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4389     0.1407        286        640:  71%|███████▏  | 577/809 [11:30<04:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4387     0.1407        165        640:  72%|███████▏  | 579/809 [11:32<04:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4386     0.1407        235        640:  72%|███████▏  | 580/809 [11:33<04:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1409        339        640:  72%|███████▏  | 582/809 [11:36<04:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1409        396        640:  72%|███████▏  | 584/809 [11:38<04:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1409         67        640:  72%|███████▏  | 585/809 [11:39<04:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1408        189        640:  73%|███████▎  | 587/809 [11:42<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1407        248        640:  73%|███████▎  | 589/809 [11:44<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4386     0.1407        253        640:  73%|███████▎  | 591/809 [11:47<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4386     0.1407        116        640:  73%|███████▎  | 592/809 [11:48<04:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4382     0.1407        437        640:  73%|███████▎  | 594/809 [11:50<04:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4382     0.1407        309        640:  74%|███████▎  | 596/809 [11:52<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4383     0.1406         69        640:  74%|███████▍  | 597/809 [11:54<04:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4382      0.141        384        640:  74%|███████▍  | 599/809 [11:56<04:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1412        265        640:  74%|███████▍  | 601/809 [11:58<04:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4381     0.1412        162        640:  74%|███████▍  | 602/809 [12:00<04:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1412        225        640:  75%|███████▍  | 604/809 [12:02<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1412        264        640:  75%|███████▍  | 606/809 [12:04<04:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1411        197        640:  75%|███████▌  | 607/809 [12:06<04:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4382      0.141        136        640:  75%|███████▌  | 609/809 [12:08<03:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4381     0.1409        223        640:  76%|███████▌  | 611/809 [12:10<03:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1409        215        640:  76%|███████▌  | 612/809 [12:12<03:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.438     0.1408        264        640:  76%|███████▌  | 614/809 [12:14<03:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4379     0.1407        268        640:  76%|███████▌  | 616/809 [12:16<03:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201      0.438     0.1407        490        640:  76%|███████▋  | 618/809 [12:19<03:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201      0.438     0.1407        222        640:  77%|███████▋  | 619/809 [12:20<03:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4381     0.1408        157        640:  77%|███████▋  | 621/809 [12:22<03:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4383     0.1407        225        640:  77%|███████▋  | 623/809 [12:24<03:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1407        169        640:  77%|███████▋  | 624/809 [12:26<03:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1406        279        640:  77%|███████▋  | 626/809 [12:28<03:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4384     0.1409        151        640:  78%|��██████▊  | 628/809 [12:30<03:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4385     0.1409        282        640:  78%|███████▊  | 629/809 [12:32<03:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4385     0.1408        332        640:  78%|███████▊  | 631/809 [12:34<03:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4386     0.1407        320        640:  78%|███████▊  | 633/809 [12:36<03:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4388     0.1407        134        640:  78%|███████▊  | 634/809 [12:37<03:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4389     0.1407        197        640:  79%|███████▊  | 636/809 [12:40<03:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1406        114        640:  79%|███████▉  | 638/809 [12:42<03:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1405        181        640:  79%|███████▉  | 640/809 [12:45<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4394     0.1405        354        640:  79%|███████▉  | 641/809 [12:46<03:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1405        327        640:  79%|███████▉  | 643/809 [12:48<03:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1404        363        640:  80%|███████▉  | 645/809 [12:51<03:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4391     0.1405        351        640:  80%|███████▉  | 646/809 [12:52<03:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4391     0.1404        209        640:  80%|████████  | 648/809 [12:54<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4392     0.1403        233        640:  80%|████████  | 650/809 [12:56<03:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1402        224        640:  80%|████████  | 651/809 [12:58<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4395     0.1401        159        640:  81%|████████  | 653/809 [13:00<03:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4393     0.1404        459        640:  81%|████████  | 655/809 [13:02<03:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1403        237        640:  81%|████████  | 657/809 [13:05<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1402        301        640:  81%|████████▏ | 658/809 [13:06<02:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1401        299        640:  82%|████████▏ | 660/809 [13:08<02:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1401        233        640:  82%|████████▏ | 662/809 [13:11<02:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4392     0.1403        415        640:  82%|████████▏ | 663/809 [13:12<02:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1402        147        640:  82%|████████▏ | 665/809 [13:14<02:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1401        207        640:  82%|████████▏ | 667/809 [13:17<02:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4395     0.1401        172        640:  83%|████████▎ | 668/809 [13:18<02:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.195     0.4395     0.1399        299        640:  83%|████████▎ | 670/809 [13:20<02:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4394     0.1399        148        640:  83%|████████▎ | 672/809 [13:23<02:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4393     0.1398        361        640:  83%|████████▎ | 673/809 [13:24<02:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.196     0.4391     0.1399        449        640:  83%|████████▎ | 675/809 [13:26<02:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4389       0.14        494        640:  84%|████████▎ | 677/809 [13:29<02:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4391       0.14        110        640:  84%|████████▍ | 678/809 [13:30<02:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4391       0.14        180        640:  84%|████████▍ | 680/809 [13:32<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.198     0.4389     0.1402        247        640:  84%|████████▍ | 682/809 [13:35<02:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197     0.4391     0.1401        174        640:  84%|████████▍ | 683/809 [13:36<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.197      0.439       0.14        314        640:  85%|████████▍ | 685/809 [13:38<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4386     0.1404        527        640:  85%|████████▍ | 687/809 [13:41<02:30,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4385     0.1404         96        640:  85%|████████▌ | 689/809 [13:43<02:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4384     0.1404        235        640:  85%|████████▌ | 690/809 [13:44<02:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4382     0.1405        360        640:  86%|████████▌ | 692/809 [13:47<02:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4382     0.1404        255        640:  86%|████████▌ | 694/809 [13:49<02:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4383     0.1403        260        640:  86%|████████▌ | 696/809 [13:51<02:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4384     0.1403        184        640:  86%|████████▌ | 697/809 [13:53<02:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4385     0.1401        151        640:  86%|████████▋ | 699/809 [13:55<02:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4381     0.1409        293        640:  87%|████████▋ | 701/809 [13:57<02:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2      0.438     0.1409        271        640:  87%|████████▋ | 702/809 [13:59<02:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4378     0.1408        221        640:  87%|████████▋ | 704/809 [14:01<02:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4378     0.1408        295        640:  87%|████████▋ | 705/809 [14:02<02:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4375      0.141        126        640:  87%|████████▋ | 707/809 [14:05<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4376     0.1409        246        640:  88%|████████▊ | 709/809 [14:07<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4376     0.1409        232        640:  88%|████████▊ | 711/809 [14:09<01:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1408        237        640:  88%|████████▊ | 712/809 [14:10<01:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4378     0.1408        187        640:  88%|████████▊ | 714/809 [14:13<01:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4378     0.1407        205        640:  89%|████████▊ | 716/809 [14:15<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1407        198        640:  89%|████████▊ | 717/809 [14:16<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1406        252        640:  89%|████████▉ | 719/809 [14:19<01:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1405        217        640:  89%|████████▉ | 721/809 [14:21<01:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4378     0.1408        551        640:  89%|████████▉ | 722/809 [14:23<01:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1407        329        640:  89%|████████▉ | 724/809 [14:25<01:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4377     0.1408        392        640:  90%|████████▉ | 726/809 [14:27<01:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1408        167        640:  90%|████████▉ | 727/809 [14:28<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1407        404        640:  90%|█████████ | 729/809 [14:31<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199     0.4381     0.1406        145        640:  90%|█████████ | 731/809 [14:33<01:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2      0.438     0.1408        347        640:  90%|█████████ | 732/809 [14:34<01:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4379     0.1408        222        640:  91%|█████████ | 734/809 [14:37<01:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2      0.438     0.1408        492        640:  91%|█████████ | 736/809 [14:39<01:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199      0.438     0.1407        213        640:  91%|█████████ | 737/809 [14:40<01:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.199      0.438     0.1412        218        640:  91%|█████████▏| 739/809 [14:43<01:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2      0.438     0.1413        362        640:  92%|█████████▏| 741/809 [14:45<01:23,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2      0.438     0.1413        545        640:  92%|█████████▏| 742/809 [14:47<01:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4378     0.1414        415        640:  92%|█████████▏| 744/809 [14:49<01:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4378     0.1413        340        640:  92%|█████████▏| 746/809 [14:51<01:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1415        518        640:  92%|█████████▏| 747/809 [14:53<01:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1416        525        640:  93%|█████████▎| 749/809 [14:55<01:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1417        273        640:  93%|█████████▎| 750/809 [14:56<01:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4376     0.1418        159        640:  93%|█████████▎| 752/809 [14:59<01:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4378     0.1417        313        640:  93%|█████████▎| 754/809 [15:01<01:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4379     0.1417        212        640:  93%|█████████▎| 756/809 [15:04<01:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4379     0.1417        239        640:  94%|█████████▎| 757/809 [15:05<01:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4377     0.1417        292        640:  94%|█████████▍| 759/809 [15:07<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4375     0.1417        297        640:  94%|█████████▍| 761/809 [15:09<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4374     0.1417        258        640:  94%|█████████▍| 762/809 [15:11<00:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4371     0.1416        368        640:  94%|█████████▍| 764/809 [15:13<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4372     0.1415        359        640:  95%|█████████▍| 766/809 [15:15<00:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1414        186        640:  95%|█████████▍| 767/809 [15:17<00:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4377     0.1413        200        640:  95%|█████████▌| 769/809 [15:19<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4377     0.1412        275        640:  95%|█████████▌| 771/809 [15:21<00:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4377     0.1411        215        640:  95%|█████████▌| 772/809 [15:23<00:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4376     0.1411        310        640:  96%|█████████▌| 774/809 [15:25<00:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4375     0.1411        344        640:  96%|█████████▌| 776/809 [15:27<00:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G        1.2     0.4375     0.1411        301        640:  96%|█████████▌| 777/809 [15:29<00:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1412        299        640:  96%|█████████▋| 779/809 [15:31<00:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4373     0.1412        256        640:  97%|█████████▋| 781/809 [15:33<00:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4373     0.1411        302        640:  97%|█████████▋| 782/809 [15:35<00:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1411        196        640:  97%|█████████▋| 784/809 [15:37<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1411        376        640:  97%|█████████▋| 786/809 [15:39<00:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4375     0.1411        261        640:  97%|█████████▋| 787/809 [15:40<00:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4374     0.1411        206        640:  98%|█████████▊| 789/809 [15:43<00:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.201     0.4373     0.1411        353        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4371     0.1411        244        640:  98%|█████████▊| 793/809 [15:48<00:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4371     0.1411        142        640:  98%|█████████▊| 794/809 [15:49<00:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4369     0.1411        312        640:  98%|█████████▊| 796/809 [15:51<00:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437     0.1411        338        640:  99%|█████████▊| 798/809 [15:54<00:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437     0.1411        294        640:  99%|█████████▉| 799/809 [15:55<00:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437      0.141        328        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4369     0.1411        235        640:  99%|█████████▉| 803/809 [16:00<00:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437     0.1411        254        640:  99%|█████████▉| 804/809 [16:01<00:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202      0.437     0.1411        189        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     38/300      13.8G      1.202     0.4368      0.141        240        640: 100%|█████████▉| 808/809 [16:06<00:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:01<00:16,  4.06it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:03<00:14,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:05<00:12,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:10,  3.85it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:10<00:07,  3.66it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:13<00:05,  3.56it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:03,  3.79it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:01,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:01,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.5G      1.442     0.3574     0.1323        139        640:   0%|          | 2/809 [00:02<15:38,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.316     0.3891     0.1188        157        640:   0%|          | 3/809 [00:03<16:03,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4407     0.1074        126        640:   1%|          | 5/809 [00:05<16:01,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.167     0.4425     0.1253        117        640:   1%|          | 7/809 [00:08<16:17,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.105     0.4573      0.122        106        640:   1%|          | 8/809 [00:09<15:56,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.08     0.4716     0.1209        227        640:   1%|          | 10/809 [00:12<16:05,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.057     0.4772     0.1171        179        640:   1%|▏         | 12/809 [00:14<15:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.072     0.4712     0.1204        371        640:   2%|▏         | 13/809 [00:15<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.122     0.4573      0.128        326        640:   2%|▏         | 15/809 [00:18<15:54,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.144     0.4523     0.1297        220        640:   2%|▏         | 17/809 [00:20<15:44,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.133     0.4545     0.1277        288        640:   2%|▏         | 18/809 [00:21<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.13     0.4552     0.1256        162        640:   2%|▏         | 20/809 [00:24<16:05,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.148     0.4479     0.1244        232        640:   3%|▎         | 22/809 [00:26<15:46,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.163     0.4425     0.1245        304        640:   3%|▎         | 23/809 [00:27<15:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.189     0.4325     0.1249        286        640:   3%|▎         | 25/809 [00:30<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4326     0.1264        159        640:   3%|▎         | 27/809 [00:32<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4335     0.1263        247        640:   4%|▎         | 29/809 [00:34<15:36,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4333     0.1274        338        640:   4%|▎         | 30/809 [00:36<15:44,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.216     0.4298     0.1268        359        640:   4%|▍         | 32/809 [00:38<15:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.221     0.4294     0.1279        440        640:   4%|▍         | 34/809 [00:40<15:34,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.226     0.4273     0.1285        431        640:   4%|▍         | 35/809 [00:42<15:52,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.238     0.4243     0.1292        276        640:   5%|▍         | 37/809 [00:44<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.252     0.4211     0.1361        218        640:   5%|▍         | 39/809 [00:46<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.25      0.422     0.1357        197        640:   5%|▍         | 40/809 [00:48<15:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.255     0.4213     0.1378        338        640:   5%|▌         | 42/809 [00:50<15:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.243     0.4235     0.1362        264        640:   5%|▌         | 44/809 [00:52<15:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.233     0.4273     0.1354        131        640:   6%|▌         | 45/809 [00:54<14:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.24     0.4247     0.1351        328        640:   6%|▌         | 47/809 [00:56<15:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.231     0.4284     0.1334        206        640:   6%|▌         | 49/809 [00:58<15:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.236     0.4285     0.1352        241        640:   6%|▌         | 50/809 [01:00<14:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.237     0.4289     0.1349        200        640:   6%|▋         | 52/809 [01:02<14:58,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.234     0.4288     0.1336        180        640:   7%|▋         | 54/809 [01:04<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4272      0.133         99        640:   7%|▋         | 56/809 [01:07<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.236     0.4262     0.1328        311        640:   7%|▋         | 57/809 [01:08<14:36,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.246     0.4235     0.1365        288        640:   7%|▋         | 59/809 [01:10<15:04,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4271     0.1356         74        640:   8%|▊         | 61/809 [01:12<14:36,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.236     0.4267      0.135        197        640:   8%|▊         | 62/809 [01:14<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.236     0.4268     0.1352        206        640:   8%|▊         | 64/809 [01:16<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.241     0.4249     0.1353        308        640:   8%|▊         | 66/809 [01:18<14:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.242     0.4248     0.1349        286        640:   8%|▊         | 67/809 [01:20<14:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.244     0.4235     0.1354        180        640:   9%|▊         | 69/809 [01:22<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.242     0.4237     0.1349        234        640:   9%|▉         | 71/809 [01:24<14:26,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.246     0.4217     0.1352        209        640:   9%|▉         | 73/809 [01:27<14:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.249     0.4202     0.1356        472        640:   9%|▉         | 74/809 [01:28<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.251       0.42     0.1367        339        640:   9%|▉         | 76/809 [01:30<14:48,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.242     0.4235     0.1357        167        640:  10%|▉         | 78/809 [01:33<14:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.241     0.4238     0.1358        346        640:  10%|▉         | 79/809 [01:34<14:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4268     0.1346        163        640:  10%|█         | 81/809 [01:36<14:16,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.231     0.4266     0.1355         79        640:  10%|█         | 83/809 [01:39<14:29,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.231     0.4266     0.1355        435        640:  10%|█         | 84/809 [01:40<14:34,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4277      0.135        278        640:  11%|█         | 86/809 [01:42<14:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.225     0.4284     0.1349        273        640:  11%|█         | 88/809 [01:45<14:28,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222     0.4295     0.1345        252        640:  11%|█         | 89/809 [01:46<14:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.225     0.4284     0.1351        304        640:  11%|█         | 91/809 [01:48<14:36,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.223     0.4299     0.1349        196        640:  11%|█▏        | 93/809 [01:51<14:19,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222       0.43     0.1353        247        640:  12%|█▏        | 94/809 [01:52<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.217     0.4317     0.1345        232        640:  12%|█▏        | 96/809 [01:54<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.217     0.4314     0.1345        163        640:  12%|█▏        | 98/809 [01:57<14:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.219     0.4303     0.1346        401        640:  12%|█▏        | 99/809 [01:58<14:27,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.213     0.4323     0.1338        107        640:  12%|█▏        | 101/809 [02:00<14:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4335      0.133        249        640:  13%|█▎        | 103/809 [02:03<13:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.212     0.4322     0.1331        471        640:  13%|█▎        | 104/809 [02:04<13:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.211     0.4325     0.1332        382        640:  13%|█▎        | 106/809 [02:06<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.211     0.4324      0.133        145        640:  13%|█▎        | 108/809 [02:09<14:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.215     0.4312     0.1343        372        640:  13%|█▎        | 109/809 [02:10<13:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.212     0.4326     0.1341        320        640:  14%|█▎        | 111/809 [02:12<13:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4339     0.1332        258        640:  14%|█▍        | 113/809 [02:15<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4336      0.133        248        640:  14%|█▍        | 114/809 [02:16<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4358     0.1324        229        640:  14%|█▍        | 116/809 [02:18<13:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4374     0.1322        125        640:  15%|█▍        | 118/809 [02:20<13:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4368     0.1318        119        640:  15%|█▍        | 120/809 [02:23<13:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4361     0.1324        420        640:  15%|█▍        | 121/809 [02:24<13:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4356     0.1326        218        640:  15%|█▌        | 123/809 [02:26<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4351     0.1349        274        640:  15%|█▌        | 125/809 [02:29<13:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203      0.435     0.1348        211        640:  16%|█▌        | 126/809 [02:30<13:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4346     0.1354        278        640:  16%|█▌        | 128/809 [02:32<13:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346     0.1354        246        640:  16%|█▌        | 130/809 [02:35<13:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4352     0.1355        159        640:  16%|█▌        | 131/809 [02:36<13:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4346     0.1365        342        640:  16%|█▋        | 133/809 [02:38<13:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4346     0.1362        120        640:  17%|█▋        | 135/809 [02:41<13:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207      0.434     0.1363        332        640:  17%|█▋        | 136/809 [02:42<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1359        210        640:  17%|█▋        | 138/809 [02:44<13:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4344     0.1385        378        640:  17%|█▋        | 140/809 [02:47<13:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4343     0.1384        278        640:  17%|█▋        | 141/809 [02:48<13:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4344     0.1383        202        640:  18%|█▊        | 143/809 [02:50<13:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4354     0.1378        199        640:  18%|█▊        | 145/809 [02:53<12:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4369     0.1373        198        640:  18%|█▊        | 147/809 [02:55<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4359     0.1375        349        640:  18%|█▊        | 148/809 [02:56<12:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4349     0.1389        312        640:  19%|█▊        | 150/809 [02:59<13:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207      0.434     0.1396        386        640:  19%|█▉        | 152/809 [03:01<13:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4338     0.1401        210        640:  19%|█▉        | 153/809 [03:02<13:46,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4342     0.1398        242        640:  19%|█▉        | 155/809 [03:05<13:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205      0.435     0.1397        125        640:  19%|█▉        | 156/809 [03:06<13:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4344     0.1396        227        640:  20%|█▉        | 158/809 [03:08<13:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4342     0.1412        326        640:  20%|█▉        | 160/809 [03:11<13:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4333     0.1424        463        640:  20%|█▉        | 161/809 [03:12<13:15,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.213     0.4322     0.1439        441        640:  20%|██        | 163/809 [03:15<13:38,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.214      0.432     0.1441        252        640:  20%|██        | 165/809 [03:17<13:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.214     0.4321      0.144        202        640:  21%|██        | 166/809 [03:18<12:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.214     0.4314     0.1436        196        640:  21%|██        | 168/809 [03:20<12:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.214     0.4311     0.1435        317        640:  21%|██        | 170/809 [03:23<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.214      0.431     0.1433        252        640:  21%|██        | 171/809 [03:24<12:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.215     0.4307     0.1431        508        640:  21%|██▏       | 173/809 [03:26<12:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.213     0.4306     0.1428        108        640:  22%|██▏       | 175/809 [03:29<12:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.213     0.4307     0.1425        452        640:  22%|██▏       | 177/809 [03:31<12:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.21     0.4309     0.1422         94        640:  22%|██▏       | 178/809 [03:32<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4313     0.1419        234        640:  22%|██▏       | 180/809 [03:35<12:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4315     0.1419        183        640:  22%|██▏       | 182/809 [03:37<12:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4317     0.1416        210        640:  23%|██▎       | 183/809 [03:38<12:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4319     0.1412        173        640:  23%|██▎       | 185/809 [03:41<12:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4322      0.141        252        640:  23%|██▎       | 187/809 [03:43<12:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4312     0.1413        268        640:  23%|██▎       | 189/809 [03:45<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.21     0.4305     0.1421        313        640:  23%|██▎       | 190/809 [03:47<12:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.215     0.4295     0.1435        243        640:  24%|██▎       | 192/809 [03:49<12:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.217     0.4289     0.1437        231        640:  24%|██▍       | 194/809 [03:51<12:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4283     0.1439        207        640:  24%|██▍       | 195/809 [03:53<12:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4284     0.1436        270        640:  24%|██▍       | 197/809 [03:55<12:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.221     0.4275     0.1438        414        640:  25%|██▍       | 199/809 [03:57<12:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.219     0.4279     0.1436        262        640:  25%|██▍       | 201/809 [04:00<12:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4284     0.1436        224        640:  25%|██▍       | 202/809 [04:01<12:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4283     0.1434        270        640:  25%|██▌       | 204/809 [04:03<12:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4281     0.1432        281        640:  25%|██▌       | 205/809 [04:05<12:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.22     0.4279     0.1435        323        640:  26%|██▌       | 207/809 [04:07<12:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.221     0.4273     0.1432        320        640:  26%|██▌       | 209/809 [04:10<12:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.223     0.4267      0.144        492        640:  26%|██▌       | 210/809 [04:11<12:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4257     0.1445        351        640:  26%|██▌       | 212/809 [04:13<12:29,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.226      0.426     0.1442        382        640:  26%|██▋       | 214/809 [04:16<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.228     0.4253     0.1443        338        640:  27%|██▋       | 215/809 [04:17<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.229      0.425     0.1448        261        640:  27%|██▋       | 217/809 [04:19<11:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.229     0.4254     0.1448        120        640:  27%|██▋       | 219/809 [04:22<11:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.229     0.4254     0.1448        328        640:  27%|██▋       | 220/809 [04:23<11:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4262     0.1443        212        640:  27%|██▋       | 222/809 [04:25<11:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.23     0.4253     0.1447        378        640:  28%|██▊       | 224/809 [04:28<11:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.23     0.4254     0.1448        279        640:  28%|██▊       | 225/809 [04:29<11:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4248     0.1447        250        640:  28%|██▊       | 227/809 [04:31<11:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.233     0.4246     0.1443        275        640:  28%|██▊       | 229/809 [04:34<11:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.233     0.4248     0.1445        276        640:  28%|██▊       | 230/809 [04:35<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.233      0.425     0.1443        289        640:  29%|██▊       | 232/809 [04:37<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4252      0.144        231        640:  29%|██▉       | 234/809 [04:40<11:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.231     0.4254     0.1439        182        640:  29%|██▉       | 235/809 [04:41<11:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4252     0.1439        398        640:  29%|██▉       | 237/809 [04:43<11:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4253     0.1438        282        640:  30%|██▉       | 239/809 [04:46<11:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.231     0.4258     0.1435        130        640:  30%|██▉       | 240/809 [04:47<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.233     0.4254     0.1439        239        640:  30%|██▉       | 242/809 [04:49<11:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.234     0.4252     0.1437        378        640:  30%|███       | 244/809 [04:52<11:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.234      0.425      0.144        438        640:  30%|███       | 245/809 [04:53<11:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.237     0.4241     0.1445        397        640:  31%|███       | 247/809 [04:55<11:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4244     0.1443        134        640:  31%|███       | 249/809 [04:58<11:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4243     0.1442        362        640:  31%|███       | 250/809 [04:59<11:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4242      0.144        172        640:  31%|███       | 252/809 [05:01<11:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.235     0.4248     0.1442         76        640:  31%|███▏      | 254/809 [05:04<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4255      0.144         98        640:  32%|███▏      | 255/809 [05:05<10:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.232     0.4253     0.1438        484        640:  32%|███▏      | 257/809 [05:07<10:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.229     0.4258     0.1433        255        640:  32%|███▏      | 259/809 [05:10<11:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.229     0.4261     0.1433        173        640:  32%|███▏      | 260/809 [05:11<10:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4265     0.1429        107        640:  32%|███▏      | 262/809 [05:13<10:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4263     0.1426        251        640:  33%|███▎      | 264/809 [05:15<10:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4264     0.1424        163        640:  33%|███▎      | 266/809 [05:18<10:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4265     0.1424        221        640:  33%|███▎      | 267/809 [05:19<10:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.228     0.4265     0.1429        259        640:  33%|███▎      | 269/809 [05:21<10:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.227     0.4267     0.1427        196        640:  33%|███▎      | 271/809 [05:24<10:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.226     0.4268     0.1426        221        640:  34%|███▎      | 272/809 [05:25<10:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.225     0.4274     0.1423        166        640:  34%|███▍      | 274/809 [05:27<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.226     0.4274     0.1428        350        640:  34%|███▍      | 276/809 [05:30<10:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.224     0.4279     0.1425        326        640:  34%|███▍      | 278/809 [05:32<10:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.225     0.4276     0.1428        466        640:  34%|███▍      | 279/809 [05:33<10:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.224     0.4278     0.1426        294        640:  35%|███▍      | 281/809 [05:36<10:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222     0.4286     0.1421        188        640:  35%|███▍      | 283/809 [05:38<10:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222     0.4286      0.142        290        640:  35%|███▌      | 284/809 [05:39<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222     0.4288     0.1421        346        640:  35%|███▌      | 286/809 [05:41<10:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.221     0.4285      0.142        319        640:  36%|███▌      | 288/809 [05:44<10:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.221     0.4286     0.1418        123        640:  36%|███▌      | 289/809 [05:45<10:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222      0.428     0.1419        290        640:  36%|███▌      | 291/809 [05:48<10:49,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.224     0.4274     0.1421        190        640:  36%|███▌      | 293/809 [05:50<10:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.223     0.4278     0.1419        199        640:  36%|███▋      | 294/809 [05:51<10:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222     0.4279     0.1417        159        640:  37%|███▋      | 296/809 [05:54<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.222      0.428     0.1416        290        640:  37%|███▋      | 298/809 [05:56<10:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.22     0.4284     0.1414         96        640:  37%|███▋      | 299/809 [05:57<10:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.22     0.4286     0.1412        274        640:  37%|███▋      | 301/809 [06:00<10:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4291     0.1411        247        640:  37%|███▋      | 303/809 [06:02<10:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.218     0.4294     0.1409        224        640:  38%|███▊      | 304/809 [06:03<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.217     0.4297     0.1407        188        640:  38%|███▊      | 306/809 [06:05<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.216     0.4299     0.1405        192        640:  38%|███▊      | 308/809 [06:08<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.215     0.4302     0.1403        186        640:  38%|███▊      | 309/809 [06:09<09:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.213     0.4307     0.1401        212        640:  38%|███▊      | 311/809 [06:11<09:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.211     0.4316     0.1397        121        640:  39%|███▊      | 313/809 [06:14<09:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.21     0.4318     0.1396        288        640:  39%|███▉      | 314/809 [06:15<09:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4322     0.1394         86        640:  39%|███▉      | 316/809 [06:17<09:59,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.21      0.432     0.1398        324        640:  39%|███▉      | 318/809 [06:20<09:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4327     0.1396        144        640:  39%|███▉      | 319/809 [06:21<10:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G       1.21     0.4323     0.1397        287        640:  40%|███▉      | 321/809 [06:23<09:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4326     0.1395        325        640:  40%|███▉      | 323/809 [06:26<09:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4331     0.1394        142        640:  40%|████      | 324/809 [06:27<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209      0.433     0.1397        293        640:  40%|████      | 326/809 [06:29<09:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209      0.433     0.1398        233        640:  41%|████      | 328/809 [06:32<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4334     0.1396        214        640:  41%|████      | 330/809 [06:34<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4333     0.1399        433        640:  41%|████      | 331/809 [06:35<09:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4335     0.1397        154        640:  41%|████      | 333/809 [06:38<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4339     0.1395        210        640:  41%|████▏     | 335/809 [06:40<09:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4341     0.1393        168        640:  42%|████▏     | 336/809 [06:41<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4336     0.1398        549        640:  42%|████▏     | 338/809 [06:44<09:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4338     0.1397        146        640:  42%|████▏     | 340/809 [06:46<09:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4341     0.1396        281        640:  42%|████▏     | 341/809 [06:47<09:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4343     0.1396        273        640:  42%|████▏     | 343/809 [06:50<09:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4348     0.1394        212        640:  43%|████▎     | 345/809 [06:52<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4345     0.1396        517        640:  43%|████▎     | 346/809 [06:53<09:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4344     0.1396        381        640:  43%|████▎     | 348/809 [06:56<09:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4342     0.1395        302        640:  43%|████▎     | 350/809 [06:58<09:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4344     0.1393        234        640:  44%|████▎     | 352/809 [07:00<08:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4338     0.1392        252        640:  44%|████▍     | 354/809 [07:03<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4339     0.1392        244        640:  44%|████▍     | 355/809 [07:04<08:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4343      0.139         73        640:  44%|████▍     | 357/809 [07:06<08:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4343     0.1388        304        640:  44%|████▍     | 359/809 [07:09<08:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4344     0.1386        256        640:  44%|████▍     | 360/809 [07:10<08:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4348     0.1385        416        640:  45%|████▍     | 362/809 [07:12<08:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4345     0.1389        508        640:  45%|████▍     | 364/809 [07:15<09:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4349     0.1389        129        640:  45%|████▌     | 365/809 [07:16<08:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4347     0.1388        330        640:  45%|████▌     | 367/809 [07:18<08:59,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4351     0.1386        126        640:  46%|████▌     | 369/809 [07:21<08:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4354     0.1384        164        640:  46%|████▌     | 370/809 [07:22<08:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4359     0.1381        233        640:  46%|████▌     | 372/809 [07:24<08:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4359      0.138        369        640:  46%|████▌     | 374/809 [07:27<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4361      0.138        162        640:  46%|████▋     | 375/809 [07:28<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4366     0.1381        383        640:  47%|████▋     | 377/809 [07:30<08:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4364     0.1384        305        640:  47%|████▋     | 379/809 [07:33<08:48,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4366     0.1383        309        640:  47%|████▋     | 380/809 [07:34<08:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4369     0.1384        239        640:  47%|████▋     | 382/809 [07:36<08:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4373     0.1382        196        640:  47%|████▋     | 384/809 [07:38<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4374     0.1382        239        640:  48%|████▊     | 386/809 [07:41<08:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4376      0.138        177        640:  48%|████▊     | 387/809 [07:42<08:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4372      0.138        216        640:  48%|████▊     | 389/809 [07:44<08:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4367     0.1383        310        640:  48%|████▊     | 391/809 [07:47<08:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4367     0.1382        146        640:  48%|████▊     | 392/809 [07:48<08:42,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4364     0.1381        160        640:  49%|████▊     | 394/809 [07:50<08:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202      0.436     0.1382        267        640:  49%|████▉     | 396/809 [07:53<08:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201      0.436     0.1381        136        640:  49%|████▉     | 397/809 [07:54<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4357      0.138        290        640:  49%|████▉     | 399/809 [07:56<08:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4353     0.1383        452        640:  50%|████▉     | 401/809 [07:59<08:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4352     0.1383        398        640:  50%|████▉     | 402/809 [08:00<08:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203      0.435     0.1382        238        640:  50%|████▉     | 404/809 [08:02<08:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4352     0.1379        250        640:  50%|█████     | 406/809 [08:05<07:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4353     0.1379        243        640:  50%|█████     | 407/809 [08:06<07:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4356     0.1377        173        640:  51%|█████     | 409/809 [08:08<07:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4356     0.1377        321        640:  51%|█████     | 411/809 [08:11<07:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4354     0.1377        306        640:  51%|█████     | 413/809 [08:13<07:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4351     0.1376        176        640:  51%|█████     | 414/809 [08:14<07:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4343     0.1379        358        640:  51%|█████▏    | 416/809 [08:17<07:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4341     0.1377        144        640:  52%|█████▏    | 418/809 [08:19<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204      0.434     0.1377        160        640:  52%|█████▏    | 419/809 [08:20<07:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4339     0.1375        381        640:  52%|█████▏    | 421/809 [08:22<07:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4342     0.1374        152        640:  52%|█████▏    | 423/809 [08:25<07:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4342     0.1373        168        640:  52%|█████▏    | 424/809 [08:26<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4343     0.1372        181        640:  53%|█████▎    | 426/809 [08:28<07:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201      0.435      0.137        164        640:  53%|█████▎    | 428/809 [08:31<07:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4349     0.1372        494        640:  53%|█████▎    | 430/809 [08:33<07:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2      0.435      0.137        121        640:  53%|█████▎    | 431/809 [08:34<07:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2      0.435     0.1369        225        640:  54%|█████▎    | 433/809 [08:36<07:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4353     0.1368         87        640:  54%|█████▍    | 435/809 [08:39<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4355     0.1367        172        640:  54%|█████▍    | 436/809 [08:40<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4357     0.1367        297        640:  54%|█████▍    | 438/809 [08:42<07:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4356     0.1366        229        640:  54%|█████▍    | 440/809 [08:45<07:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4357     0.1366        203        640:  55%|█████▍    | 441/809 [08:46<07:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4355     0.1367        234        640:  55%|█████▍    | 443/809 [08:48<07:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4355     0.1366        208        640:  55%|█████▌    | 445/809 [08:51<07:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4357     0.1364        195        640:  55%|█████▌    | 447/809 [08:53<07:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4356     0.1365        311        640:  55%|█████▌    | 448/809 [08:54<07:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4358     0.1364        179        640:  56%|█████▌    | 450/809 [08:56<06:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4361     0.1363        206        640:  56%|█████▌    | 452/809 [08:59<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4363     0.1362        153        640:  56%|█████▌    | 453/809 [09:00<06:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4365     0.1361        295        640:  56%|█████▌    | 455/809 [09:02<06:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4367      0.136        201        640:  56%|█████▋    | 457/809 [09:05<07:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4367     0.1361        268        640:  57%|█████▋    | 458/809 [09:06<06:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4371      0.136         99        640:  57%|█████▋    | 460/809 [09:08<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4371     0.1365        461        640:  57%|█████▋    | 462/809 [09:11<07:05,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4374     0.1364        150        640:  57%|█████▋    | 463/809 [09:12<06:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4376     0.1364        407        640:  57%|█████▋    | 465/809 [09:14<06:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1363        345        640:  58%|█████▊    | 467/809 [09:17<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194      0.438     0.1363        206        640:  58%|█████▊    | 468/809 [09:18<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1363        260        640:  58%|█████▊    | 470/809 [09:20<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4381     0.1363        180        640:  58%|█████▊    | 472/809 [09:23<06:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1362        313        640:  58%|█████▊    | 473/809 [09:24<06:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4377     0.1364        235        640:  59%|█████▊    | 475/809 [09:27<06:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4378     0.1363        173        640:  59%|█████▉    | 477/809 [09:29<06:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4377     0.1363        363        640:  59%|█████▉    | 478/809 [09:30<06:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4376     0.1363        212        640:  59%|█████▉    | 480/809 [09:32<06:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4375     0.1364        114        640:  60%|█████▉    | 482/809 [09:35<06:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4375     0.1363        257        640:  60%|█████▉    | 483/809 [09:36<06:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4373     0.1364         90        640:  60%|█████▉    | 485/809 [09:38<06:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4374     0.1362        146        640:  60%|██████    | 487/809 [09:41<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4373     0.1364        131        640:  60%|██████    | 489/809 [09:43<06:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197      0.437     0.1365        535        640:  61%|██████    | 490/809 [09:44<06:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4367     0.1366        435        640:  61%|██████    | 492/809 [09:47<06:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4365     0.1367        340        640:  61%|██████    | 493/809 [09:48<06:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4362     0.1367        262        640:  61%|██████    | 495/809 [09:50<06:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4358     0.1366        208        640:  61%|██████▏   | 497/809 [09:53<06:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4356     0.1369         87        640:  62%|██████▏   | 499/809 [09:55<06:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4352     0.1374        555        640:  62%|██████▏   | 500/809 [09:57<06:27,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4352     0.1374        162        640:  62%|██████▏   | 502/809 [09:59<06:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4352     0.1374        355        640:  62%|██████▏   | 504/809 [10:01<06:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4353     0.1374        331        640:  63%|██████▎   | 506/809 [10:04<06:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4354     0.1373        221        640:  63%|██████▎   | 507/809 [10:05<06:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203      0.435      0.138        287        640:  63%|██████▎   | 509/809 [10:07<06:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347      0.138        210        640:  63%|██████▎   | 511/809 [10:10<05:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346     0.1382        404        640:  63%|██████▎   | 512/809 [10:11<05:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4343     0.1386        208        640:  64%|██████▎   | 514/809 [10:13<05:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4344     0.1385         98        640:  64%|██████▎   | 515/809 [10:15<05:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4341     0.1388        267        640:  64%|██████▍   | 517/809 [10:17<05:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4339     0.1392        332        640:  64%|██████▍   | 519/809 [10:20<05:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4338     0.1392        339        640:  64%|██████▍   | 520/809 [10:21<05:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4338      0.139        438        640:  65%|██████▍   | 522/809 [10:23<05:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4337     0.1392        523        640:  65%|██████▍   | 524/809 [10:26<05:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4335     0.1394        464        640:  65%|██████▍   | 525/809 [10:27<05:50,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4333     0.1393        355        640:  65%|██████▌   | 527/809 [10:29<05:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4332     0.1393        466        640:  65%|██████▌   | 529/809 [10:32<05:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209      0.433     0.1395        437        640:  66%|██████▌   | 530/809 [10:33<05:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4332     0.1393         80        640:  66%|██████▌   | 532/809 [10:35<05:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4334     0.1393        166        640:  66%|██████▌   | 534/809 [10:38<05:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4332     0.1392        210        640:  66%|██████▌   | 535/809 [10:39<05:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4332     0.1392        166        640:  66%|██████▋   | 537/809 [10:41<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208      0.433     0.1392        277        640:  67%|██████▋   | 539/809 [10:44<05:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208      0.433     0.1392        262        640:  67%|██████▋   | 540/809 [10:45<05:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.209     0.4328     0.1392        357        640:  67%|██████▋   | 542/809 [10:47<05:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.208     0.4331     0.1391        173        640:  67%|██████▋   | 544/809 [10:50<05:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4334      0.139         87        640:  67%|██████▋   | 545/809 [10:51<05:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.207     0.4333     0.1389        506        640:  68%|██████▊   | 547/809 [10:53<05:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4335     0.1387        231        640:  68%|██████▊   | 549/809 [10:56<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.206     0.4334     0.1386        327        640:  68%|██████▊   | 551/809 [10:58<05:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4337     0.1385        104        640:  68%|██████▊   | 552/809 [10:59<05:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4337     0.1384        111        640:  68%|██████▊   | 554/809 [11:01<05:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4338     0.1384        164        640:  69%|██████▊   | 556/809 [11:04<04:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4339     0.1384        206        640:  69%|██████▉   | 557/809 [11:05<04:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4345     0.1383         73        640:  69%|██████▉   | 559/809 [11:07<04:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4342     0.1383        358        640:  69%|██████▉   | 561/809 [11:10<04:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4345     0.1381        145        640:  69%|██████▉   | 562/809 [11:11<04:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4348      0.138        206        640:  70%|██████▉   | 564/809 [11:13<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1379        168        640:  70%|██████▉   | 566/809 [11:16<04:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347      0.138        480        640:  70%|███████   | 567/809 [11:17<04:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346     0.1381        253        640:  70%|███████   | 569/809 [11:19<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346      0.138        115        640:  71%|███████   | 571/809 [11:22<04:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346      0.138        285        640:  71%|███████   | 572/809 [11:23<04:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4345     0.1379        315        640:  71%|███████   | 574/809 [11:25<04:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4344     0.1381        146        640:  71%|███████   | 576/809 [11:28<04:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4343     0.1383        206        640:  71%|███████▏  | 578/809 [11:30<04:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.205     0.4344     0.1383        215        640:  72%|███████▏  | 579/809 [11:31<04:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1381        246        640:  72%|███████▏  | 581/809 [11:34<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1381        323        640:  72%|███████▏  | 582/809 [11:35<04:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4351     0.1382        153        640:  72%|███████▏  | 584/809 [11:37<04:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1381        244        640:  72%|███████▏  | 586/809 [11:40<04:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4348     0.1381        312        640:  73%|███████▎  | 587/809 [11:41<04:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347     0.1394        460        640:  73%|███████▎  | 589/809 [11:43<04:34,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4349     0.1393        122        640:  73%|███████▎  | 591/809 [11:46<04:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4348     0.1393        299        640:  73%|███████▎  | 592/809 [11:47<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4347     0.1393        344        640:  73%|███████▎  | 594/809 [11:49<04:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4348     0.1393        172        640:  74%|███████▎  | 596/809 [11:52<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346     0.1396        302        640:  74%|███████▍  | 597/809 [11:53<04:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4346     0.1396        235        640:  74%|███████▍  | 599/809 [11:55<04:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4345     0.1397         87        640:  74%|███████▍  | 601/809 [11:58<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347     0.1396         94        640:  75%|███████▍  | 603/809 [12:00<04:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347     0.1396        306        640:  75%|███████▍  | 604/809 [12:01<04:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204     0.4347     0.1396        190        640:  75%|███████▍  | 606/809 [12:04<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.204      0.435     0.1395        198        640:  75%|███████▌  | 608/809 [12:06<03:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4351     0.1394        167        640:  75%|███████▌  | 609/809 [12:07<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4352     0.1395        340        640:  76%|███████▌  | 611/809 [12:09<03:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.203     0.4355     0.1394        212        640:  76%|███████▌  | 613/809 [12:12<03:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4356     0.1393        222        640:  76%|███████▌  | 614/809 [12:13<03:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.202     0.4355     0.1393        279        640:  76%|███████▌  | 616/809 [12:15<03:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201     0.4359     0.1391         98        640:  76%|███████▋  | 618/809 [12:18<03:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.201      0.436      0.139        196        640:  77%|███████▋  | 619/809 [12:19<03:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4362      0.139        222        640:  77%|███████▋  | 621/809 [12:21<03:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G        1.2     0.4364     0.1389        283        640:  77%|███████▋  | 623/809 [12:24<03:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4365     0.1388        180        640:  77%|███████▋  | 625/809 [12:26<03:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4368     0.1387         96        640:  77%|███████▋  | 626/809 [12:27<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4368     0.1389        161        640:  78%|███████▊  | 628/809 [12:30<03:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.199     0.4367     0.1394        525        640:  78%|███████▊  | 630/809 [12:32<03:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198      0.437     0.1393         93        640:  78%|███████▊  | 631/809 [12:33<03:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4373     0.1392        124        640:  78%|███████▊  | 633/809 [12:36<03:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4372     0.1391        278        640:  78%|███████▊  | 635/809 [12:38<03:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4374      0.139        161        640:  79%|███████▊  | 636/809 [12:39<03:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4374     0.1389        372        640:  79%|███████▉  | 638/809 [12:42<03:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4372     0.1392        200        640:  79%|███████▉  | 640/809 [12:44<03:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4373     0.1392        117        640:  79%|███████▉  | 641/809 [12:45<03:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4377      0.139        141        640:  79%|███████▉  | 643/809 [12:48<03:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4378      0.139        134        640:  80%|███████▉  | 645/809 [12:50<03:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4381     0.1388        165        640:  80%|███████▉  | 647/809 [12:52<03:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4383     0.1387        198        640:  80%|████████  | 649/809 [12:54<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4382     0.1388        441        640:  80%|████████  | 650/809 [12:56<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4381      0.139        371        640:  81%|████████  | 652/809 [12:58<03:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1391        361        640:  81%|████████  | 654/809 [13:01<03:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4381     0.1391        240        640:  81%|████████  | 655/809 [13:02<03:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4383     0.1389        181        640:  81%|████████  | 657/809 [13:04<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4386     0.1388        221        640:  81%|████████▏ | 659/809 [13:07<03:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4387     0.1387        139        640:  82%|████████▏ | 660/809 [13:08<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4388     0.1386        194        640:  82%|████████▏ | 662/809 [13:10<02:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4387     0.1386        370        640:  82%|████████▏ | 664/809 [13:13<02:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4386     0.1386        369        640:  82%|████████▏ | 665/809 [13:14<02:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4385     0.1386        239        640:  82%|████████▏ | 667/809 [13:16<02:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4387     0.1385        131        640:  83%|████████▎ | 669/809 [13:18<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4388     0.1384        116        640:  83%|████████▎ | 671/809 [13:21<02:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4387     0.1385        559        640:  83%|████████▎ | 672/809 [13:22<02:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1383         93        640:  83%|████████▎ | 674/809 [13:24<02:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1383        358        640:  83%|████████�� | 675/809 [13:26<02:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392     0.1383        496        640:  84%|████████▎ | 677/809 [13:28<02:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392     0.1383        433        640:  84%|████████▍ | 679/809 [13:31<02:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1385        452        640:  84%|████████▍ | 680/809 [13:32<02:39,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4391     0.1386        250        640:  84%|████████▍ | 682/809 [13:34<02:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4392     0.1387        453        640:  85%|████████▍ | 684/809 [13:37<02:35,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4391     0.1392        527        640:  85%|████████▍ | 685/809 [13:38<02:38,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.439     0.1392        363        640:  85%|████████▍ | 687/809 [13:41<02:31,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4389     0.1392        437        640:  85%|████████▌ | 688/809 [13:42<02:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194      0.439     0.1391        196        640:  85%|████████▌ | 690/809 [13:44<02:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1391        457        640:  86%|████████▌ | 692/809 [13:47<02:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1391        345        640:  86%|████████▌ | 693/809 [13:48<02:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4391      0.139        192        640:  86%|████████▌ | 695/809 [13:50<02:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1389        123        640:  86%|████████▌ | 697/809 [13:52<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1389        237        640:  86%|████████▋ | 699/809 [13:55<02:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1388        128        640:  87%|████████▋ | 700/809 [13:56<02:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4388     0.1388        240        640:  87%|████████▋ | 702/809 [13:58<02:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193      0.439     0.1387        135        640:  87%|████████▋ | 704/809 [14:01<02:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4391     0.1386         87        640:  87%|████████▋ | 705/809 [14:02<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392     0.1385        269        640:  87%|████████▋ | 707/809 [14:04<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4393     0.1384        261        640:  88%|████████▊ | 709/809 [14:07<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4394     0.1383        191        640:  88%|████████▊ | 710/809 [14:08<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4394     0.1383        181        640:  88%|████████▊ | 712/809 [14:10<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4395     0.1382        275        640:  88%|████████▊ | 714/809 [14:13<01:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4396     0.1382        230        640:  88%|████████▊ | 715/809 [14:14<01:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4396     0.1381        198        640:  89%|████████▊ | 717/809 [14:16<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.192     0.4395      0.138        258        640:  89%|████████▉ | 719/809 [14:18<01:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392     0.1381        284        640:  89%|████████▉ | 721/809 [14:21<01:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392      0.138        193        640:  89%|████████▉ | 722/809 [14:22<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4392     0.1381        231        640:  89%|████████▉ | 724/809 [14:24<01:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1384        463        640:  90%|████████▉ | 726/809 [14:27<01:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1384        312        640:  90%|████████▉ | 727/809 [14:28<01:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194      0.439     0.1384        218        640:  90%|█████████ | 729/809 [14:30<01:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1385        186        640:  90%|█████████ | 731/809 [14:33<01:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4389     0.1384        148        640:  90%|█████████ | 732/809 [14:34<01:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4388     0.1384        163        640:  91%|█████████ | 734/809 [14:36<01:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4386     0.1389        571        640:  91%|█████████ | 736/809 [14:39<01:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4385     0.1388        199        640:  91%|█████████ | 737/809 [14:40<01:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4384     0.1388        299        640:  91%|█████████▏| 739/809 [14:42<01:25,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4383     0.1392        453        640:  92%|█████████▏| 741/809 [14:45<01:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4384     0.1391        137        640:  92%|█████████▏| 742/809 [14:46<01:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4384     0.1391        321        640:  92%|█████████▏| 744/809 [14:49<01:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4385      0.139        130        640:  92%|█████████▏| 746/809 [14:51<01:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4384     0.1391        394        640:  92%|█████████▏| 747/809 [14:52<01:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4381     0.1392        384        640:  93%|█████████▎| 749/809 [14:55<01:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196      0.438     0.1392        173        640:  93%|█████████▎| 750/809 [14:56<01:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4377     0.1394        201        640:  93%|█████████▎| 752/809 [14:58<01:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4376     0.1394        228        640:  93%|█████████▎| 754/809 [15:01<01:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4374     0.1396        524        640:  93%|█████████▎| 755/809 [15:02<01:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.198     0.4375     0.1395        279        640:  94%|█████████▎| 757/809 [15:04<01:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4375     0.1393        159        640:  94%|█████████▍| 759/809 [15:07<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4376     0.1393        451        640:  94%|█████████▍| 761/809 [15:09<00:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4378     0.1393        124        640:  94%|█████████▍| 762/809 [15:10<00:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4378     0.1392        306        640:  94%|█████████▍| 764/809 [15:13<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4375     0.1394        349        640:  95%|█████████▍| 766/809 [15:15<00:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4376     0.1394         38        640:  95%|█████████▍| 767/809 [15:16<00:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4375     0.1394        327        640:  95%|█████████▌| 769/809 [15:18<00:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4375     0.1394        373        640:  95%|█████████▌| 771/809 [15:21<00:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4376     0.1393        267        640:  95%|█████████▌| 772/809 [15:22<00:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4378     0.1392        152        640:  96%|█████████▌| 774/809 [15:24<00:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1391        154        640:  96%|█████████▌| 776/809 [15:27<00:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4381      0.139        264        640:  96%|█████████▌| 778/809 [15:29<00:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4382      0.139        258        640:  96%|█████████▋| 779/809 [15:30<00:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438     0.1391        462        640:  97%|█████████▋| 781/809 [15:33<00:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4381     0.1389        252        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4384     0.1388        135        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.193     0.4385     0.1388        198        640:  97%|█████████▋| 786/809 [15:39<00:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4382      0.139        265        640:  97%|█████████▋| 788/809 [15:41<00:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195      0.438      0.139        101        640:  98%|█████████▊| 790/809 [15:44<00:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4379      0.139        259        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4376     0.1392        239        640:  98%|█████████▊| 793/809 [15:47<00:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4373     0.1393        281        640:  98%|█████████▊| 794/809 [15:49<00:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.197     0.4372     0.1393        189        640:  98%|█████████▊| 796/809 [15:51<00:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.196     0.4373     0.1392         73        640:  99%|█████████▊| 798/809 [15:53<00:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4375     0.1391        181        640:  99%|█████████▉| 800/809 [15:56<00:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4377      0.139         80        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.195     0.4376     0.1394        278        640:  99%|█████████▉| 803/809 [15:59<00:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4379     0.1393        192        640: 100%|█████████▉| 805/809 [16:02<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194      0.438     0.1393        238        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     39/300      13.6G      1.194     0.4379     0.1393        381        640: 100%|█████████▉| 808/809 [16:05<00:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.04it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  41%|████      | 28/69 [00:07<00:10,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:09<00:08,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:11<00:06,  3.68it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:04,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.83it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.83it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.022     0.4683     0.1062        142        640:   0%|          | 2/809 [00:02<16:08,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G     0.9738     0.4774    0.09552        244        640:   0%|          | 4/809 [00:04<15:45,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      0.939     0.4933    0.09579        220        640:   1%|          | 5/809 [00:05<15:37,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G     0.8802     0.5048    0.09315         61        640:   1%|          | 7/809 [00:08<15:28,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G     0.9414     0.4931     0.1024        314        640:   1%|          | 9/809 [00:10<15:44,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G     0.9642     0.4866      0.103        242        640:   1%|          | 10/809 [00:11<15:54,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.057     0.4603     0.1178        468        640:   1%|▏         | 12/809 [00:14<16:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.084     0.4503     0.1209        173        640:   2%|▏         | 14/809 [00:16<15:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.111     0.4415     0.1254        418        640:   2%|▏         | 15/809 [00:17<15:55,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.131     0.4356     0.1256        210        640:   2%|▏         | 17/809 [00:20<15:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.153     0.4319      0.133        327        640:   2%|▏         | 19/809 [00:22<16:12,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.149     0.4353     0.1315        112        640:   2%|▏         | 20/809 [00:23<15:59,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.149     0.4352      0.129        180        640:   3%|▎         | 22/809 [00:26<15:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.163     0.4324     0.1333        202        640:   3%|▎         | 24/809 [00:28<15:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14      0.438     0.1311        153        640:   3%|▎         | 25/809 [00:29<15:35,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.139     0.4402     0.1299        428        640:   3%|▎         | 27/809 [00:32<15:43,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.147     0.4381     0.1309        368        640:   4%|▎         | 29/809 [00:34<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.133     0.4441      0.131        243        640:   4%|▍         | 31/809 [00:36<15:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.124     0.4487     0.1295        146        640:   4%|▍         | 32/809 [00:38<15:03,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.123     0.4518     0.1299        136        640:   4%|▍         | 34/809 [00:40<15:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.127     0.4501     0.1302        168        640:   4%|▍         | 36/809 [00:42<14:56,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.134     0.4484     0.1305        522        640:   5%|▍         | 37/809 [00:43<15:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.151     0.4451     0.1362        227        640:   5%|▍         | 39/809 [00:46<15:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.146     0.4478     0.1343        216        640:   5%|▌         | 41/809 [00:48<15:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.141     0.4496     0.1333        215        640:   5%|▌         | 42/809 [00:49<15:07,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4501     0.1317        305        640:   5%|▌         | 44/809 [00:52<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.143     0.4501     0.1321        239        640:   6%|▌         | 46/809 [00:54<15:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4508     0.1313        223        640:   6%|▌         | 48/809 [00:56<14:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.133     0.4532     0.1306        150        640:   6%|▌         | 49/809 [00:58<14:43,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.129     0.4539     0.1305        119        640:   6%|▋         | 51/809 [01:00<14:49,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.138     0.4522     0.1312        535        640:   7%|▋         | 53/809 [01:02<14:56,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4514     0.1318        212        640:   7%|▋         | 54/809 [01:03<14:48,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4509      0.131        316        640:   7%|▋         | 56/809 [01:06<14:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.143     0.4503     0.1308        250        640:   7%|▋         | 58/809 [01:08<15:02,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.145     0.4494     0.1309        313        640:   7%|▋         | 59/809 [01:09<14:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.143     0.4495     0.1304        295        640:   8%|▊         | 61/809 [01:12<14:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.147     0.4479     0.1302        291        640:   8%|▊         | 63/809 [01:14<14:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.147     0.4478     0.1294        179        640:   8%|▊         | 65/809 [01:17<14:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.149     0.4475     0.1294        414        640:   8%|▊         | 66/809 [01:18<14:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.146     0.4479     0.1296        144        640:   8%|▊         | 68/809 [01:20<14:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.15     0.4462     0.1291        274        640:   9%|▊         | 70/809 [01:22<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.149     0.4464     0.1286        282        640:   9%|▉         | 71/809 [01:24<14:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156     0.4439     0.1286        336        640:   9%|▉         | 73/809 [01:26<14:51,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.148     0.4464     0.1276         93        640:   9%|▉         | 75/809 [01:28<14:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.143     0.4474     0.1269        183        640:   9%|▉         | 76/809 [01:30<14:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.138     0.4493     0.1259        171        640:  10%|▉         | 78/809 [01:32<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.133     0.4517     0.1254        129        640:  10%|▉         | 80/809 [01:34<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.133     0.4516     0.1253        274        640:  10%|█         | 81/809 [01:35<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.137     0.4505     0.1259        206        640:  10%|█         | 83/809 [01:38<14:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.137     0.4508     0.1259        208        640:  11%|█         | 85/809 [01:40<14:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.135      0.451     0.1253        218        640:  11%|█         | 87/809 [01:43<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4498     0.1267        298        640:  11%|█         | 88/809 [01:44<14:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4494     0.1262        225        640:  11%|█         | 90/809 [01:46<14:35,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.133     0.4517     0.1253        151        640:  11%|█▏        | 92/809 [01:49<14:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.137     0.4502      0.127        507        640:  11%|█▏        | 93/809 [01:50<14:31,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4499     0.1278        390        640:  12%|█▏        | 95/809 [01:52<14:24,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.141     0.4493     0.1279        335        640:  12%|█▏        | 96/809 [01:53<14:19,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.141     0.4491     0.1278        186        640:  12%|█▏        | 98/809 [01:56<14:30,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.14     0.4493     0.1273        206        640:  12%|█▏        | 100/809 [01:58<14:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.142     0.4484     0.1269        231        640:  12%|█▏        | 101/809 [02:00<14:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.145     0.4471     0.1276        301        640:  13%|█▎        | 103/809 [02:02<14:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.144     0.4468      0.127        155        640:  13%|█▎        | 105/809 [02:04<13:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.147     0.4456     0.1281        561        640:  13%|█▎        | 106/809 [02:06<14:33,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.146     0.4458     0.1275        207        640:  13%|█▎        | 108/809 [02:08<13:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.151     0.4444     0.1296        324        640:  14%|█▎        | 110/809 [02:10<14:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.153     0.4436     0.1297        443        640:  14%|█▎        | 111/809 [02:12<14:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156     0.4429     0.1303        474        640:  14%|█▍        | 113/809 [02:14<14:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.154     0.4436     0.1296        188        640:  14%|█▍        | 115/809 [02:17<14:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.153     0.4437     0.1295        283        640:  14%|█▍        | 116/809 [02:18<13:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156      0.443     0.1302        265        640:  15%|█▍        | 118/809 [02:20<13:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156     0.4428     0.1303        348        640:  15%|█▍        | 120/809 [02:23<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.161     0.4417     0.1329        293        640:  15%|█▍        | 121/809 [02:24<13:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.161      0.442     0.1328        308        640:  15%|█▌        | 123/809 [02:26<13:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.158     0.4429     0.1327         66        640:  15%|█▌        | 125/809 [02:29<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.157     0.4435     0.1325        150        640:  16%|█▌        | 126/809 [02:30<13:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.159     0.4433     0.1321        228        640:  16%|█▌        | 128/809 [02:32<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.16     0.4426     0.1319        290        640:  16%|█▌        | 130/809 [02:35<13:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.163     0.4416     0.1318        312        640:  16%|█▌        | 131/809 [02:36<13:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.16     0.4426     0.1313        200        640:  16%|█▋        | 133/809 [02:38<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156     0.4441     0.1307        157        640:  17%|█▋        | 135/809 [02:40<13:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.153      0.445     0.1302        224        640:  17%|█▋        | 137/809 [02:43<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.152     0.4452     0.1298         87        640:  17%|█▋        | 138/809 [02:44<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.156     0.4446     0.1309        174        640:  17%|█▋        | 140/809 [02:46<13:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.159     0.4438     0.1314        322        640:  18%|█▊        | 142/809 [02:49<13:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4423     0.1319        448        640:  18%|█▊        | 144/809 [02:51<13:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.166     0.4418     0.1333        411        640:  18%|█▊        | 145/809 [02:52<13:36,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4417      0.133        261        640:  18%|█▊        | 147/809 [02:55<13:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.17     0.4413     0.1355        231        640:  18%|█▊        | 149/809 [02:57<13:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.168     0.4422     0.1351        167        640:  19%|█▊        | 150/809 [02:58<13:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167      0.442      0.135        240        640:  19%|█▉        | 152/809 [03:01<13:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.168     0.4421     0.1348        186        640:  19%|█▉        | 154/809 [03:03<13:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.169     0.4414      0.135        265        640:  19%|█▉        | 155/809 [03:04<13:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.165     0.4429     0.1346        105        640:  19%|█▉        | 157/809 [03:07<13:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4427     0.1358        276        640:  20%|█▉        | 159/809 [03:09<12:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.169      0.442     0.1361        263        640:  20%|█▉        | 160/809 [03:10<12:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4424     0.1359        275        640:  20%|██        | 162/809 [03:13<12:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4433     0.1355        155        640:  20%|██        | 164/809 [03:15<12:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.165     0.4431     0.1354        242        640:  20%|██        | 165/809 [03:16<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.162     0.4439     0.1349         97        640:  21%|██        | 167/809 [03:19<12:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4433     0.1347        241        640:  21%|██        | 169/809 [03:21<12:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4426     0.1363        397        640:  21%|██        | 170/809 [03:22<13:10,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.166     0.4429     0.1359        211        640:  21%|██▏       | 172/809 [03:25<12:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.166     0.4428     0.1356         75        640:  22%|██▏       | 174/809 [03:27<12:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.165      0.443     0.1352        205        640:  22%|██▏       | 176/809 [03:29<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.166     0.4428     0.1352        176        640:  22%|██▏       | 177/809 [03:31<12:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4434     0.1347        260        640:  22%|██▏       | 179/809 [03:33<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4438     0.1353         70        640:  22%|██▏       | 181/809 [03:35<12:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.164     0.4434     0.1352        189        640:  22%|██▏       | 182/809 [03:36<12:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4427     0.1364        469        640:  23%|██▎       | 184/809 [03:39<12:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.165     0.4432      0.136        202        640:  23%|██▎       | 186/809 [03:41<12:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.167     0.4426     0.1365        445        640:  23%|██▎       | 187/809 [03:43<12:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.17     0.4419     0.1365        339        640:  23%|██▎       | 189/809 [03:45<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.172     0.4412     0.1364        359        640:  24%|██▎       | 191/809 [03:47<12:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4407     0.1365        240        640:  24%|██▎       | 192/809 [03:48<12:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.176     0.4395     0.1369        298        640:  24%|██▍       | 194/809 [03:51<12:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4404     0.1365        165        640:  24%|██▍       | 196/809 [03:53<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4406     0.1364        236        640:  24%|██▍       | 197/809 [03:54<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.171     0.4413      0.136        107        640:  25%|██▍       | 199/809 [03:57<11:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.171     0.4413     0.1358        247        640:  25%|██▍       | 201/809 [03:59<11:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.172     0.4409     0.1359        425        640:  25%|██▍       | 202/809 [04:00<12:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4403     0.1358        347        640:  25%|██▌       | 204/809 [04:03<12:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4401     0.1358        119        640:  25%|██▌       | 206/809 [04:05<12:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4406     0.1357        144        640:  26%|██▌       | 208/809 [04:08<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4405     0.1357        215        640:  26%|██▌       | 209/809 [04:09<11:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.177       0.44     0.1357        145        640:  26%|██▌       | 211/809 [04:11<11:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.179     0.4398     0.1368        349        640:  26%|██▋       | 213/809 [04:14<11:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.18     0.4398     0.1372        339        640:  26%|██▋       | 214/809 [04:15<11:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183     0.4391     0.1373        160        640:  27%|██▋       | 216/809 [04:17<11:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183     0.4393     0.1376        177        640:  27%|██▋       | 218/809 [04:20<11:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183     0.4391     0.1375        267        640:  27%|██▋       | 219/809 [04:21<11:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183      0.439     0.1373        177        640:  27%|██▋       | 221/809 [04:23<11:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.18     0.4405     0.1367        214        640:  28%|██▊       | 223/809 [04:25<11:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.179      0.441     0.1366        245        640:  28%|██▊       | 224/809 [04:27<11:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.179     0.4409     0.1364        124        640:  28%|██▊       | 226/809 [04:29<11:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.179      0.441     0.1365        256        640:  28%|██▊       | 228/809 [04:31<11:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.177     0.4411     0.1362        154        640:  28%|██▊       | 230/809 [04:34<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.177     0.4412      0.136        192        640:  29%|██▊       | 231/809 [04:35<11:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.177     0.4413     0.1358        128        640:  29%|██▉       | 233/809 [04:37<11:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4417     0.1354        177        640:  29%|██▉       | 235/809 [04:40<11:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4419     0.1354        302        640:  29%|██▉       | 236/809 [04:41<11:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.17     0.4423     0.1352        213        640:  29%|██▉       | 238/809 [04:43<11:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.169     0.4427      0.135        271        640:  30%|██▉       | 240/809 [04:45<11:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.169     0.4424      0.135        229        640:  30%|██▉       | 241/809 [04:47<11:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.17      0.442     0.1351        281        640:  30%|███       | 243/809 [04:49<11:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.171     0.4415     0.1352        208        640:  30%|███       | 245/809 [04:51<11:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.172     0.4412     0.1355        342        640:  30%|███       | 246/809 [04:53<11:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4409     0.1356        291        640:  31%|███       | 248/809 [04:55<11:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.175     0.4405     0.1358        293        640:  31%|███       | 250/809 [04:57<11:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4413     0.1357        100        640:  31%|███       | 251/809 [04:59<10:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4417     0.1356        287        640:  31%|███▏      | 253/809 [05:01<10:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4414     0.1354        102        640:  32%|███▏      | 255/809 [05:03<10:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4417     0.1354        277        640:  32%|███▏      | 257/809 [05:06<10:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.174     0.4414     0.1355        204        640:  32%|███▏      | 258/809 [05:07<10:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4419     0.1353        199        640:  32%|███▏      | 260/809 [05:09<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4422     0.1353        191        640:  32%|███▏      | 261/809 [05:11<11:33,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.172     0.4425      0.135        270        640:  33%|███▎      | 263/809 [05:13<11:22,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.173     0.4421     0.1355        203        640:  33%|███▎      | 265/809 [05:15<11:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.175     0.4418     0.1358        311        640:  33%|███▎      | 266/809 [05:17<11:08,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.176     0.4415     0.1357        147        640:  33%|███▎      | 268/809 [05:19<10:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.177      0.441     0.1357        278        640:  33%|███▎      | 270/809 [05:21<10:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.179     0.4404     0.1359        583        640:  33%|███▎      | 271/809 [05:23<11:20,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.181     0.4396     0.1362        437        640:  34%|███▎      | 273/809 [05:25<10:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.182     0.4393     0.1363        349        640:  34%|███▍      | 275/809 [05:28<11:14,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183     0.4393     0.1364        413        640:  34%|███▍      | 276/809 [05:29<11:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.182     0.4394     0.1362        200        640:  34%|███▍      | 278/809 [05:31<10:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.183      0.439     0.1363        200        640:  35%|███▍      | 280/809 [05:34<10:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.185     0.4385     0.1368        457        640:  35%|███▍      | 281/809 [05:35<10:54,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.184     0.4385     0.1367        321        640:  35%|███▍      | 283/809 [05:37<10:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.184     0.4384     0.1366        295        640:  35%|███▌      | 284/809 [05:39<10:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.185     0.4385     0.1371        164        640:  35%|███▌      | 286/809 [05:41<10:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.185     0.4388     0.1374        111        640:  36%|███▌      | 288/809 [05:43<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.185     0.4386     0.1374        149        640:  36%|███▌      | 290/809 [05:46<10:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.184     0.4388     0.1373        133        640:  36%|███▌      | 291/809 [05:47<10:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.184      0.439      0.137        107        640:  36%|███▌      | 293/809 [05:49<10:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.187     0.4379     0.1376        463        640:  36%|███▋      | 295/809 [05:52<10:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.187     0.4382     0.1376         64        640:  37%|███▋      | 296/809 [05:53<10:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.187     0.4379     0.1375        336        640:  37%|███▋      | 298/809 [05:55<10:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.188     0.4376     0.1374        265        640:  37%|███▋      | 300/809 [05:58<10:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.187     0.4376     0.1373        283        640:  37%|███▋      | 301/809 [05:59<10:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.187     0.4376     0.1373        136        640:  37%|███▋      | 303/809 [06:01<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.186     0.4381     0.1371        295        640:  38%|███▊      | 305/809 [06:04<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.186     0.4381     0.1371        220        640:  38%|███▊      | 306/809 [06:05<10:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.189     0.4372     0.1383        341        640:  38%|███▊      | 308/809 [06:07<10:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.189     0.4368     0.1383        307        640:  38%|███▊      | 310/809 [06:10<10:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.19     0.4365     0.1385        246        640:  39%|███▊      | 312/809 [06:12<10:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.192     0.4359     0.1396        397        640:  39%|███▊      | 313/809 [06:14<10:20,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4357     0.1397        302        640:  39%|███▉      | 315/809 [06:16<10:12,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4353     0.1398        428        640:  39%|███▉      | 316/809 [06:17<10:10,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195      0.435     0.1403        420        640:  39%|███▉      | 318/809 [06:20<10:09,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4349     0.1404        320        640:  40%|███▉      | 320/809 [06:22<09:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4346     0.1407        448        640:  40%|███▉      | 321/809 [06:23<09:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.198     0.4339     0.1416        287        640:  40%|███▉      | 323/809 [06:26<10:02,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4331     0.1423        283        640:  40%|████      | 325/809 [06:28<09:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4333     0.1422        225        640:  40%|████      | 326/809 [06:30<09:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4334     0.1421        258        640:  41%|████      | 328/809 [06:32<09:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4328     0.1423        513        640:  41%|████      | 330/809 [06:34<09:49,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4331     0.1422        209        640:  41%|████      | 331/809 [06:36<09:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4325     0.1423        437        640:  41%|████      | 333/809 [06:38<09:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204      0.432     0.1427        499        640:  41%|████▏     | 335/809 [06:40<09:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4319     0.1426        342        640:  42%|████▏     | 336/809 [06:42<09:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4322     0.1423        192        640:  42%|████▏     | 338/809 [06:44<09:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4323     0.1423        262        640:  42%|████▏     | 340/809 [06:46<09:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4326     0.1423        225        640:  42%|████▏     | 341/809 [06:47<09:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4325     0.1423        259        640:  42%|████▏     | 343/809 [06:50<09:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4322     0.1422        271        640:  43%|████▎     | 345/809 [06:52<09:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4321     0.1421        325        640:  43%|████▎     | 346/809 [06:54<09:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4322     0.1421        209        640:  43%|████▎     | 348/809 [06:56<09:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4323      0.142        342        640:  43%|████▎     | 350/809 [06:58<09:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4327     0.1417        168        640:  44%|████▎     | 352/809 [07:01<08:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4329     0.1416        236        640:  44%|████▎     | 353/809 [07:02<08:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4324     0.1421        592        640:  44%|████▍     | 355/809 [07:04<09:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4326     0.1419        283        640:  44%|████▍     | 357/809 [07:07<08:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4324     0.1418        228        640:  44%|████▍     | 358/809 [07:08<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199      0.433     0.1415        205        640:  44%|████▍     | 360/809 [07:10<08:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199      0.433     0.1414        299        640:  45%|████▍     | 362/809 [07:12<08:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4329     0.1416        382        640:  45%|████▍     | 363/809 [07:14<08:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2      0.433     0.1423        447        640:  45%|████▌     | 365/809 [07:16<09:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4329     0.1425        578        640:  45%|████▌     | 367/809 [07:19<09:01,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4329     0.1424        324        640:  45%|████▌     | 368/809 [07:20<08:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4326     0.1426        286        640:  46%|████▌     | 370/809 [07:22<08:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4329     0.1424        314        640:  46%|████▌     | 372/809 [07:25<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2      0.433     0.1424        343        640:  46%|████▌     | 373/809 [07:26<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4331     0.1424        269        640:  46%|████▋     | 375/809 [07:28<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4325     0.1426        215        640:  47%|████▋     | 377/809 [07:30<08:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4324     0.1425        246        640:  47%|████▋     | 378/809 [07:32<08:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4319     0.1428        208        640:  47%|████▋     | 380/809 [07:34<08:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203      0.432     0.1426        269        640:  47%|████▋     | 382/809 [07:37<08:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4317     0.1428        361        640:  47%|████▋     | 383/809 [07:38<08:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4318     0.1427        170        640:  48%|████▊     | 385/809 [07:40<08:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4316     0.1428        491        640:  48%|████▊     | 387/809 [07:43<08:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4319     0.1427        179        640:  48%|████▊     | 388/809 [07:44<08:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4317     0.1427        179        640:  48%|████▊     | 390/809 [07:46<08:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4318     0.1425        196        640:  48%|████▊     | 392/809 [07:49<08:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4319     0.1424        227        640:  49%|████▊     | 393/809 [07:50<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4322     0.1421         83        640:  49%|████▉     | 395/809 [07:52<08:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4324     0.1419        142        640:  49%|████▉     | 397/809 [07:54<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4325     0.1417        188        640:  49%|████▉     | 399/809 [07:57<07:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4326     0.1417        266        640:  49%|████▉     | 400/809 [07:58<08:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.198      0.433     0.1415        136        640:  50%|████▉     | 402/809 [08:00<08:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4333     0.1413        102        640:  50%|████▉     | 404/809 [08:03<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4334     0.1413        235        640:  50%|█████     | 405/809 [08:04<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4335     0.1411        162        640:  50%|█████     | 407/809 [08:06<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4336     0.1409        243        640:  51%|█████     | 409/809 [08:08<07:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4339     0.1407         57        640:  51%|█████     | 410/809 [08:10<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4338     0.1409         81        640:  51%|█████     | 412/809 [08:12<07:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4342     0.1407        211        640:  51%|█████     | 414/809 [08:14<07:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4343     0.1406        258        640:  51%|█████▏    | 416/809 [08:17<07:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4345     0.1404        138        640:  52%|█████▏    | 417/809 [08:18<07:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4347     0.1402        192        640:  52%|█████▏    | 419/809 [08:20<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4344     0.1401        198        640:  52%|█████▏    | 421/809 [08:23<07:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4344       0.14        236        640:  52%|█████▏    | 422/809 [08:24<07:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4345     0.1398        349        640:  52%|█████▏    | 424/809 [08:26<07:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4343     0.1397        159        640:  53%|█████▎    | 426/809 [08:29<07:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195      0.434     0.1397        401        640:  53%|█████▎    | 427/809 [08:30<07:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4341     0.1396        389        640:  53%|█████▎    | 429/809 [08:32<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195      0.434     0.1395        268        640:  53%|█████▎    | 431/809 [08:35<07:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4337     0.1395        429        640:  53%|█████▎    | 432/809 [08:36<07:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4331       0.14        654        640:  54%|█████▎    | 434/809 [08:38<07:51,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4329     0.1399        349        640:  54%|█████▍    | 436/809 [08:41<07:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4328     0.1398        275        640:  54%|█████▍    | 437/809 [08:42<07:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4325     0.1399        411        640:  54%|█████▍    | 439/809 [08:44<07:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4318       0.14        552        640:  55%|█████▍    | 441/809 [08:47<07:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4319       0.14        163        640:  55%|█████▍    | 442/809 [08:48<07:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4321     0.1398        140        640:  55%|█████▍    | 444/809 [08:50<07:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.432     0.1397        151        640:  55%|█████▌    | 446/809 [08:53<07:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4321     0.1396        243        640:  55%|█████▌    | 447/809 [08:54<07:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4321     0.1396        294        640:  56%|█████▌    | 449/809 [08:56<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203      0.432     0.1395        359        640:  56%|█████▌    | 451/809 [08:59<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4321     0.1394        311        640:  56%|█████▌    | 452/809 [09:00<07:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4321     0.1393        390        640:  56%|█████▌    | 454/809 [09:02<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4319     0.1393        353        640:  56%|█████▋    | 456/809 [09:04<06:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4319     0.1391        220        640:  57%|█████▋    | 458/809 [09:07<06:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4318     0.1393        510        640:  57%|█████▋    | 459/809 [09:08<07:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4317     0.1392         76        640:  57%|█████▋    | 461/809 [09:10<06:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205      0.432     0.1391        186        640:  57%|█████▋    | 463/809 [09:13<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205      0.432     0.1391        258        640:  57%|█████▋    | 464/809 [09:14<06:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4323      0.139        154        640:  58%|█████▊    | 466/809 [09:16<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4328     0.1387        114        640:  58%|█████▊    | 468/809 [09:19<06:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4331     0.1385        150        640:  58%|█████▊    | 470/809 [09:21<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4332     0.1385        365        640:  58%|█████▊    | 471/809 [09:22<06:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4337     0.1383        103        640:  58%|█████▊    | 473/809 [09:25<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4336     0.1382        198        640:  59%|█████▊    | 475/809 [09:27<06:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4335     0.1387        340        640:  59%|█████▉    | 476/809 [09:28<06:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4337     0.1386        195        640:  59%|█████▉    | 478/809 [09:31<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.198     0.4342     0.1385        134        640:  59%|█████▉    | 480/809 [09:33<06:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4347     0.1384        230        640:  60%|█████▉    | 482/809 [09:35<06:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4349     0.1383        174        640:  60%|█████▉    | 483/809 [09:37<06:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4351     0.1381        178        640:  60%|█████▉    | 485/809 [09:39<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4352      0.138        249        640:  60%|██████    | 487/809 [09:41<06:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4352      0.138        482        640:  60%|██████    | 488/809 [09:43<06:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4357     0.1378        239        640:  61%|██████    | 490/809 [09:45<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4359     0.1377        274        640:  61%|██████    | 492/809 [09:47<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4361     0.1377        165        640:  61%|██████    | 493/809 [09:48<06:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.192     0.4362     0.1375        184        640:  61%|██████    | 495/809 [09:51<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193      0.436     0.1376        427        640:  61%|██████▏   | 497/809 [09:53<06:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4359     0.1377        521        640:  62%|██████▏   | 498/809 [09:55<06:26,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4356      0.138        333        640:  62%|██████▏   | 500/809 [09:57<06:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4356      0.138        187        640:  62%|██████▏   | 502/809 [09:59<06:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4354      0.138        227        640:  62%|██████▏   | 503/809 [10:01<06:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4355      0.138        233        640:  62%|██████▏   | 505/809 [10:03<05:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4356     0.1379        271        640:  63%|██████▎   | 507/809 [10:05<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4356     0.1378        200        640:  63%|██████▎   | 509/809 [10:08<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4354     0.1379        288        640:  63%|██████▎   | 510/809 [10:09<05:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4352     0.1379        174        640:  63%|██████▎   | 512/809 [10:11<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196      0.435     0.1382        388        640:  64%|██████▎   | 514/809 [10:14<05:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4353     0.1381         92        640:  64%|██████▎   | 515/809 [10:15<05:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4356      0.138        214        640:  64%|██████▍   | 517/809 [10:17<05:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4359     0.1378        241        640:  64%|██████▍   | 519/809 [10:19<05:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4356      0.138        686        640:  64%|██████▍   | 520/809 [10:21<05:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4354     0.1381        438        640:  65%|██████▍   | 522/809 [10:23<05:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4353     0.1381        180        640:  65%|██████▍   | 524/809 [10:25<05:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4355      0.138        199        640:  65%|██████▍   | 525/809 [10:27<05:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4358      0.138        141        640:  65%|██████▌   | 527/809 [10:29<05:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4358     0.1383        149        640:  65%|██████▌   | 529/809 [10:32<05:48,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4358     0.1383        347        640:  66%|██████▌   | 530/809 [10:33<05:48,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4358     0.1383        257        640:  66%|██████▌   | 532/809 [10:35<05:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4363     0.1382         65        640:  66%|██████▌   | 534/809 [10:38<05:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4362     0.1383        325        640:  66%|██████▌   | 535/809 [10:39<05:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4365     0.1382        246        640:  66%|██████▋   | 537/809 [10:41<05:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4366      0.138        120        640:  67%|██████▋   | 539/809 [10:44<05:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4368      0.138         59        640:  67%|██████▋   | 540/809 [10:45<05:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4364     0.1382        396        640:  67%|██████▋   | 542/809 [10:47<05:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4363     0.1382        144        640:  67%|██████▋   | 544/809 [10:49<05:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4362     0.1381        336        640:  67%|██████▋   | 545/809 [10:51<05:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196      0.436     0.1384        478        640:  68%|██████▊   | 547/809 [10:53<05:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4358     0.1385        179        640:  68%|██████▊   | 549/809 [10:56<05:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4359     0.1386        192        640:  68%|██████▊   | 550/809 [10:57<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.198     0.4357     0.1395        219        640:  68%|██████▊   | 552/809 [10:59<05:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4359     0.1395        323        640:  68%|██████▊   | 554/809 [11:02<05:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4361     0.1394        216        640:  69%|██████▊   | 555/809 [11:03<05:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4361     0.1394        278        640:  69%|██████▉   | 557/809 [11:05<04:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4363     0.1395        451        640:  69%|██████▉   | 559/809 [11:08<05:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4363     0.1395        170        640:  69%|██████▉   | 560/809 [11:09<04:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4363     0.1394        184        640:  69%|██████▉   | 562/809 [11:11<04:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4367     0.1393        151        640:  70%|██████▉   | 564/809 [11:13<04:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4368     0.1392        283        640:  70%|██████▉   | 565/809 [11:15<04:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195      0.437     0.1391        257        640:  70%|███████   | 567/809 [11:17<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4372      0.139        441        640:  70%|███████   | 569/809 [11:19<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4374     0.1391        192        640:  71%|███████   | 571/809 [11:22<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377      0.139        117        640:  71%|███████   | 572/809 [11:23<04:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377     0.1388        261        640:  71%|███████   | 574/809 [11:25<04:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.192     0.4379     0.1387        179        640:  71%|███████   | 576/809 [11:28<04:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4381     0.1386        148        640:  71%|███████▏  | 577/809 [11:29<04:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191      0.438     0.1387        446        640:  72%|███████▏  | 579/809 [11:31<04:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4381     0.1386        261        640:  72%|███████▏  | 581/809 [11:34<04:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191      0.438     0.1386        237        640:  72%|███████▏  | 582/809 [11:35<04:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4383      0.139        119        640:  72%|███████▏  | 584/809 [11:37<04:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4384     0.1389        388        640:  72%|███████▏  | 586/809 [11:40<04:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4383      0.139        204        640:  73%|███████▎  | 587/809 [11:41<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.19     0.4384     0.1389        201        640:  73%|███████▎  | 589/809 [11:43<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G       1.19     0.4385     0.1388        304        640:  73%|███████▎  | 591/809 [11:46<04:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4383     0.1391        439        640:  73%|███████▎  | 592/809 [11:47<04:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191     0.4381     0.1394        296        640:  73%|███████▎  | 594/809 [11:49<04:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191      0.438     0.1394        331        640:  74%|███████▎  | 596/809 [11:52<04:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.191      0.438     0.1394        362        640:  74%|███████▍  | 597/809 [11:53<04:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.192      0.438     0.1393        210        640:  74%|███████▍  | 599/809 [11:55<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.192     0.4377     0.1394        339        640:  74%|███████▍  | 601/809 [11:58<04:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4376     0.1397        341        640:  74%|███████▍  | 602/809 [11:59<04:14,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4375     0.1397        344        640:  75%|███████▍  | 604/809 [12:01<04:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4375     0.1397        276        640:  75%|███████▍  | 606/809 [12:04<04:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4375     0.1397        199        640:  75%|███████▌  | 607/809 [12:05<04:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4374     0.1396        332        640:  75%|███████▌  | 609/809 [12:07<04:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377     0.1396        254        640:  76%|███████▌  | 611/809 [12:10<03:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377     0.1395        180        640:  76%|███████▌  | 612/809 [12:11<03:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4375     0.1397        198        640:  76%|███████▌  | 614/809 [12:13<03:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4376     0.1397        280        640:  76%|███████▌  | 616/809 [12:16<03:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4374     0.1397        319        640:  76%|███████▋  | 617/809 [12:17<03:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4373     0.1397        170        640:  77%|███████▋  | 619/809 [12:19<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4372     0.1396        218        640:  77%|███████▋  | 621/809 [12:22<03:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195      0.437     0.1398        559        640:  77%|███████▋  | 622/809 [12:23<03:51,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4371     0.1398        151        640:  77%|███████▋  | 624/809 [12:25<03:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4373     0.1397        232        640:  77%|███████▋  | 626/809 [12:28<03:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4372       0.14        584        640:  78%|███████▊  | 628/809 [12:30<03:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4375     0.1399        166        640:  78%|███████▊  | 629/809 [12:31<03:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4376     0.1397         92        640:  78%|███████▊  | 631/809 [12:34<03:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4374     0.1399        192        640:  78%|███████▊  | 633/809 [12:36<03:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4373     0.1399        188        640:  78%|███████▊  | 634/809 [12:37<03:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4373     0.1401        348        640:  79%|███████▊  | 636/809 [12:40<03:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4376     0.1399        149        640:  79%|███████▉  | 638/809 [12:42<03:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377     0.1398        229        640:  79%|███████▉  | 640/809 [12:45<03:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.193     0.4377     0.1399        374        640:  79%|███████▉  | 641/809 [12:46<03:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4373     0.1404        220        640:  79%|███████▉  | 643/809 [12:48<03:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.194     0.4372     0.1404        224        640:  80%|███████▉  | 644/809 [12:49<03:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.195     0.4371     0.1406         54        640:  80%|███████▉  | 646/809 [12:52<03:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.196     0.4366     0.1406        285        640:  80%|████████  | 648/809 [12:54<03:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.197     0.4363     0.1407        345        640:  80%|████████  | 649/809 [12:55<03:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4358     0.1413        470        640:  80%|████████  | 651/809 [12:58<03:15,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4354     0.1415        321        640:  81%|████████  | 653/809 [13:00<03:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4354     0.1414        262        640:  81%|████████  | 654/809 [13:02<03:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4355     0.1413        216        640:  81%|████████  | 656/809 [13:04<03:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4354     0.1413        138        640:  81%|████████▏ | 658/809 [13:06<03:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4355     0.1413        293        640:  81%|████████▏ | 659/809 [13:08<03:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.199     0.4356     0.1411        232        640:  82%|████████▏ | 663/809 [13:12<02:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4354     0.1413        460        640:  82%|████████▏ | 664/809 [13:14<02:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4353     0.1413        323        640:  82%|████████▏ | 666/809 [13:16<02:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4352     0.1414        256        640:  83%|████████▎ | 668/809 [13:18<02:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4352     0.1413        218        640:  83%|████████▎ | 669/809 [13:20<02:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4353     0.1413        242        640:  83%|████████▎ | 671/809 [13:22<02:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4353     0.1413        270        640:  83%|████████▎ | 673/809 [13:24<02:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G        1.2     0.4354     0.1412        166        640:  83%|████████▎ | 674/809 [13:25<02:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201      0.435     0.1413        370        640:  84%|████████▎ | 676/809 [13:28<02:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4347     0.1414        197        640:  84%|████████▍ | 678/809 [13:30<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4344     0.1414        235        640:  84%|████████▍ | 680/809 [13:33<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204      0.434     0.1421        733        640:  84%|████████▍ | 681/809 [13:34<02:39,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4339     0.1421        310        640:  84%|████████▍ | 683/809 [13:36<02:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204      0.434      0.142        238        640:  85%|████████▍ | 684/809 [13:38<02:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4343     0.1419        215        640:  85%|████████▍ | 686/809 [13:40<02:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4345     0.1418        212        640:  85%|████████▌ | 688/809 [13:42<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4351     0.1416        114        640:  85%|████████▌ | 690/809 [13:45<02:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4349     0.1416        206        640:  85%|████████▌ | 691/809 [13:46<02:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4345     0.1415        170        640:  86%|████████▌ | 693/809 [13:48<02:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4339     0.1419        528        640:  86%|████████▌ | 695/809 [13:51<02:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4335      0.142        387        640:  86%|████████▌ | 696/809 [13:52<02:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.207     0.4332     0.1419        149        640:  86%|████████▋ | 698/809 [13:54<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1418        337        640:  87%|████████▋ | 700/809 [13:57<02:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1418        333        640:  87%|████████▋ | 701/809 [13:58<02:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4335     0.1416        211        640:  87%|████████▋ | 703/809 [14:00<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4333     0.1416        224        640:  87%|████████▋ | 705/809 [14:03<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1417        330        640:  87%|████████▋ | 706/809 [14:04<02:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206      0.433     0.1419        360        640:  88%|████████▊ | 708/809 [14:06<02:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1417        220        640:  88%|████████▊ | 710/809 [14:09<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4331     0.1418        370        640:  88%|████████▊ | 711/809 [14:10<01:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.207     0.4328     0.1418        293        640:  88%|████████▊ | 713/809 [14:12<01:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4333     0.1416        124        640:  88%|████████▊ | 715/809 [14:15<01:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1418        518        640:  89%|████████▊ | 716/809 [14:16<01:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4332     0.1419        260        640:  89%|████████▉ | 718/809 [14:18<01:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4331     0.1418        326        640:  89%|████████▉ | 720/809 [14:21<01:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4331     0.1417        101        640:  89%|████████▉ | 721/809 [14:22<01:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4331     0.1416         76        640:  89%|████████▉ | 723/809 [14:24<01:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205      0.433     0.1415        330        640:  90%|████████▉ | 725/809 [14:26<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4329     0.1416        375        640:  90%|████████▉ | 726/809 [14:28<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4328     0.1415        257        640:  90%|████████▉ | 728/809 [14:30<01:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4329     0.1413        108        640:  90%|█████████ | 730/809 [14:32<01:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4327     0.1419        572        640:  90%|█████████ | 731/809 [14:34<01:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4327     0.1418        382        640:  91%|█████████ | 733/809 [14:36<01:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.207     0.4327     0.1423        502        640:  91%|█████████ | 735/809 [14:39<01:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.207     0.4328     0.1422        350        640:  91%|█████████ | 736/809 [14:40<01:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4328     0.1421        306        640:  91%|█████████ | 738/809 [14:42<01:27,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206      0.433      0.142        124        640:  91%|█████████▏| 740/809 [14:45<01:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4331      0.142        265        640:  92%|█████████▏| 741/809 [14:46<01:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4331     0.1419        545        640:  92%|█████████▏| 743/809 [14:48<01:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4331     0.1418        185        640:  92%|█████████▏| 745/809 [14:51<01:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206      0.433     0.1419        402        640:  92%|█████████▏| 746/809 [14:52<01:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206      0.433     0.1419        313        640:  92%|█████████▏| 748/809 [14:54<01:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.207     0.4328     0.1425        403        640:  93%|█████████▎| 750/809 [14:57<01:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206     0.4328     0.1424        262        640:  93%|█████████▎| 751/809 [14:58<01:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.206      0.433     0.1423        277        640:  93%|█████████▎| 753/809 [15:00<01:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.205     0.4331     0.1424        442        640:  93%|█████████▎| 755/809 [15:03<01:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4334     0.1422        245        640:  94%|█████████▎| 757/809 [15:05<01:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.204     0.4335     0.1422        188        640:  94%|█████████▎| 758/809 [15:06<00:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4338      0.142        137        640:  94%|█████████▍| 760/809 [15:08<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4337     0.1422        298        640:  94%|█████████▍| 762/809 [15:11<00:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.203     0.4337     0.1422        202        640:  94%|█████████▍| 763/809 [15:12<00:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.434      0.142        160        640:  95%|█████████▍| 765/809 [15:14<00:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.434     0.1421        253        640:  95%|█████████▍| 767/809 [15:17<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4341      0.142        220        640:  95%|█████████▍| 768/809 [15:18<00:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4342     0.1419        221        640:  95%|█████████▌| 770/809 [15:20<00:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339     0.1421        361        640:  95%|█████████▌| 772/809 [15:23<00:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339     0.1421        161        640:  96%|█████████▌| 774/809 [15:25<00:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.434     0.1421        238        640:  96%|█████████▌| 775/809 [15:26<00:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339      0.142        293        640:  96%|█████████▌| 777/809 [15:28<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339      0.142        351        640:  96%|█████████▋| 779/809 [15:31<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339      0.142        280        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4337      0.142        244        640:  97%|█████████▋| 782/809 [15:34<00:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4338      0.142        190        640:  97%|█████████▋| 784/809 [15:37<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339     0.1419        151        640:  97%|█████████▋| 785/809 [15:38<00:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201      0.434     0.1419        369        640:  97%|█████████▋| 787/809 [15:40<00:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4342     0.1419        266        640:  98%|█████████▊| 789/809 [15:43<00:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4342     0.1418        200        640:  98%|█████████▊| 791/809 [15:45<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.434      0.142        387        640:  98%|█████████▊| 792/809 [15:46<00:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202     0.4339      0.142        452        640:  98%|█████████▊| 794/809 [15:49<00:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201      0.434      0.142        198        640:  98%|█████████▊| 795/809 [15:51<00:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201      0.434      0.142        264        640:  99%|█████████▊| 797/809 [15:52<00:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.202      0.434     0.1421        542        640:  99%|█████████▉| 799/809 [15:55<00:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201      0.434     0.1421        279        640:  99%|█████████▉| 801/809 [15:57<00:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4342      0.142        155        640:  99%|█████████▉| 802/809 [15:58<00:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4343      0.142        230        640:  99%|█████████▉| 804/809 [16:01<00:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4343      0.142        247        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     40/300      13.7G      1.201     0.4342     0.1421        491        640: 100%|█████████▉| 807/809 [16:04<00:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   3%|▎         | 2/69 [00:00<00:17,  3.83it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  38%|███▊      | 26/69 [00:06<00:10,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:09,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:07,  3.69it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:05,  3.52it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.248     0.4297     0.1365        332        640:   0%|          | 1/809 [00:01<16:24,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.066     0.4519     0.1252        220        640:   0%|          | 3/809 [00:03<16:07,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.057     0.4608     0.1244        221        640:   1%|          | 5/809 [00:06<16:21,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.057     0.4495     0.1159        179        640:   1%|          | 7/809 [00:08<15:52,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.031     0.4497     0.1118        196        640:   1%|          | 8/809 [00:09<15:55,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.051     0.4474     0.1116        447        640:   1%|          | 10/809 [00:12<16:07,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.036     0.4552     0.1124         99        640:   1%|▏         | 12/809 [00:14<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.041     0.4548     0.1101        272        640:   2%|▏         | 14/809 [00:16<15:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.079     0.4454     0.1167        366        640:   2%|▏         | 15/809 [00:18<15:58,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.053     0.4582      0.114        157        640:   2%|▏         | 17/809 [00:20<15:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.064     0.4564     0.1141        258        640:   2%|▏         | 18/809 [00:21<15:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.065     0.4587     0.1134        232        640:   2%|▏         | 20/809 [00:24<15:47,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.066     0.4638     0.1131        257        640:   3%|▎         | 22/809 [00:26<15:48,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.06      0.467     0.1122        114        640:   3%|▎         | 24/809 [00:28<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.063     0.4665     0.1122        251        640:   3%|▎         | 25/809 [00:30<16:02,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.047     0.4741     0.1103        138        640:   3%|▎         | 27/809 [00:32<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.043     0.4747       0.11        162        640:   4%|▎         | 29/809 [00:34<15:19,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.043     0.4752     0.1099        221        640:   4%|▎         | 30/809 [00:36<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.048     0.4734     0.1119        428        640:   4%|▍         | 32/809 [00:38<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.053     0.4734     0.1116        196        640:   4%|▍         | 34/809 [00:40<15:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.047     0.4763     0.1121        116        640:   4%|▍         | 35/809 [00:41<15:02,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.05     0.4769      0.112        297        640:   5%|▍         | 37/809 [00:44<15:07,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.053     0.4751     0.1132        223        640:   5%|▍         | 39/809 [00:46<15:02,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.052     0.4756     0.1129        165        640:   5%|▍         | 40/809 [00:47<14:54,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.066     0.4723     0.1168        660        640:   5%|▌         | 42/809 [00:50<15:39,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.058      0.473      0.116        268        640:   5%|▌         | 44/809 [00:52<15:10,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.06     0.4725     0.1158        429        640:   6%|▌         | 45/809 [00:53<15:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.052     0.4747     0.1151        312        640:   6%|▌         | 47/809 [00:56<15:21,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.062     0.4726     0.1187        523        640:   6%|▌         | 49/809 [00:58<15:50,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.067     0.4712     0.1202        309        640:   6%|▌         | 50/809 [00:59<15:37,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.073     0.4706      0.124        428        640:   6%|▋         | 52/809 [01:02<15:36,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.084     0.4668     0.1256        391        640:   7%|▋         | 54/809 [01:04<15:22,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.076      0.469      0.125        104        640:   7%|▋         | 55/809 [01:06<15:04,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.085     0.4664     0.1267        267        640:   7%|▋         | 57/809 [01:08<15:15,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.092     0.4644     0.1275        231        640:   7%|▋         | 59/809 [01:10<14:47,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.097     0.4631     0.1293        314        640:   7%|▋         | 60/809 [01:11<14:46,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.103     0.4611     0.1308        552        640:   8%|▊         | 62/809 [01:14<15:07,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.113     0.4578     0.1318        261        640:   8%|▊         | 64/809 [01:16<15:01,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.111      0.459     0.1312        170        640:   8%|▊         | 65/809 [01:18<15:02,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.111     0.4583     0.1306        252        640:   8%|▊         | 67/809 [01:20<14:55,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.117     0.4571     0.1317        404        640:   9%|▊         | 69/809 [01:22<14:50,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.117     0.4566     0.1314        232        640:   9%|▊         | 70/809 [01:24<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.12     0.4551      0.132        389        640:   9%|▉         | 72/809 [01:26<14:54,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.122     0.4545     0.1316        448        640:   9%|▉         | 74/809 [01:28<14:48,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.121     0.4543     0.1311        235        640:   9%|▉         | 75/809 [01:30<14:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.13     0.4516     0.1363        569        640:  10%|▉         | 77/809 [01:32<15:17,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.132     0.4511     0.1365        414        640:  10%|▉         | 78/809 [01:33<15:11,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.132     0.4508     0.1359        226        640:  10%|▉         | 80/809 [01:36<14:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.136     0.4496     0.1359        401        640:  10%|█         | 82/809 [01:38<14:54,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.134       0.45     0.1351        238        640:  10%|█         | 83/809 [01:39<14:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.137     0.4489     0.1353        189        640:  11%|█         | 85/809 [01:42<15:10,  1.26s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.135     0.4497     0.1344        263        640:  11%|█         | 87/809 [01:44<14:31,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.132     0.4511     0.1341        288        640:  11%|█         | 88/809 [01:46<14:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.129     0.4522     0.1347        314        640:  11%|█         | 90/809 [01:48<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.132     0.4515      0.136        296        640:  11%|█▏        | 92/809 [01:50<14:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.13     0.4524     0.1354        152        640:  11%|█▏        | 93/809 [01:51<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.131     0.4519     0.1354        328        640:  12%|█▏        | 95/809 [01:54<14:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.136     0.4499     0.1359        310        640:  12%|█▏        | 97/809 [01:56<14:20,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.139     0.4491     0.1359        272        640:  12%|█▏        | 98/809 [01:58<14:19,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.143     0.4483     0.1371        434        640:  12%|█▏        | 100/809 [02:00<14:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.145     0.4483     0.1367        235        640:  13%|█▎        | 102/809 [02:02<14:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.149     0.4469     0.1367        392        640:  13%|█▎        | 103/809 [02:03<13:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.153     0.4461     0.1367        506        640:  13%|█▎        | 105/809 [02:06<14:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.158     0.4446     0.1366        145        640:  13%|█▎        | 107/809 [02:08<13:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.16     0.4438     0.1365        234        640:  13%|█▎        | 108/809 [02:09<13:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.165     0.4423     0.1369        274        640:  14%|█▎        | 110/809 [02:12<13:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4398     0.1369        378        640:  14%|█▍        | 112/809 [02:14<13:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.173     0.4396     0.1366        240        640:  14%|█▍        | 114/809 [02:17<13:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4392     0.1372        256        640:  14%|█▍        | 115/809 [02:18<13:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4393     0.1378        126        640:  14%|█▍        | 117/809 [02:20<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4391     0.1376        271        640:  15%|█▍        | 119/809 [02:22<13:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4392     0.1373        253        640:  15%|█▍        | 120/809 [02:24<13:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4393     0.1368        301        640:  15%|█▌        | 122/809 [02:26<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4397      0.137        208        640:  15%|█▌        | 124/809 [02:28<13:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4396     0.1367        288        640:  15%|█▌        | 125/809 [02:30<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177       0.44     0.1372        119        640:  16%|█▌        | 127/809 [02:32<13:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4408      0.137        232        640:  16%|█▌        | 129/809 [02:34<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181       0.44     0.1389        250        640:  16%|█▌        | 131/809 [02:37<13:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4408     0.1384        155        640:  16%|█▋        | 132/809 [02:38<13:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.18     0.4401     0.1388        325        640:  17%|█▋        | 134/809 [02:40<13:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4413     0.1383        255        640:  17%|█▋        | 136/809 [02:43<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.175     0.4422     0.1378        154        640:  17%|█▋        | 137/809 [02:44<13:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4422     0.1381        208        640:  17%|█▋        | 139/809 [02:46<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4412     0.1385        306        640:  17%|█▋        | 141/809 [02:48<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.178     0.4407     0.1384        231        640:  18%|█▊        | 142/809 [02:50<12:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.18     0.4399     0.1393        260        640:  18%|█▊        | 144/809 [02:52<13:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.179     0.4401     0.1385        154        640:  18%|█▊        | 146/809 [02:54<12:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.177     0.4418     0.1384        284        640:  18%|█▊        | 148/809 [02:57<12:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4426     0.1383         54        640:  18%|█▊        | 149/809 [02:58<12:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.173     0.4435     0.1381        167        640:  19%|█▊        | 151/809 [03:00<12:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4433     0.1379        308        640:  19%|█▉        | 153/809 [03:03<13:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4426     0.1379        308        640:  19%|█▉        | 155/809 [03:05<13:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.175     0.4419     0.1379        221        640:  19%|█▉        | 156/809 [03:06<13:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4424     0.1375         61        640:  20%|█▉        | 158/809 [03:09<12:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4412     0.1373        154        640:  20%|█▉        | 160/809 [03:11<12:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4414     0.1368        181        640:  20%|██        | 162/809 [03:13<12:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4417     0.1367        115        640:  20%|██        | 163/809 [03:14<12:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.175      0.441      0.137        379        640:  20%|██        | 165/809 [03:17<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.175     0.4413     0.1368        175        640:  21%|██        | 167/809 [03:19<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4411     0.1364        136        640:  21%|██        | 168/809 [03:20<12:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.173     0.4415      0.136        187        640:  21%|██        | 170/809 [03:23<12:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4423     0.1359        140        640:  21%|██▏       | 172/809 [03:25<12:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4425     0.1359        291        640:  21%|██▏       | 173/809 [03:26<12:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4427     0.1356        199        640:  22%|██▏       | 175/809 [03:29<12:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4432     0.1368        433        640:  22%|██▏       | 177/809 [03:31<12:54,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.168     0.4435     0.1366        111        640:  22%|██▏       | 178/809 [03:32<12:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4424     0.1366        249        640:  22%|██▏       | 180/809 [03:35<12:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.169     0.4428     0.1364        271        640:  22%|██▏       | 182/809 [03:37<12:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172      0.442     0.1365        397        640:  23%|██▎       | 184/809 [03:39<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4422     0.1364        197        640:  23%|██▎       | 185/809 [03:41<12:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.169     0.4428     0.1361        169        640:  23%|██▎       | 187/809 [03:43<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.17     0.4428      0.136        322        640:  23%|██▎       | 189/809 [03:45<12:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4423     0.1359        335        640:  23%|██▎       | 190/809 [03:46<12:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.17     0.4427     0.1356        153        640:  24%|██▎       | 192/809 [03:49<12:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4423     0.1355        235        640:  24%|██▍       | 194/809 [03:51<12:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.171     0.4424     0.1353        206        640:  24%|██▍       | 195/809 [03:52<12:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4424     0.1352        356        640:  24%|██▍       | 197/809 [03:55<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.172     0.4422      0.135        384        640:  25%|██▍       | 199/809 [03:57<11:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.175     0.4413     0.1358        409        640:  25%|██▍       | 200/809 [03:58<12:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.178     0.4408     0.1362        254        640:  25%|██▍       | 202/809 [04:01<12:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.179     0.4404     0.1362        438        640:  25%|██▌       | 204/809 [04:03<12:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.179     0.4402     0.1361        278        640:  25%|██▌       | 205/809 [04:04<12:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181     0.4399     0.1364        421        640:  26%|██▌       | 207/809 [04:07<13:17,  1.32s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181     0.4402     0.1362        270        640:  26%|██▌       | 208/809 [04:08<12:47,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.182     0.4399      0.136        366        640:  26%|██▌       | 210/809 [04:11<12:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181       0.44     0.1358        358        640:  26%|██▌       | 212/809 [04:13<11:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.18     0.4401      0.136        247        640:  26%|██▋       | 213/809 [04:14<12:28,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.183     0.4391     0.1359        204        640:  27%|██▋       | 215/809 [04:17<12:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181     0.4398     0.1355        172        640:  27%|██▋       | 217/809 [04:19<11:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181     0.4397     0.1353        236        640:  27%|██▋       | 218/809 [04:20<11:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.18     0.4399     0.1351        224        640:  27%|██▋       | 220/809 [04:23<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.178     0.4411     0.1346        135        640:  27%|██▋       | 222/809 [04:25<11:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4418     0.1342        140        640:  28%|██▊       | 224/809 [04:27<11:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.173     0.4423     0.1339        151        640:  28%|██▊       | 225/809 [04:28<11:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174      0.442      0.134        239        640:  28%|██▊       | 227/809 [04:31<11:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.173     0.4418     0.1337        289        640:  28%|██▊       | 229/809 [04:33<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174     0.4415      0.134        353        640:  28%|██▊       | 230/809 [04:34<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.174      0.442     0.1341        541        640:  29%|██▊       | 232/809 [04:37<11:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.176     0.4413     0.1342        304        640:  29%|██▉       | 234/809 [04:39<11:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.178     0.4407     0.1343        217        640:  29%|██▉       | 235/809 [04:40<11:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.183     0.4395     0.1355        382        640:  29%|██▉       | 237/809 [04:43<11:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4384      0.136        309        640:  30%|██▉       | 239/809 [04:45<11:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4383     0.1359        296        640:  30%|██▉       | 240/809 [04:46<11:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4383      0.136        100        640:  30%|██▉       | 242/809 [04:49<11:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4384     0.1357        233        640:  30%|███       | 244/809 [04:51<11:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4381     0.1359        332        640:  30%|███       | 245/809 [04:52<11:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4379     0.1367        201        640:  31%|███       | 247/809 [04:55<11:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4387     0.1364        175        640:  31%|███       | 249/809 [04:57<11:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4381     0.1366        413        640:  31%|███       | 250/809 [04:58<11:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4383     0.1366         86        640:  31%|███       | 252/809 [05:01<10:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4381     0.1364        175        640:  31%|███▏      | 254/809 [05:03<10:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4386     0.1361        144        640:  32%|███▏      | 256/809 [05:05<10:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4384      0.136        318        640:  32%|███▏      | 257/809 [05:07<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185     0.4387     0.1359        199        640:  32%|███▏      | 259/809 [05:09<11:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185     0.4387      0.136        396        640:  32%|███▏      | 261/809 [05:11<10:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185      0.439     0.1358        183        640:  32%|███▏      | 262/809 [05:13<10:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.183     0.4395     0.1355        219        640:  33%|███▎      | 264/809 [05:15<10:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.182     0.4399     0.1353        291        640:  33%|███▎      | 266/809 [05:17<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.182       0.44     0.1352        222        640:  33%|███▎      | 267/809 [05:18<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181     0.4403     0.1349        124        640:  33%|███▎      | 269/809 [05:21<10:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.181       0.44     0.1347        199        640:  33%|███▎      | 271/809 [05:23<10:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.182     0.4397     0.1349        270        640:  34%|███▎      | 272/809 [05:24<10:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.184     0.4391     0.1359        143        640:  34%|███▍      | 274/809 [05:27<10:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185      0.439     0.1363        205        640:  34%|███▍      | 276/809 [05:29<10:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4385     0.1369        683        640:  34%|███▍      | 277/809 [05:31<11:02,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4389     0.1369        158        640:  34%|███▍      | 279/809 [05:33<10:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185     0.4389     0.1369        321        640:  35%|███▍      | 281/809 [05:35<10:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185     0.4392     0.1369        226        640:  35%|███▍      | 282/809 [05:37<10:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.185     0.4398     0.1368        187        640:  35%|███▌      | 284/809 [05:39<10:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.184       0.44     0.1366        292        640:  35%|███▌      | 286/809 [05:41<10:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.184       0.44     0.1367        317        640:  36%|███▌      | 288/809 [05:44<10:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4394      0.137        578        640:  36%|███▌      | 289/809 [05:45<10:58,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4389     0.1375        444        640:  36%|███▌      | 291/809 [05:47<10:36,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188      0.439     0.1373        259        640:  36%|███▌      | 292/809 [05:49<10:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.19     0.4383     0.1384        336        640:  36%|███▋      | 294/809 [05:51<10:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4385     0.1381        153        640:  37%|███▋      | 296/809 [05:54<10:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4385      0.138        290        640:  37%|███▋      | 297/809 [05:55<10:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4387     0.1379        365        640:  37%|███▋      | 299/809 [05:57<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4385     0.1378        256        640:  37%|███▋      | 301/809 [05:59<10:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4386     0.1377        252        640:  37%|███▋      | 302/809 [06:01<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4388     0.1375        269        640:  38%|███▊      | 304/809 [06:03<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.186     0.4394     0.1371         85        640:  38%|███▊      | 306/809 [06:05<09:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4386     0.1373        525        640:  38%|███▊      | 308/809 [06:08<09:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4384     0.1373        297        640:  38%|███▊      | 309/809 [06:09<09:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4382     0.1372        144        640:  38%|███▊      | 311/809 [06:11<09:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4373      0.138        263        640:  39%|███▊      | 313/809 [06:14<10:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4368     0.1384        286        640:  39%|███▉      | 315/809 [06:16<09:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4366     0.1384        272        640:  39%|███▉      | 316/809 [06:17<09:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4365     0.1383         49        640:  39%|███▉      | 318/809 [06:20<09:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4366     0.1383        234        640:  40%|███▉      | 320/809 [06:22<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4364     0.1382        237        640:  40%|███▉      | 321/809 [06:23<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4371     0.1379        116        640:  40%|███▉      | 323/809 [06:26<09:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4374     0.1378        280        640:  40%|████      | 325/809 [06:28<09:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192      0.438     0.1375         75        640:  40%|████      | 327/809 [06:30<09:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4379     0.1376        258        640:  41%|████      | 328/809 [06:31<09:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4386     0.1372        297        640:  41%|████      | 330/809 [06:34<09:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.189     0.4385      0.137        370        640:  41%|████      | 332/809 [06:36<09:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.188     0.4389     0.1369        178        640:  41%|████      | 333/809 [06:37<09:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4389     0.1368        240        640:  41%|████▏     | 335/809 [06:40<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.187     0.4388     0.1372        388        640:  42%|████▏     | 337/809 [06:42<09:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.19     0.4381     0.1375        421        640:  42%|████▏     | 338/809 [06:43<09:47,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.19     0.4378     0.1374        125        640:  42%|████▏     | 340/809 [06:46<09:40,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4376     0.1375         85        640:  42%|████▏     | 342/809 [06:48<09:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4373     0.1376        336        640:  42%|████▏     | 343/809 [06:50<09:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4364     0.1379        455        640:  43%|████▎     | 345/809 [06:52<09:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195      0.437     0.1378        198        640:  43%|████▎     | 347/809 [06:54<09:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195      0.437     0.1378        212        640:  43%|████▎     | 348/809 [06:56<09:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4375     0.1376        149        640:  43%|████▎     | 350/809 [06:58<08:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4374     0.1377        126        640:  44%|████▎     | 352/809 [07:00<08:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4375     0.1375        188        640:  44%|████▎     | 353/809 [07:01<09:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4378     0.1374        180        640:  44%|████▍     | 355/809 [07:04<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4376     0.1374        166        640:  44%|████▍     | 357/809 [07:06<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4375     0.1374        277        640:  44%|████▍     | 359/809 [07:08<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.192     0.4374     0.1374        289        640:  44%|████▍     | 360/809 [07:10<08:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4374     0.1372        225        640:  45%|████▍     | 362/809 [07:12<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4372     0.1374        360        640:  45%|████▍     | 364/809 [07:14<08:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4371     0.1375        414        640:  45%|████▌     | 365/809 [07:16<08:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4374     0.1377        212        640:  45%|████▌     | 367/809 [07:18<08:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4373     0.1377        391        640:  45%|████▌     | 368/809 [07:19<08:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4376     0.1375        241        640:  46%|████▌     | 370/809 [07:22<08:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4376     0.1373        120        640:  46%|████▌     | 372/809 [07:24<08:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.193     0.4376     0.1373        105        640:  46%|████▌     | 373/809 [07:25<08:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4373     0.1373        274        640:  46%|████▋     | 375/809 [07:28<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4373     0.1374        425        640:  47%|████▋     | 377/809 [07:30<08:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4377     0.1372        138        640:  47%|████▋     | 379/809 [07:32<08:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.194     0.4377     0.1372        272        640:  47%|████▋     | 380/809 [07:34<08:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4371     0.1374        285        640:  47%|████▋     | 382/809 [07:36<08:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4372     0.1372        199        640:  47%|████▋     | 384/809 [07:38<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4373     0.1372        229        640:  48%|████▊     | 385/809 [07:40<08:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4374     0.1371        191        640:  48%|████▊     | 387/809 [07:42<08:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.195     0.4374     0.1371        434        640:  48%|████▊     | 389/809 [07:44<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4371     0.1377        399        640:  48%|████▊     | 390/809 [07:46<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4372     0.1376        243        640:  48%|████▊     | 392/809 [07:48<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196      0.437     0.1375        252        640:  49%|████▊     | 394/809 [07:50<08:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4369     0.1374        230        640:  49%|████▉     | 395/809 [07:52<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1375        316        640:  49%|████▉     | 397/809 [07:54<08:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.199     0.4358     0.1378        405        640:  49%|████▉     | 399/809 [07:56<08:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.199     0.4357     0.1377        261        640:  49%|████▉     | 400/809 [07:58<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4352     0.1391        285        640:  50%|████▉     | 402/809 [08:00<08:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202      0.435     0.1391        187        640:  50%|████▉     | 404/809 [08:02<08:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4351      0.139        162        640:  50%|█████     | 405/809 [08:04<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346     0.1392        337        640:  50%|█████     | 407/809 [08:06<08:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346     0.1391        230        640:  51%|█████     | 409/809 [08:08<07:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346     0.1391        297        640:  51%|█████     | 410/809 [08:10<07:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1398        136        640:  51%|█████     | 412/809 [08:12<07:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4345     0.1397         90        640:  51%|█████     | 414/809 [08:14<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4344       0.14        188        640:  51%|█████▏    | 416/809 [08:17<07:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4341     0.1403        214        640:  52%|█████▏    | 417/809 [08:18<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4343     0.1402        255        640:  52%|█████▏    | 419/809 [08:20<07:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205      0.434     0.1402        243        640:  52%|█████▏    | 421/809 [08:23<07:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205      0.434     0.1401        220        640:  52%|█████▏    | 422/809 [08:24<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4335     0.1403        301        640:  52%|█████▏    | 424/809 [08:26<07:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4333     0.1402        467        640:  53%|█████▎    | 426/809 [08:29<07:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4335     0.1402        242        640:  53%|█████▎    | 427/809 [08:30<07:49,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4341     0.1399        362        640:  53%|█████▎    | 429/809 [08:32<07:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4342     0.1399        424        640:  53%|█████▎    | 431/809 [08:35<07:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4343     0.1399        318        640:  53%|█████▎    | 432/809 [08:36<07:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4345     0.1398        124        640:  54%|█████▎    | 434/809 [08:38<07:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4342     0.1399        391        640:  54%|█████▍    | 435/809 [08:39<07:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4343     0.1398        192        640:  54%|█████▍    | 437/809 [08:42<07:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4342     0.1397        256        640:  54%|█████▍    | 439/809 [08:44<07:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4341     0.1396        185        640:  55%|█████▍    | 441/809 [08:47<07:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1396        298        640:  55%|█████▍    | 442/809 [08:48<07:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4339     0.1395        292        640:  55%|█████▍    | 444/809 [08:50<07:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1402        119        640:  55%|█████▌    | 446/809 [08:53<07:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1401        238        640:  55%|█████▌    | 447/809 [08:54<07:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4339     0.1402        370        640:  56%|█████▌    | 449/809 [08:56<07:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4338     0.1402        182        640:  56%|█████▌    | 451/809 [08:59<07:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4339     0.1401        240        640:  56%|█████▌    | 452/809 [09:00<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4337     0.1402        222        640:  56%|█████▌    | 454/809 [09:02<06:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4336       0.14        258        640:  56%|█████▋    | 456/809 [09:05<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4335     0.1399        331        640:  56%|█████▋    | 457/809 [09:06<07:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4338     0.1397        215        640:  57%|█████▋    | 459/809 [09:08<06:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4339     0.1397        343        640:  57%|█████▋    | 461/809 [09:11<06:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1395        120        640:  57%|█████▋    | 462/809 [09:12<06:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4343     0.1393        114        640:  57%|█████▋    | 464/809 [09:14<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4348     0.1391        171        640:  58%|█████▊    | 466/809 [09:16<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346     0.1399        379        640:  58%|█████▊    | 467/809 [09:18<06:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346       0.14        271        640:  58%|█████▊    | 469/809 [09:20<06:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347     0.1398        350        640:  58%|█████▊    | 471/809 [09:22<06:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4351     0.1397        249        640:  58%|█████▊    | 473/809 [09:25<06:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201      0.435     0.1395        292        640:  59%|█████▊    | 475/809 [09:27<06:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4351     0.1395        124        640:  59%|█████▉    | 476/809 [09:28<06:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4351     0.1394        405        640:  59%|█████▉    | 478/809 [09:31<06:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4351     0.1394        274        640:  59%|█████▉    | 480/809 [09:33<06:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4353     0.1393        252        640:  59%|█████▉    | 481/809 [09:35<06:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4353     0.1392        287        640:  60%|█████▉    | 483/809 [09:37<06:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4349     0.1393        378        640:  60%|█████▉    | 485/809 [09:39<06:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4352     0.1393         90        640:  60%|██████    | 486/809 [09:40<06:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4349     0.1394        329        640:  60%|██████    | 488/809 [09:43<06:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347     0.1395        236        640:  61%|██████    | 490/809 [09:45<06:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4348     0.1395        234        640:  61%|██████    | 491/809 [09:47<06:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4348     0.1394        132        640:  61%|██████    | 493/809 [09:49<06:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4348     0.1392        150        640:  61%|██████    | 495/809 [09:51<06:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347     0.1392        227        640:  61%|██████▏   | 496/809 [09:52<06:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4349     0.1391        221        640:  62%|██████▏   | 498/809 [09:55<06:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4347     0.1392        143        640:  62%|██████▏   | 500/809 [09:57<06:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4345     0.1393        404        640:  62%|██████▏   | 501/809 [09:58<06:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4345     0.1393         68        640:  62%|██████▏   | 503/809 [10:01<05:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4343     0.1395        365        640:  62%|██████▏   | 505/809 [10:03<06:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4343     0.1394        134        640:  63%|██████▎   | 506/809 [10:04<06:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4344     0.1393        254        640:  63%|██████▎   | 508/809 [10:07<05:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4344     0.1395        317        640:  63%|██████▎   | 510/809 [10:09<05:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4344     0.1394        421        640:  63%|██████▎   | 511/809 [10:10<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4347     0.1393        212        640:  63%|██████▎   | 513/809 [10:13<05:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4348     0.1392        256        640:  64%|██████▎   | 515/809 [10:15<05:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4349     0.1391        297        640:  64%|██████▍   | 516/809 [10:16<05:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4346     0.1399        196        640:  64%|██████▍   | 518/809 [10:19<05:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4349     0.1397        247        640:  64%|██████▍   | 520/809 [10:21<05:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4347     0.1396        263        640:  65%|██████▍   | 522/809 [10:23<05:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4345     0.1396        413        640:  65%|██████▍   | 523/809 [10:25<05:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347     0.1398        277        640:  65%|██████▍   | 525/809 [10:27<05:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4349     0.1396        103        640:  65%|██████▌   | 527/809 [10:29<05:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347     0.1398        456        640:  65%|██████▌   | 528/809 [10:31<05:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4345     0.1397        255        640:  66%|██████▌   | 530/809 [10:33<05:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4341     0.1396        275        640:  66%|██████▌   | 532/809 [10:35<05:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4341     0.1396        238        640:  66%|██████▌   | 533/809 [10:36<05:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204      0.434     0.1396        178        640:  66%|██████▌   | 535/809 [10:39<05:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4339     0.1395        126        640:  66%|██████▋   | 537/809 [10:41<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4342     0.1393        127        640:  67%|██████▋   | 539/809 [10:43<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4342     0.1393        280        640:  67%|██████▋   | 540/809 [10:45<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4346     0.1393        178        640:  67%|██████▋   | 542/809 [10:47<05:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4347     0.1396        325        640:  67%|██████▋   | 544/809 [10:49<05:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4345     0.1403        392        640:  67%|██████▋   | 545/809 [10:51<05:28,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4343     0.1409        446        640:  68%|██████▊   | 547/809 [10:53<05:27,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4345     0.1408        214        640:  68%|██████▊   | 548/809 [10:54<05:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4349     0.1406         90        640:  68%|██████▊   | 550/809 [10:57<05:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4348     0.1406        178        640:  68%|██████▊   | 552/809 [10:59<05:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4348     0.1406        437        640:  68%|██████▊   | 553/809 [11:01<05:16,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4353     0.1405        127        640:  69%|██████▊   | 555/809 [11:03<05:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203      0.435     0.1406        437        640:  69%|██████▉   | 557/809 [11:05<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203      0.435     0.1405        301        640:  69%|██████▉   | 559/809 [11:08<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4347     0.1408        389        640:  69%|██████▉   | 560/809 [11:09<05:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4349     0.1408        147        640:  69%|██████▉   | 562/809 [11:11<04:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4349     0.1407        238        640:  70%|██████▉   | 563/809 [11:12<04:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4348     0.1406        273        640:  70%|██████▉   | 565/809 [11:15<04:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4349     0.1407        251        640:  70%|███████   | 567/809 [11:17<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4351     0.1405        279        640:  70%|███████   | 569/809 [11:20<04:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4351     0.1405        209        640:  70%|███████   | 570/809 [11:21<04:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4352     0.1404        223        640:  71%|███████   | 572/809 [11:23<04:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202      0.435     0.1405        259        640:  71%|███████   | 574/809 [11:26<04:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4347     0.1405        233        640:  71%|███████   | 575/809 [11:27<04:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4348     0.1405        134        640:  71%|███████▏  | 577/809 [11:29<04:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4346      0.141        209        640:  72%|███████▏  | 579/809 [11:32<04:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4344     0.1411        289        640:  72%|███████▏  | 580/809 [11:33<04:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4343      0.141        273        640:  72%|███████▏  | 582/809 [11:35<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4342     0.1411        183        640:  72%|███████▏  | 584/809 [11:38<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4339     0.1412        407        640:  72%|███████▏  | 585/809 [11:39<04:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4339     0.1412        391        640:  73%|███████▎  | 587/809 [11:41<04:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4341      0.141        205        640:  73%|███████▎  | 589/809 [11:44<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4342      0.141        298        640:  73%|███████▎  | 590/809 [11:45<04:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207      0.434     0.1413        408        640:  73%|███████▎  | 592/809 [11:47<04:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4339     0.1413        271        640:  73%|███████▎  | 593/809 [11:48<04:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4339     0.1411        264        640:  74%|███████▎  | 595/809 [11:51<04:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4336     0.1416        298        640:  74%|███████▍  | 597/809 [11:53<04:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4336     0.1414        256        640:  74%|███████▍  | 599/809 [11:56<04:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4335     0.1415        207        640:  74%|███████▍  | 600/809 [11:57<04:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4335     0.1415        272        640:  74%|███████▍  | 602/809 [11:59<04:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.209     0.4332     0.1419        446        640:  75%|███████▍  | 603/809 [12:01<04:13,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.209     0.4332     0.1419        225        640:  75%|███████▍  | 605/809 [12:03<04:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.209     0.4331     0.1418        315        640:  75%|███████▌  | 607/809 [12:05<04:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.21     0.4329     0.1423        610        640:  75%|███████▌  | 608/809 [12:07<04:14,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.209     0.4331     0.1421        194        640:  75%|███████▌  | 610/809 [12:09<04:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.21      0.433     0.1422        392        640:  76%|███████▌  | 612/809 [12:12<04:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.21      0.433     0.1422        229        640:  76%|███████▌  | 613/809 [12:13<03:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G       1.21      0.433     0.1421        519        640:  76%|███████▌  | 615/809 [12:15<03:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.209     0.4331      0.142        243        640:  76%|███████▋  | 617/809 [12:18<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4333     0.1419        105        640:  76%|███████▋  | 618/809 [12:19<03:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4334     0.1417        253        640:  77%|███████▋  | 620/809 [12:21<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4336     0.1416        283        640:  77%|███████▋  | 622/809 [12:23<03:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4335     0.1417        260        640:  77%|███████▋  | 624/809 [12:26<03:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4335     0.1417        606        640:  77%|███████▋  | 626/809 [12:28<03:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.208     0.4335     0.1417        266        640:  78%|███████▊  | 627/809 [12:29<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.207     0.4338     0.1415        254        640:  78%|███████▊  | 629/809 [12:32<03:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4341     0.1414        159        640:  78%|███████▊  | 631/809 [12:34<03:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4342     0.1413        192        640:  78%|███████▊  | 632/809 [12:35<03:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.206     0.4341     0.1414        427        640:  78%|███████▊  | 634/809 [12:38<03:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4344     0.1413        165        640:  79%|███████▊  | 636/809 [12:40<03:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4346     0.1413        356        640:  79%|███████▉  | 638/809 [12:42<03:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4346     0.1413        337        640:  79%|███████▉  | 639/809 [12:44<03:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4346     0.1414        534        640:  79%|███████▉  | 641/809 [12:46<03:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4344     0.1414        342        640:  79%|███████▉  | 642/809 [12:47<03:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4345     0.1414        255        640:  80%|███████▉  | 644/809 [12:50<03:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4344     0.1413        179        640:  80%|███████▉  | 646/809 [12:52<03:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4344     0.1412        315        640:  80%|████████  | 648/809 [12:54<03:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4345     0.1411        226        640:  80%|████████  | 649/809 [12:56<03:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4349     0.1411        300        640:  80%|████████  | 651/809 [12:58<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4352     0.1416         96        640:  81%|████████  | 653/809 [13:00<03:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4353     0.1416        255        640:  81%|████████  | 654/809 [13:01<03:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4355     0.1415        262        640:  81%|████████  | 656/809 [13:04<03:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4353     0.1415        267        640:  81%|████████▏ | 658/809 [13:06<03:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4353     0.1415        222        640:  81%|████████▏ | 659/809 [13:07<03:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4354     0.1414        241        640:  82%|████████▏ | 661/809 [13:10<02:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4354     0.1414        274        640:  82%|████████▏ | 663/809 [13:12<02:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4354     0.1414        329        640:  82%|████████▏ | 664/809 [13:13<02:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4355     0.1413        262        640:  82%|████████▏ | 666/809 [13:16<02:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4356     0.1413        265        640:  83%|████████▎ | 668/809 [13:18<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4356     0.1412        252        640:  83%|████████▎ | 669/809 [13:20<02:51,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4356     0.1412        248        640:  83%|████████▎ | 671/809 [13:22<02:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4354     0.1412        133        640:  83%|████████▎ | 673/809 [13:24<02:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4355     0.1411         99        640:  83%|████████▎ | 674/809 [13:25<02:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4351     0.1413        263        640:  84%|████████▎ | 676/809 [13:28<02:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4346     0.1414        221        640:  84%|████████▍ | 678/809 [13:30<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4347     0.1414        165        640:  84%|████████▍ | 679/809 [13:31<02:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.205     0.4347     0.1413        115        640:  84%|████████▍ | 681/809 [13:34<02:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4347     0.1413        243        640:  84%|████████▍ | 683/809 [13:36<02:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4348     0.1412        237        640:  85%|████████▍ | 685/809 [13:38<02:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4347     0.1413        199        640:  85%|████████▍ | 686/809 [13:40<02:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4348     0.1412        317        640:  85%|████████▌ | 688/809 [13:42<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4348     0.1411        150        640:  85%|████████▌ | 690/809 [13:44<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4349     0.1411        158        640:  85%|████████▌ | 691/809 [13:46<02:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203      0.435     0.1411        176        640:  86%|████████▌ | 693/809 [13:48<02:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4351      0.141        269        640:  86%|████████▌ | 695/809 [13:50<02:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4353     0.1409         97        640:  86%|████████▌ | 696/809 [13:51<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4349      0.141        376        640:  86%|████████▋ | 698/809 [13:54<02:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4348     0.1414        187        640:  87%|████████▋ | 700/809 [13:56<02:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4347     0.1413        231        640:  87%|████████▋ | 701/809 [13:57<02:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4347     0.1413        157        640:  87%|████████▋ | 703/809 [14:00<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4347     0.1412        154        640:  87%|████████▋ | 705/809 [14:02<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4346     0.1412        280        640:  87%|████████▋ | 707/809 [14:05<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.203     0.4346     0.1411        154        640:  88%|████████▊ | 708/809 [14:06<01:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.204     0.4343     0.1412        117        640:  88%|████████▊ | 710/809 [14:08<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346     0.1411        106        640:  88%|████████▊ | 712/809 [14:10<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4346      0.141        206        640:  88%|████████▊ | 713/809 [14:12<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4347      0.141        190        640:  88%|████████▊ | 715/809 [14:14<01:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4349     0.1408        170        640:  89%|████████▊ | 717/809 [14:16<01:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201      0.435     0.1408        253        640:  89%|████████▉ | 718/809 [14:17<01:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4349     0.1411        115        640:  89%|████████▉ | 720/809 [14:20<01:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4349     0.1412        404        640:  89%|████████▉ | 722/809 [14:22<01:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202      0.435     0.1412        281        640:  89%|████████▉ | 724/809 [14:25<01:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4348     0.1413        451        640:  90%|████████▉ | 725/809 [14:26<01:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.202     0.4349     0.1412        213        640:  90%|████████▉ | 727/809 [14:28<01:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4352     0.1411        168        640:  90%|█████████ | 729/809 [14:31<01:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4352      0.141        175        640:  90%|█████████ | 730/809 [14:32<01:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4353      0.141        284        640:  90%|█████████ | 732/809 [14:34<01:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.201     0.4354     0.1411         58        640:  91%|█████████ | 734/809 [14:36<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4355     0.1411        111        640:  91%|█████████ | 735/809 [14:38<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G        1.2     0.4356     0.1411        133        640:  91%|█████████ | 737/809 [14:40<01:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.199     0.4358     0.1411         88        640:  91%|█████████▏| 739/809 [14:42<01:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4361     0.1409         73        640:  92%|█████████▏| 741/809 [14:45<01:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1409        258        640:  92%|█████████▏| 742/809 [14:46<01:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4361     0.1408        249        640:  92%|█████████▏| 744/809 [14:48<01:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198      0.436     0.1409        271        640:  92%|█████████▏| 746/809 [14:51<01:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1408        141        640:  92%|█████████▏| 747/809 [14:52<01:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1408        176        640:  93%|█████████▎| 749/809 [14:54<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1407        348        640:  93%|█████████▎| 751/809 [14:56<01:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1408        358        640:  93%|█████████▎| 752/809 [14:58<01:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4363     0.1407        141        640:  93%|█████████▎| 754/809 [15:00<01:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4362     0.1407        449        640:  93%|█████████▎| 756/809 [15:03<01:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4363     0.1407        356        640:  94%|█████████▎| 757/809 [15:04<01:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4365     0.1407        228        640:  94%|█████████▍| 759/809 [15:06<00:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4363     0.1408        294        640:  94%|█████████▍| 761/809 [15:09<00:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4366     0.1407        125        640:  94%|█████████▍| 762/809 [15:10<00:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4367     0.1406        303        640:  94%|█████████▍| 764/809 [15:12<00:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4369     0.1405        208        640:  95%|█████████▍| 766/809 [15:14<00:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4369     0.1405        228        640:  95%|█████████▍| 767/809 [15:16<00:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4369     0.1405        289        640:  95%|█████████▌| 769/809 [15:18<00:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4369     0.1404         87        640:  95%|█████████▌| 771/809 [15:20<00:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4368     0.1404        157        640:  96%|█████████▌| 773/809 [15:23<00:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4367     0.1404        320        640:  96%|█████████▌| 774/809 [15:24<00:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4366     0.1404        262        640:  96%|█████████▌| 776/809 [15:26<00:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4366     0.1412        102        640:  96%|█████████▌| 778/809 [15:29<00:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4368     0.1412        205        640:  96%|█████████▋| 780/809 [15:31<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.198     0.4367     0.1412        348        640:  97%|█████████▋| 781/809 [15:32<00:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197      0.437     0.1411         80        640:  97%|█████████▋| 783/809 [15:35<00:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4371      0.141        336        640:  97%|█████████▋| 785/809 [15:37<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197      0.437      0.141        257        640:  97%|█████████▋| 786/809 [15:38<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4372     0.1409        193        640:  97%|█████████▋| 788/809 [15:41<00:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4371     0.1408        358        640:  98%|█████████▊| 790/809 [15:43<00:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4371      0.141        306        640:  98%|█████████▊| 792/809 [15:45<00:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4372      0.141        262        640:  98%|█████████▊| 793/809 [15:47<00:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4372     0.1409        169        640:  98%|█████████▊| 795/809 [15:49<00:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4372     0.1409        326        640:  99%|█████████▊| 797/809 [15:51<00:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.197     0.4371     0.1408        198        640:  99%|█████████▊| 798/809 [15:52<00:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4373     0.1408        163        640:  99%|█████████▉| 800/809 [15:55<00:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4374     0.1407        488        640:  99%|█████████▉| 802/809 [15:57<00:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4375     0.1406        209        640:  99%|█████████▉| 803/809 [15:58<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4377     0.1406        151        640: 100%|��████████▉| 805/809 [16:01<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4377     0.1406        297        640: 100%|█████████▉| 807/809 [16:03<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     41/300      13.9G      1.196     0.4376     0.1406        238        640: 100%|██████████| 809/809 [16:05<00:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  10%|█         | 7/69 [00:01<00:15,  4.03it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  22%|██▏       | 15/69 [00:03<00:13,  3.95it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  33%|███▎      | 23/69 [00:05<00:11,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  45%|████▍     | 31/69 [00:07<00:09,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  57%|█████▋    | 39/69 [00:09<00:07,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.42it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  78%|███████▊  | 54/69 [00:14<00:03,  3.78it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  88%|████████▊ | 61/69 [00:15<00:02,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.422      0.278      0.282       0.13
-     42/300      13.4G      1.229     0.4028     0.1157        229        640:   0%|          | 1/809 [00:01<15:30,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.203     0.4228      0.116        200        640:   0%|          | 3/809 [00:03<15:38,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.284     0.4011      0.134        273        640:   0%|          | 4/809 [00:04<15:47,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.252     0.4067     0.1221        167        640:   1%|          | 6/809 [00:06<15:31,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4323     0.1176        116        640:   1%|          | 8/809 [00:09<15:47,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.126     0.4499     0.1116        153        640:   1%|          | 9/809 [00:10<15:39,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.131      0.452     0.1148        279        640:   1%|▏         | 11/809 [00:12<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.129     0.4511     0.1137        288        640:   2%|▏         | 13/809 [00:15<15:28,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.14     0.4419     0.1159        175        640:   2%|▏         | 15/809 [00:17<15:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.135     0.4428     0.1159        199        640:   2%|▏         | 16/809 [00:18<15:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.156      0.437     0.1197        442        640:   2%|▏         | 18/809 [00:21<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4365     0.1188        407        640:   2%|▏         | 20/809 [00:23<15:50,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.149     0.4431     0.1176        177        640:   3%|▎         | 21/809 [00:24<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.154     0.4407     0.1193        163        640:   3%|▎         | 23/809 [00:27<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.139     0.4462     0.1179        268        640:   3%|▎         | 25/809 [00:29<15:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.134      0.448     0.1166        245        640:   3%|▎         | 26/809 [00:30<15:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.145     0.4462     0.1212        193        640:   3%|▎         | 28/809 [00:33<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.134     0.4501     0.1194        259        640:   4%|▎         | 30/809 [00:35<15:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.151     0.4463     0.1279        611        640:   4%|▍         | 31/809 [00:36<15:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.138     0.4505      0.126        203        640:   4%|▍         | 33/809 [00:38<15:19,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.144     0.4484     0.1277        381        640:   4%|▍         | 35/809 [00:41<15:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.14       0.45     0.1264        239        640:   4%|▍         | 36/809 [00:42<15:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.13     0.4538     0.1255        256        640:   5%|▍         | 38/809 [00:44<15:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.13     0.4527     0.1261        473        640:   5%|▍         | 40/809 [00:47<15:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.13     0.4554     0.1299        201        640:   5%|▌         | 42/809 [00:49<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.131     0.4557     0.1298        193        640:   5%|▌         | 43/809 [00:50<14:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.121     0.4579      0.128        216        640:   6%|▌         | 45/809 [00:53<14:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.131     0.4546     0.1293        555        640:   6%|▌         | 47/809 [00:55<15:05,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.132     0.4541     0.1289        314        640:   6%|▌         | 48/809 [00:56<15:18,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.13     0.4549     0.1304        139        640:   6%|▌         | 50/809 [00:59<15:01,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.137     0.4539     0.1337        305        640:   6%|▋         | 52/809 [01:01<14:50,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.132     0.4548     0.1327        203        640:   7%|▋         | 53/809 [01:02<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.139     0.4524      0.133        260        640:   7%|▋         | 55/809 [01:05<14:55,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.144     0.4508     0.1325        281        640:   7%|▋         | 57/809 [01:07<14:54,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.152     0.4479     0.1336        501        640:   7%|▋         | 58/809 [01:08<15:01,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4454     0.1345        348        640:   7%|▋         | 60/809 [01:11<14:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4467     0.1342        322        640:   8%|▊         | 62/809 [01:13<14:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.165     0.4454     0.1364        256        640:   8%|▊         | 63/809 [01:14<14:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.165     0.4456     0.1374        318        640:   8%|▊         | 65/809 [01:17<14:53,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4442     0.1421        248        640:   8%|▊         | 67/809 [01:19<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4438     0.1423        253        640:   8%|▊         | 68/809 [01:20<14:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4417     0.1454        242        640:   9%|▊         | 70/809 [01:22<14:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.183     0.4403     0.1469        392        640:   9%|▉         | 72/809 [01:25<14:54,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4396     0.1464        100        640:   9%|▉         | 74/809 [01:27<14:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4399      0.146        216        640:   9%|▉         | 75/809 [01:29<14:38,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4412     0.1453        263        640:  10%|▉         | 77/809 [01:31<14:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4393     0.1453        475        640:  10%|▉         | 79/809 [01:33<14:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4393     0.1451        228        640:  10%|▉         | 80/809 [01:34<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4387     0.1447        485        640:  10%|█         | 82/809 [01:37<14:29,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.184     0.4392     0.1445        133        640:  10%|█         | 84/809 [01:39<14:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4386      0.145        464        640:  11%|█         | 85/809 [01:40<14:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4385     0.1455        439        640:  11%|█         | 87/809 [01:43<14:24,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187     0.4386     0.1448        342        640:  11%|█         | 89/809 [01:45<14:25,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.19     0.4377     0.1455        227        640:  11%|█         | 90/809 [01:46<14:19,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.196     0.4353     0.1458        178        640:  11%|█▏        | 92/809 [01:49<14:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4346     0.1453        166        640:  12%|█▏        | 94/809 [01:51<13:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G        1.2     0.4331     0.1457        310        640:  12%|█▏        | 95/809 [01:52<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.208      0.431     0.1463        274        640:  12%|█▏        | 97/809 [01:55<14:21,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.204     0.4325     0.1462        176        640:  12%|█▏        | 99/809 [01:57<14:09,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.203     0.4324     0.1457        154        640:  12%|█▏        | 100/809 [01:58<14:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G        1.2     0.4331     0.1448        202        640:  13%|█▎        | 102/809 [02:01<13:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4348     0.1443        191        640:  13%|█▎        | 104/809 [02:03<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G        1.2     0.4341     0.1467        266        640:  13%|█▎        | 105/809 [02:04<14:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.199     0.4341     0.1463        317        640:  13%|█▎        | 107/809 [02:07<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.195     0.4355     0.1452        157        640:  13%|█▎        | 109/809 [02:09<13:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.198     0.4347     0.1467        362        640:  14%|█▎        | 111/809 [02:11<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4353     0.1463        240        640:  14%|█▍        | 112/809 [02:13<13:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4346     0.1458        302        640:  14%|█▍        | 114/809 [02:15<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4341     0.1452        257        640:  14%|█▍        | 116/809 [02:17<13:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.196     0.4342     0.1447        223        640:  14%|█▍        | 117/809 [02:18<13:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.196     0.4338     0.1445        469        640:  15%|█▍        | 119/809 [02:21<13:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.193     0.4345     0.1436        202        640:  15%|█▍        | 121/809 [02:23<13:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4366     0.1431        147        640:  15%|█▌        | 123/809 [02:25<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187      0.437     0.1428        183        640:  15%|█▌        | 124/809 [02:27<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187     0.4372     0.1423        160        640:  16%|█▌        | 126/809 [02:29<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4373      0.142        280        640:  16%|█▌        | 128/809 [02:32<14:16,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.189     0.4368     0.1417        256        640:  16%|█▌        | 130/809 [02:34<13:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.189     0.4369     0.1416        319        640:  16%|█▌        | 131/809 [02:35<13:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.191     0.4369     0.1434        295        640:  16%|█▋        | 133/809 [02:37<13:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.192     0.4367     0.1435        395        640:  17%|█▋        | 135/809 [02:40<13:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.191     0.4374     0.1432        261        640:  17%|█▋        | 136/809 [02:41<14:00,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4387     0.1425        216        640:  17%|█▋        | 138/809 [02:43<13:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4403     0.1417        180        640:  17%|█▋        | 140/809 [02:46<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4408     0.1415        167        640:  17%|█▋        | 141/809 [02:47<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4406     0.1409        238        640:  18%|█▊        | 143/809 [02:49<12:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4397     0.1408        204        640:  18%|█▊        | 145/809 [02:52<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.183     0.4388     0.1409        396        640:  18%|█▊        | 146/809 [02:53<13:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4387     0.1407        171        640:  18%|█▊        | 148/809 [02:55<13:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4388     0.1403        231        640:  19%|█▊        | 150/809 [02:58<12:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.183     0.4383     0.1401        273        640:  19%|█▊        | 151/809 [02:59<13:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.19     0.4362     0.1416        382        640:  19%|█▉        | 153/809 [03:01<13:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4366      0.142         47        640:  19%|█▉        | 155/809 [03:04<13:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.185     0.4374     0.1413        149        640:  19%|█▉        | 157/809 [03:06<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4373     0.1415        247        640:  20%|█▉        | 158/809 [03:07<12:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4366     0.1427        249        640:  20%|█▉        | 160/809 [03:10<13:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.185     0.4375     0.1421        195        640:  20%|██        | 162/809 [03:12<12:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4373     0.1421        375        640:  20%|██        | 163/809 [03:13<12:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4372     0.1426        250        640:  20%|██        | 165/809 [03:16<12:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187     0.4379     0.1422        242        640:  21%|██        | 167/809 [03:18<12:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4378     0.1421        267        640:  21%|██        | 168/809 [03:19<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188      0.437      0.142        210        640:  21%|██        | 170/809 [03:21<12:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.189     0.4364     0.1418        364        640:  21%|██▏       | 172/809 [03:24<12:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.19     0.4361      0.142        382        640:  22%|██▏       | 174/809 [03:26<12:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.19     0.4361     0.1422        296        640:  22%|██▏       | 175/809 [03:27<12:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.189     0.4368     0.1421        317        640:  22%|██▏       | 177/809 [03:30<12:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187     0.4375      0.142        171        640:  22%|██▏       | 179/809 [03:32<12:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.185     0.4387     0.1419        124        640:  22%|██▏       | 180/809 [03:33<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4383     0.1419        279        640:  22%|██▏       | 182/809 [03:36<12:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4384     0.1418        207        640:  23%|██▎       | 184/809 [03:38<12:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.186     0.4383     0.1417        285        640:  23%|██▎       | 185/809 [03:39<12:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4377      0.142        384        640:  23%|██▎       | 187/809 [03:42<12:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.188     0.4377     0.1417        138        640:  23%|██▎       | 189/809 [03:44<12:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.191     0.4367     0.1425        465        640:  23%|██▎       | 190/809 [03:45<12:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.195     0.4353     0.1424        342        640:  24%|██▎       | 192/809 [03:48<12:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.194     0.4355     0.1422        358        640:  24%|██▍       | 194/809 [03:50<12:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.197     0.4347     0.1436        445        640:  24%|██▍       | 195/809 [03:51<12:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.195     0.4351     0.1432        223        640:  24%|██▍       | 197/809 [03:54<12:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.193     0.4364     0.1427        152        640:  25%|██▍       | 199/809 [03:56<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.191      0.437     0.1423        194        640:  25%|██▍       | 200/809 [03:57<12:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.192     0.4365     0.1424        207        640:  25%|██▍       | 202/809 [04:00<11:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.193     0.4359     0.1423        146        640:  25%|██▌       | 204/809 [04:02<11:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.192     0.4361     0.1421        125        640:  25%|██▌       | 205/809 [04:04<11:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.191     0.4368     0.1418         78        640:  26%|██▌       | 207/809 [04:05<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.187     0.4379     0.1412        124        640:  26%|██▌       | 209/809 [04:08<11:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.184     0.4383     0.1409        136        640:  26%|██▌       | 211/809 [04:10<11:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.185     0.4382     0.1409        289        640:  26%|██▌       | 212/809 [04:11<11:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.184     0.4387     0.1407        268        640:  26%|██▋       | 214/809 [04:14<11:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4393     0.1404         85        640:  27%|██▋       | 216/809 [04:16<11:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4393     0.1402        206        640:  27%|██▋       | 217/809 [04:17<11:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4398     0.1396        192        640:  27%|██▋       | 219/809 [04:20<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4397     0.1394        202        640:  27%|██▋       | 221/809 [04:22<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4391     0.1394        383        640:  28%|██▊       | 223/809 [04:24<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4397     0.1392        111        640:  28%|██▊       | 224/809 [04:26<11:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4398     0.1387        258        640:  28%|██▊       | 226/809 [04:28<11:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4395     0.1384        139        640:  28%|██▊       | 228/809 [04:30<11:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4391     0.1384        308        640:  28%|██▊       | 229/809 [04:31<11:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4379     0.1386        470        640:  29%|██▊       | 231/809 [04:34<11:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.183     0.4375     0.1384        201        640:  29%|██▉       | 233/809 [04:36<11:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.183     0.4372     0.1382        273        640:  29%|██▉       | 234/809 [04:37<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181      0.438     0.1378        227        640:  29%|██▉       | 236/809 [04:40<11:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4382     0.1376        250        640:  29%|██▉       | 238/809 [04:42<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181      0.438     0.1375        360        640:  30%|██▉       | 239/809 [04:43<11:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4385     0.1373         75        640:  30%|██▉       | 241/809 [04:46<11:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4382     0.1376        489        640:  30%|███       | 243/809 [04:48<11:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4377     0.1375        339        640:  30%|███       | 245/809 [04:50<11:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4379     0.1375        169        640:  30%|███       | 246/809 [04:51<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4378     0.1374        368        640:  31%|███       | 248/809 [04:54<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4388      0.137        181        640:  31%|███       | 250/809 [04:56<10:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4388     0.1371        296        640:  31%|███       | 251/809 [04:57<10:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4397     0.1366        115        640:  31%|███▏      | 253/809 [05:00<10:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4396     0.1366        223        640:  32%|███▏      | 255/809 [05:02<10:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175       0.44     0.1364        238        640:  32%|███▏      | 257/809 [05:04<10:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4402     0.1364        173        640:  32%|███▏      | 258/809 [05:05<10:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4396     0.1364        335        640:  32%|███▏      | 260/809 [05:08<10:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179      0.439     0.1367        321        640:  32%|███▏      | 262/809 [05:10<10:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4384     0.1365        327        640:  33%|███▎      | 264/809 [05:13<10:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4389     0.1364        221        640:  33%|███▎      | 265/809 [05:14<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4397     0.1359        213        640:  33%|███▎      | 267/809 [05:16<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4396     0.1359        215        640:  33%|███▎      | 269/809 [05:18<10:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4397      0.136        244        640:  33%|███▎      | 271/809 [05:21<10:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4402     0.1358        159        640:  34%|███▎      | 272/809 [05:22<10:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176       0.44     0.1356        372        640:  34%|███▍      | 274/809 [05:24<10:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4405     0.1352        223        640:  34%|███▍      | 276/809 [05:27<10:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4409     0.1351        141        640:  34%|███▍      | 277/809 [05:28<10:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4411     0.1347        411        640:  34%|███▍      | 279/809 [05:30<10:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4412     0.1345        342        640:  35%|███▍      | 281/809 [05:33<10:50,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4412     0.1347        350        640:  35%|███▍      | 282/809 [05:34<10:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174      0.441     0.1354        218        640:  35%|███▌      | 284/809 [05:37<10:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4414     0.1354        132        640:  35%|███▌      | 286/809 [05:39<10:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4417     0.1353        202        640:  35%|███▌      | 287/809 [05:40<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4414     0.1354        370        640:  36%|███▌      | 289/809 [05:43<10:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4415     0.1352        265        640:  36%|███▌      | 290/809 [05:44<10:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4413     0.1351        209        640:  36%|███▌      | 292/809 [05:46<10:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4416      0.135        303        640:  36%|███▋      | 294/809 [05:49<10:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4415      0.135        202        640:  36%|███▋      | 295/809 [05:50<10:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4409     0.1351        323        640:  37%|███▋      | 297/809 [05:52<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4398     0.1352        533        640:  37%|███▋      | 299/809 [05:55<10:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18      0.439     0.1351        332        640:  37%|███▋      | 301/809 [05:57<10:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4392     0.1351        151        640:  37%|███▋      | 302/809 [05:58<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18      0.439     0.1349        266        640:  38%|███▊      | 304/809 [06:01<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4388     0.1353        499        640:  38%|███▊      | 306/809 [06:03<10:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4387     0.1359        284        640:  38%|███▊      | 307/809 [06:04<10:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4386     0.1358        176        640:  38%|███▊      | 309/809 [06:07<10:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4392     0.1356        240        640:  38%|███▊      | 311/809 [06:09<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4392     0.1358        402        640:  39%|███▊      | 312/809 [06:10<09:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4395     0.1357        295        640:  39%|███▉      | 314/809 [06:13<09:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4398     0.1356        283        640:  39%|███▉      | 316/809 [06:15<09:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4395     0.1362        288        640:  39%|███▉      | 317/809 [06:16<09:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4393     0.1367        322        640:  39%|███▉      | 319/809 [06:19<09:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4392     0.1373        265        640:  40%|███▉      | 321/809 [06:21<09:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4398     0.1372        163        640:  40%|███▉      | 322/809 [06:22<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4404      0.137        133        640:  40%|████      | 324/809 [06:24<09:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4407     0.1369        197        640:  40%|████      | 326/809 [06:27<09:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4403     0.1369        530        640:  40%|████      | 327/809 [06:28<09:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4399     0.1376        600        640:  41%|████      | 329/809 [06:31<09:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4403     0.1377        171        640:  41%|████      | 331/809 [06:33<09:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4404     0.1375        365        640:  41%|████      | 332/809 [06:34<09:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4408     0.1374        169        640:  41%|████▏     | 334/809 [06:37<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4409     0.1373        309        640:  42%|████▏     | 336/809 [06:39<09:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4408     0.1374        359        640:  42%|████▏     | 337/809 [06:40<09:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4408     0.1375        279        640:  42%|████▏     | 339/809 [06:43<09:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4409     0.1374        211        640:  42%|████▏     | 341/809 [06:45<09:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178      0.441     0.1374        226        640:  42%|████▏     | 342/809 [06:46<09:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4416     0.1371        195        640:  43%|████▎     | 344/809 [06:49<09:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174      0.442     0.1369        269        640:  43%|████▎     | 346/809 [06:51<09:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4419      0.137        317        640:  43%|████▎     | 347/809 [06:52<09:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4418      0.137        404        640:  43%|████▎     | 349/809 [06:55<09:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174      0.442     0.1371        492        640:  43%|████▎     | 351/809 [06:57<09:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4421      0.137        384        640:  44%|████▎     | 352/809 [06:58<09:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4417     0.1371        176        640:  44%|████▍     | 354/809 [07:01<09:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4409     0.1375        218        640:  44%|████▍     | 356/809 [07:03<09:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177      0.441     0.1375        139        640:  44%|████▍     | 357/809 [07:04<08:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4411     0.1373        249        640:  44%|████▍     | 359/809 [07:07<08:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4406     0.1382        456        640:  45%|████▍     | 361/809 [07:09<09:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4404     0.1381        312        640:  45%|███���▍     | 362/809 [07:10<09:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181     0.4404      0.138        120        640:  45%|████▍     | 364/809 [07:13<08:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4403     0.1379        169        640:  45%|████▌     | 366/809 [07:15<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181      0.441     0.1378         95        640:  45%|████▌     | 367/809 [07:16<08:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4407     0.1378        244        640:  46%|████▌     | 369/809 [07:18<08:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4406     0.1378        318        640:  46%|████▌     | 371/809 [07:21<08:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.182     0.4406     0.1376        210        640:  46%|████▌     | 373/809 [07:23<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.181      0.441     0.1375        108        640:  46%|████▌     | 374/809 [07:24<08:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18     0.4413     0.1373        231        640:  46%|████▋     | 376/809 [07:27<08:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4416     0.1371        214        640:  47%|████▋     | 378/809 [07:29<08:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4419     0.1369        279        640:  47%|████▋     | 379/809 [07:30<08:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4424     0.1367        103        640:  47%|████▋     | 381/809 [07:33<08:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4428      0.137         95        640:  47%|████▋     | 383/809 [07:35<08:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4429     0.1369        236        640:  47%|████▋     | 384/809 [07:36<08:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4434     0.1367        115        640:  48%|████▊     | 386/809 [07:39<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4434     0.1366        305        640:  48%|████▊     | 388/809 [07:41<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4436     0.1365        109        640:  48%|████▊     | 389/809 [07:42<08:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4437     0.1364        102        640:  48%|████▊     | 391/809 [07:45<08:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4437     0.1362        223        640:  49%|████▊     | 393/809 [07:47<08:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4435     0.1367        684        640:  49%|████▊     | 394/809 [07:48<08:35,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4432     0.1369        829        640:  49%|████▉     | 396/809 [07:51<08:46,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176      0.443     0.1376        433        640:  49%|████▉     | 398/809 [07:53<08:29,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4428     0.1376        364        640:  49%|████▉     | 399/809 [07:55<08:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4429     0.1375         57        640:  50%|████▉     | 401/809 [07:57<08:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4432     0.1373        248        640:  50%|████▉     | 403/809 [07:59<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4431     0.1374        480        640:  50%|████▉     | 404/809 [08:00<08:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4435     0.1371        199        640:  50%|█████     | 406/809 [08:03<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4437     0.1376        309        640:  50%|█████     | 408/809 [08:05<07:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4437     0.1376        258        640:  51%|█████     | 409/809 [08:06<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4434     0.1375        301        640:  51%|█████     | 411/809 [08:09<08:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4439     0.1373        186        640:  51%|█████     | 413/809 [08:11<07:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4442     0.1372        222        640:  51%|█████▏    | 415/809 [08:14<07:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4438     0.1373        482        640:  51%|█████▏    | 416/809 [08:15<07:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4433     0.1373        394        640:  52%|█████▏    | 418/809 [08:17<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4433     0.1372        278        640:  52%|█████▏    | 420/809 [08:20<07:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177      0.443     0.1375        384        640:  52%|█████▏    | 421/809 [08:21<07:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.178     0.4427     0.1375        324        640:  52%|█████▏    | 423/809 [08:23<07:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4423     0.1376        234        640:  53%|█████▎    | 425/809 [08:26<07:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18      0.442     0.1378        382        640:  53%|█████▎    | 426/809 [08:27<07:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18      0.442     0.1376        167        640:  53%|█████▎    | 428/809 [08:29<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4422     0.1374        175        640:  53%|█████▎    | 430/809 [08:32<07:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.18      0.442     0.1374        270        640:  53%|█████▎    | 431/809 [08:33<07:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.179     0.4423     0.1373        117        640:  54%|█████▎    | 433/809 [08:35<07:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4429      0.137         71        640:  54%|█████▍    | 435/809 [08:38<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.176     0.4431     0.1369        160        640:  54%|█████▍    | 437/809 [08:40<07:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.177     0.4431     0.1369        279        640:  54%|█████▍    | 438/809 [08:41<07:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4436     0.1367        173        640:  54%|█████▍    | 440/809 [08:43<07:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173      0.444     0.1364        225        640:  55%|█████▍    | 442/809 [08:46<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173      0.444     0.1364        212        640:  55%|█████▍    | 443/809 [08:47<07:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173      0.444     0.1364        384        640:  55%|█████▌    | 445/809 [08:49<07:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4441     0.1363        122        640:  55%|█████▌    | 447/809 [08:52<07:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4446     0.1362        101        640:  55%|█████▌    | 448/809 [08:53<07:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4446      0.137        238        640:  56%|█████▌    | 450/809 [08:55<07:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4446     0.1368        329        640:  56%|█████▌    | 452/809 [08:58<07:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4446     0.1368        238        640:  56%|█████▌    | 453/809 [08:59<07:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4448     0.1366        237        640:  56%|█████▌    | 455/809 [09:01<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4448     0.1366        353        640:  56%|█████▋    | 457/809 [09:04<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171      0.445     0.1364        313        640:  57%|█████▋    | 459/809 [09:06<06:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4452     0.1363        179        640:  57%|█████▋    | 460/809 [09:07<06:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171      0.445     0.1363        224        640:  57%|█████▋    | 462/809 [09:10<06:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4446     0.1363        206        640:  57%|█████▋    | 464/809 [09:12<06:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173     0.4443     0.1367        445        640:  57%|█████▋    | 465/809 [09:13<07:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4439     0.1367        330        640:  58%|█████▊    | 467/809 [09:16<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4439     0.1366        200        640:  58%|█████▊    | 469/809 [09:18<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4437     0.1366        369        640:  58%|█████▊    | 470/809 [09:19<06:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4437     0.1364        232        640:  58%|█████▊    | 472/809 [09:22<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4436     0.1364        527        640:  58%|█████▊    | 473/809 [09:23<06:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4434     0.1367        633        640:  59%|█████▊    | 475/809 [09:26<06:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.175     0.4436     0.1365        334        640:  59%|█████▉    | 477/809 [09:28<06:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.174     0.4437     0.1365        213        640:  59%|█████▉    | 478/809 [09:29<06:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.173      0.444     0.1363        234        640:  59%|█████▉    | 480/809 [09:31<06:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4442     0.1362        119        640:  60%|█████▉    | 482/809 [09:34<06:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4443      0.136        185        640:  60%|█████▉    | 484/809 [09:36<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4441     0.1361        273        640:  60%|█████▉    | 485/809 [09:37<06:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4444     0.1361        320        640:  60%|██████    | 487/809 [09:40<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4443     0.1361        356        640:  60%|██████    | 489/809 [09:42<06:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.172     0.4441     0.1361        341        640:  61%|██████    | 490/809 [09:43<06:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4444      0.136         67        640:  61%|██████    | 492/809 [09:46<06:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4447     0.1359        202        640:  61%|██████    | 494/809 [09:48<06:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4446     0.1358        191        640:  61%|██████    | 495/809 [09:49<06:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4445     0.1358        326        640:  61%|██████▏   | 497/809 [09:52<06:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4448     0.1357        126        640:  62%|██████▏   | 499/809 [09:54<06:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4452     0.1355        146        640:  62%|██████▏   | 500/809 [09:56<06:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4453     0.1355        241        640:  62%|██████▏   | 502/809 [09:57<06:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4453     0.1355        235        640:  62%|██████▏   | 504/809 [10:00<06:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4453     0.1355        243        640:  62%|██████▏   | 505/809 [10:01<06:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4453     0.1357        127        640:  63%|██████▎   | 507/809 [10:03<06:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4454     0.1356        186        640:  63%|██████▎   | 509/809 [10:06<05:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4456     0.1355        348        640:  63%|██████▎   | 511/809 [10:08<05:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4455     0.1357        469        640:  63%|██████▎   | 512/809 [10:09<06:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4455     0.1357        216        640:  64%|██████▎   | 514/809 [10:12<05:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4456     0.1357        242        640:  64%|██████▍   | 516/809 [10:14<05:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4456     0.1358        168        640:  64%|██████▍   | 517/809 [10:15<05:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4456     0.1358        356        640:  64%|██████▍   | 519/809 [10:18<05:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4454     0.1358        272        640:  64%|██████▍   | 521/809 [10:20<05:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4453     0.1359        274        640:  65%|██████▍   | 522/809 [10:21<05:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4452     0.1359        288        640:  65%|██████▍   | 524/809 [10:24<05:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.168     0.4457     0.1358        121        640:  65%|██████▌   | 526/809 [10:26<05:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4456     0.1358        276        640:  65%|██████▌   | 527/809 [10:27<05:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4455     0.1357        189        640:  65%|██████▌   | 529/809 [10:30<05:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4451     0.1363        619        640:  66%|██████▌   | 531/809 [10:32<05:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4449     0.1374        410        640:  66%|██████▌   | 532/809 [10:33<05:45,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4449     0.1374        134        640:  66%|██████▌   | 534/809 [10:36<05:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4451     0.1373        108        640:  66%|██████▋   | 536/809 [10:38<05:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4449     0.1373        362        640:  66%|██████▋   | 537/809 [10:39<05:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4453     0.1372        230        640:  67%|██████▋   | 539/809 [10:42<05:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4452     0.1371        248        640:  67%|██████▋   | 541/809 [10:44<05:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4451     0.1371        274        640:  67%|██████▋   | 542/809 [10:45<05:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4449     0.1371        258        640:  67%|██████▋   | 544/809 [10:48<05:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4446     0.1371        426        640:  67%|██████▋   | 546/809 [10:50<05:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4448     0.1371        180        640:  68%|██████▊   | 548/809 [10:52<05:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4447     0.1371        335        640:  68%|██████▊   | 549/809 [10:54<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.171     0.4447     0.1381        211        640:  68%|██████▊   | 551/809 [10:56<05:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4448     0.1381        214        640:  68%|██████▊   | 552/809 [10:57<05:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.17     0.4448      0.138        282        640:  68%|██████▊   | 554/809 [11:00<05:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169      0.445     0.1378        237        640:  69%|██████▊   | 556/809 [11:02<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4451     0.1377        174        640:  69%|██████▉   | 557/809 [11:03<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4452     0.1378        154        640:  69%|██████▉   | 559/809 [11:06<04:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.169     0.4454     0.1377        282        640:  69%|██████▉   | 561/809 [11:08<04:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.168     0.4455     0.1376        186        640:  70%|██████▉   | 563/809 [11:10<04:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.168     0.4457     0.1375        164        640:  70%|██████▉   | 564/809 [11:12<04:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.166     0.4461     0.1373        260        640:  70%|██████▉   | 566/809 [11:14<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.166     0.4461     0.1373        333        640:  70%|███████   | 568/809 [11:16<04:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.166     0.4461     0.1372        292        640:  70%|███████   | 569/809 [11:18<04:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.165     0.4463     0.1371         65        640:  71%|███████   | 571/809 [11:20<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4465     0.1369        220        640:  71%|███████   | 573/809 [11:22<04:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162      0.447     0.1368        138        640:  71%|███████   | 575/809 [11:25<04:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4469     0.1368        366        640:  71%|███████   | 576/809 [11:26<04:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4466     0.1369        493        640:  71%|███████▏  | 578/809 [11:28<04:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4465     0.1368        276        640:  72%|███████▏  | 580/809 [11:31<04:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4465     0.1368        193        640:  72%|███████▏  | 581/809 [11:32<04:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4464     0.1367        359        640:  72%|███████▏  | 583/809 [11:34<04:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4465     0.1366        132        640:  72%|███████▏  | 585/809 [11:37<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4463     0.1366        321        640:  73%|███████▎  | 587/809 [11:39<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.165     0.4462     0.1365        248        640:  73%|███████▎  | 588/809 [11:40<04:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4463     0.1364        257        640:  73%|███████▎  | 590/809 [11:42<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4463     0.1362        175        640:  73%|███████▎  | 592/809 [11:45<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4464     0.1362         64        640:  73%|███████▎  | 593/809 [11:46<04:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4465      0.136        318        640:  74%|███████▎  | 595/809 [11:48<04:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4466     0.1359        261        640:  74%|███████▍  | 597/809 [11:51<04:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4466     0.1359        334        640:  74%|███████▍  | 599/809 [11:53<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4466     0.1358        304        640:  74%|███████▍  | 600/809 [11:54<04:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4468     0.1357        323        640:  74%|███████▍  | 602/809 [11:57<04:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4468     0.1357        297        640:  75%|███████▍  | 604/809 [11:59<04:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161      0.447     0.1356         75        640:  75%|███████▍  | 605/809 [12:00<04:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4469     0.1357        347        640:  75%|███████▌  | 607/809 [12:02<03:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4471     0.1356        107        640:  75%|███████▌  | 609/809 [12:05<03:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4471     0.1355        277        640:  75%|███████▌  | 610/809 [12:06<03:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4471     0.1355        127        640:  76%|███████▌  | 612/809 [12:08<03:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4472     0.1353        288        640:  76%|███████▌  | 614/809 [12:11<03:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4472     0.1355        475        640:  76%|███████▌  | 615/809 [12:12<03:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4472     0.1354        324        640:  76%|███████▋  | 617/809 [12:14<03:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4469     0.1356        259        640:  77%|███████▋  | 619/809 [12:17<03:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4469     0.1355        225        640:  77%|███████▋  | 620/809 [12:18<03:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16      0.447     0.1355        125        640:  77%|███████▋  | 622/809 [12:20<03:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4469     0.1355        259        640:  77%|███████▋  | 624/809 [12:23<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4468     0.1355        479        640:  77%|███████▋  | 625/809 [12:24<03:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4467     0.1356        262        640:  78%|███████▊  | 627/809 [12:26<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4467     0.1355        245        640:  78%|███████▊  | 629/809 [12:29<03:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4468     0.1355        180        640:  78%|███████▊  | 630/809 [12:30<03:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4469     0.1355        346        640:  78%|███████▊  | 632/809 [12:33<03:39,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4469     0.1354        202        640:  78%|███████▊  | 634/809 [12:35<03:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4467     0.1356        598        640:  78%|███████▊  | 635/809 [12:36<03:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4466     0.1355        261        640:  79%|███████▊  | 637/809 [12:38<03:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161      0.447     0.1354        223        640:  79%|███████▉  | 639/809 [12:41<03:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4468     0.1355        507        640:  79%|███████▉  | 640/809 [12:42<03:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4469     0.1355        338        640:  79%|███████▉  | 642/809 [12:45<03:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4473     0.1354        137        640:  80%|███████▉  | 644/809 [12:47<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4475     0.1353        266        640:  80%|███████▉  | 646/809 [12:49<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4476     0.1352        239        640:  80%|███████▉  | 647/809 [12:50<03:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16      0.448     0.1353        166        640:  80%|████████  | 649/809 [12:53<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4481     0.1352        244        640:  80%|████████  | 651/809 [12:55<03:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4482     0.1351        197        640:  81%|████████  | 652/809 [12:56<03:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4486     0.1349        161        640:  81%|████████  | 654/809 [12:59<03:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4485     0.1349        324        640:  81%|████████  | 656/809 [13:01<03:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4486     0.1349        247        640:  81%|████████  | 657/809 [13:02<03:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.157     0.4487     0.1348        241        640:  81%|████████▏ | 659/809 [13:04<02:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4485     0.1349        214        640:  82%|████████▏ | 661/809 [13:07<02:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4485     0.1348        283        640:  82%|████████▏ | 662/809 [13:08<02:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4484     0.1348        182        640:  82%|████████▏ | 664/809 [13:11<02:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4482     0.1348        408        640:  82%|████████▏ | 666/809 [13:13<02:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4484     0.1348        112        640:  82%|████████▏ | 667/809 [13:14<02:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4484     0.1346        200        640:  83%|████████▎ | 669/809 [13:17<02:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4483     0.1345        309        640:  83%|████████▎ | 671/809 [13:19<02:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4482     0.1345         68        640:  83%|████████▎ | 672/809 [13:20<02:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.157     0.4485     0.1343        198        640:  83%|████████▎ | 674/809 [13:23<02:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4483     0.1346        390        640:  84%|████████▎ | 676/809 [13:25<02:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4483     0.1345        204        640:  84%|████████▍ | 678/809 [13:27<02:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.157     0.4486     0.1344        208        640:  84%|████████▍ | 679/809 [13:28<02:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.157     0.4485     0.1344        161        640:  84%|████████▍ | 681/809 [13:31<02:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4484     0.1344        283        640:  84%|████████▍ | 683/809 [13:33<02:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4484     0.1343        104        640:  85%|████████▍ | 684/809 [13:34<02:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.157     0.4486     0.1343        189        640:  85%|████████▍ | 686/809 [13:37<02:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4482     0.1343        101        640:  85%|████████▌ | 688/809 [13:39<02:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.158     0.4482     0.1343        292        640:  85%|████████▌ | 689/809 [13:40<02:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.159     0.4481     0.1343        288        640:  85%|████████▌ | 691/809 [13:43<02:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4478     0.1344        468        640:  86%|████████▌ | 693/809 [13:45<02:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4476     0.1345        368        640:  86%|████████▌ | 694/809 [13:46<02:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4473     0.1345        312        640:  86%|████████▌ | 696/809 [13:49<02:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4475     0.1344        195        640:  86%|████████▋ | 698/809 [13:51<02:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4474     0.1345        436        640:  86%|████████▋ | 699/809 [13:52<02:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4472     0.1346        411        640:  87%|████████▋ | 701/809 [13:55<02:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4472     0.1346        171        640:  87%|████████▋ | 703/809 [13:57<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162      0.447      0.135        349        640:  87%|████████▋ | 704/809 [13:58<02:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4467      0.135        277        640:  87%|████████▋ | 706/809 [14:01<02:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4466     0.1352         84        640:  88%|████████▊ | 708/809 [14:03<02:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4466     0.1352        208        640:  88%|████████▊ | 709/809 [14:04<01:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4465     0.1351        233        640:  88%|████████▊ | 711/809 [14:07<01:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4466     0.1351        245        640:  88%|████████▊ | 713/809 [14:09<01:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4467      0.135         58        640:  88%|████████▊ | 714/809 [14:10<01:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4465     0.1351        187        640:  89%|████████▊ | 716/809 [14:13<01:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4465     0.1351        149        640:  89%|████████▉ | 718/809 [14:15<01:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4462     0.1351         74        640:  89%|████████▉ | 720/809 [14:17<01:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4464      0.135        107        640:  89%|████████▉ | 722/809 [14:20<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4465      0.135        153        640:  89%|████████▉ | 723/809 [14:21<01:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4464      0.135        287        640:  90%|████████▉ | 725/809 [14:23<01:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4464      0.135        261        640:  90%|████████▉ | 727/809 [14:26<01:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4464     0.1349        232        640:  90%|████████▉ | 728/809 [14:27<01:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4464     0.1349        188        640:  90%|█████████ | 730/809 [14:29<01:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4462     0.1352        159        640:  90%|█████████ | 732/809 [14:32<01:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163      0.446     0.1354        361        640:  91%|█████████ | 734/809 [14:34<01:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4459     0.1354        348        640:  91%|█████████ | 735/809 [14:35<01:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4457     0.1355        190        640:  91%|█████████ | 737/809 [14:38<01:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4457     0.1355        202        640:  91%|█████████▏| 739/809 [14:40<01:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.164     0.4458     0.1356        230        640:  91%|█████████▏| 740/809 [14:41<01:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4459     0.1355        271        640:  92%|█████████▏| 742/809 [14:43<01:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4462     0.1355        140        640:  92%|█████████▏| 744/809 [14:46<01:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4463     0.1355        102        640:  92%|█████████▏| 745/809 [14:47<01:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4462     0.1356        263        640:  92%|█████████▏| 747/809 [14:50<01:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4461     0.1356        310        640:  93%|█████████▎| 749/809 [14:52<01:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162      0.446     0.1356        172        640:  93%|█████████▎| 750/809 [14:53<01:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4461     0.1355        286        640:  93%|█████████▎| 752/809 [14:55<01:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4461     0.1355        239        640:  93%|█████████▎| 754/809 [14:58<01:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4462     0.1355        284        640:  93%|█████████▎| 755/809 [14:59<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4462     0.1355        190        640:  94%|█████████▎| 757/809 [15:01<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4462     0.1355        144        640:  94%|█████████▍| 759/809 [15:04<00:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4462     0.1355        215        640:  94%|█████████▍| 761/809 [15:06<00:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4461     0.1355        233        640:  94%|█████████▍| 762/809 [15:07<00:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4462     0.1354        207        640:  94%|█████████▍| 764/809 [15:10<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4462     0.1354        218        640:  95%|█████████▍| 766/809 [15:12<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4461     0.1354        297        640:  95%|█████████▍| 767/809 [15:13<00:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4461     0.1354        237        640:  95%|█████████▌| 769/809 [15:16<00:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161      0.446     0.1353        217        640:  95%|█████████▌| 770/809 [15:17<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4457     0.1361         72        640:  95%|█████████▌| 772/809 [15:19<00:46,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4455     0.1361        217        640:  96%|█████████▌| 774/809 [15:22<00:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.163     0.4455     0.1361        230        640:  96%|█████████▌| 775/809 [15:23<00:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4456      0.136        245        640:  96%|█████████▌| 777/809 [15:25<00:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4456     0.1359        214        640:  96%|█████████▋| 779/809 [15:28<00:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4457     0.1359        224        640:  97%|█████████▋| 781/809 [15:30<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4457     0.1358        233        640:  97%|█████████▋| 782/809 [15:31<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4456     0.1358        191        640:  97%|█████████▋| 784/809 [15:34<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4456     0.1358        157        640:  97%|█████████▋| 786/809 [15:36<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4456     0.1357        204        640:  97%|█████████▋| 787/809 [15:37<00:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.162     0.4455     0.1358        332        640:  98%|█████████▊| 789/809 [15:39<00:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4459     0.1357         91        640:  98%|█████████▊| 791/809 [15:42<00:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4459     0.1357        246        640:  98%|█████████▊| 792/809 [15:43<00:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1356        328        640:  98%|█████████▊| 794/809 [15:45<00:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4459     0.1355        172        640:  98%|█████████▊| 796/809 [15:48<00:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1355        284        640:  99%|█████████▊| 797/809 [15:49<00:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1354         90        640:  99%|█████████▉| 799/809 [15:51<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1354        278        640:  99%|█████████▉| 801/809 [15:54<00:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1353        315        640:  99%|█████████▉| 803/809 [15:56<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G      1.161     0.4458     0.1353        277        640:  99%|█████████▉| 804/809 [15:57<00:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4459     0.1352        221        640: 100%|█████████▉| 806/809 [16:00<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     42/300      13.4G       1.16     0.4459     0.1352        279        640: 100%|█████████▉| 808/809 [16:02<00:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   4%|▍         | 3/69 [00:00<00:16,  4.02it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  16%|█▌        | 11/69 [00:02<00:14,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  28%|██▊       | 19/69 [00:04<00:12,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:10,  3.86it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  3.82it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  61%|██████    | 42/69 [00:10<00:06,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:12<00:05,  3.67it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:03,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:01,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.4G      1.309     0.4082     0.1896        288        640:   0%|          | 2/809 [00:02<16:06,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.4G       1.38      0.391     0.3238        328        640:   0%|          | 3/809 [00:03<16:04,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.4G      1.255     0.4326     0.2503        249        640:   1%|          | 5/809 [00:05<15:52,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.224     0.4432     0.2123        366        640:   1%|          | 7/809 [00:08<16:12,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.165     0.4539     0.1975        145        640:   1%|          | 8/809 [00:09<15:50,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.174     0.4501      0.201         99        640:   1%|          | 10/809 [00:12<16:05,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4361     0.2029        299        640:   1%|▏         | 12/809 [00:14<16:07,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.197     0.4376     0.1938        160        640:   2%|▏         | 13/809 [00:15<16:03,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.171      0.444     0.1809        208        640:   2%|▏         | 15/809 [00:18<15:58,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.166     0.4472     0.1774        254        640:   2%|▏         | 16/809 [00:19<15:50,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.198       0.44      0.189        143        640:   2%|▏         | 18/809 [00:21<16:05,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401      0.183         77        640:   2%|▏         | 20/809 [00:24<15:46,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4294     0.1842        745        640:   3%|▎         | 21/809 [00:25<16:40,  1.27s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.254     0.4192      0.188        516        640:   3%|▎         | 23/809 [00:28<16:41,  1.27s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.247     0.4201     0.1808        216        640:   3%|▎         | 25/809 [00:30<15:55,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.252     0.4179     0.1782        208        640:   3%|▎         | 26/809 [00:31<15:41,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.227     0.4238     0.1719        116        640:   3%|▎         | 28/809 [00:33<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4267     0.1678        277        640:   4%|▎         | 30/809 [00:36<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.222     0.4251     0.1664        180        640:   4%|▍         | 31/809 [00:37<15:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4269      0.162        265        640:   4%|▍         | 33/809 [00:39<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4298     0.1586        202        640:   4%|▍         | 35/809 [00:42<15:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.225     0.4258     0.1618        112        640:   5%|▍         | 37/809 [00:44<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4284     0.1604        115        640:   5%|▍         | 38/809 [00:45<15:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4279     0.1587        235        640:   5%|▍         | 40/809 [00:48<15:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4277     0.1564        179        640:   5%|▌         | 42/809 [00:50<15:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4307     0.1547        121        640:   5%|▌         | 43/809 [00:51<14:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206      0.432     0.1521        209        640:   6%|▌         | 45/809 [00:53<14:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.195     0.4352     0.1498        252        640:   6%|▌         | 47/809 [00:56<14:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.197     0.4346     0.1493        339        640:   6%|▌         | 48/809 [00:57<15:13,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.202     0.4343      0.148        361        640:   6%|▌         | 50/809 [00:59<15:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.198     0.4356     0.1469        161        640:   6%|▋         | 52/809 [01:02<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4377     0.1448        153        640:   7%|▋         | 54/809 [01:04<14:34,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4385     0.1442        166        640:   7%|▋         | 56/809 [01:06<14:42,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4368     0.1447        469        640:   7%|▋         | 57/809 [01:08<14:51,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4398     0.1426        297        640:   7%|▋         | 59/809 [01:10<14:48,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4422     0.1413        372        640:   8%|▊         | 61/809 [01:12<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.176      0.443     0.1413        329        640:   8%|▊         | 62/809 [01:13<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4423     0.1431        148        640:   8%|▊         | 64/809 [01:16<14:55,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.174     0.4431     0.1421        144        640:   8%|▊         | 66/809 [01:18<14:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.168     0.4447     0.1412        121        640:   8%|▊         | 68/809 [01:21<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.169     0.4444     0.1413        297        640:   9%|▊         | 69/809 [01:22<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.171     0.4436     0.1408        409        640:   9%|▉         | 71/809 [01:24<14:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.175     0.4419     0.1409        285        640:   9%|▉         | 73/809 [01:27<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.172     0.4432     0.1403        181        640:   9%|▉         | 74/809 [01:28<14:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.174     0.4421       0.14        311        640:   9%|▉         | 76/809 [01:30<14:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.173     0.4418     0.1395        267        640:  10%|▉         | 78/809 [01:32<14:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.173     0.4419     0.1399        324        640:  10%|▉         | 79/809 [01:34<14:35,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.177     0.4403     0.1401        359        640:  10%|█         | 81/809 [01:36<14:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4372     0.1402        320        640:  10%|█         | 83/809 [01:39<14:38,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.182     0.4376     0.1403         64        640:  10%|█         | 84/809 [01:40<14:25,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4339     0.1418        469        640:  11%|█         | 86/809 [01:42<14:32,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4345     0.1409        268        640:  11%|█         | 88/809 [01:45<14:22,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4363     0.1404         97        640:  11%|█         | 89/809 [01:46<14:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4356     0.1407        264        640:  11%|█         | 91/809 [01:48<14:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4365     0.1399        179        640:  11%|█▏        | 93/809 [01:50<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187      0.436     0.1402        234        640:  12%|█▏        | 95/809 [01:53<14:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4352     0.1401        268        640:  12%|█▏        | 96/809 [01:54<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.197     0.4337     0.1413         91        640:  12%|█▏        | 98/809 [01:56<14:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.196     0.4338     0.1411        191        640:  12%|█▏        | 99/809 [01:58<14:13,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206     0.4313     0.1446        308        640:  12%|█▏        | 101/809 [02:00<14:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206     0.4314     0.1442        229        640:  13%|█▎        | 103/809 [02:03<14:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206     0.4314      0.144        281        640:  13%|█▎        | 104/809 [02:04<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209      0.431     0.1445        328        640:  13%|█▎        | 106/809 [02:06<14:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206      0.432     0.1445        169        640:  13%|█▎        | 108/809 [02:09<14:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.208     0.4317     0.1453        284        640:  13%|█▎        | 109/809 [02:10<14:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.207     0.4317     0.1448        330        640:  14%|█▎        | 111/809 [02:12<14:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4302     0.1449        192        640:  14%|█▍        | 113/809 [02:15<14:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4294     0.1454        528        640:  14%|█▍        | 114/809 [02:16<14:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4282     0.1458        198        640:  14%|█▍        | 116/809 [02:18<13:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4286     0.1456        301        640:  15%|█▍        | 118/809 [02:20<13:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.219     0.4279      0.146        393        640:  15%|█▍        | 119/809 [02:22<13:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4286     0.1459        291        640:  15%|█▍        | 121/809 [02:24<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4315      0.145        108        640:  15%|█▌        | 123/809 [02:26<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4305     0.1461        323        640:  15%|█▌        | 125/809 [02:29<13:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4312     0.1458        203        640:  16%|█▌        | 126/809 [02:30<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4315     0.1466        587        640:  16%|█▌        | 128/809 [02:33<14:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4319     0.1466        171        640:  16%|█▌        | 129/809 [02:34<13:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4317     0.1462        129        640:  16%|█▌        | 131/809 [02:36<13:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4301     0.1468        333        640:  16%|█▋        | 133/809 [02:39<13:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216       0.43     0.1471        267        640:  17%|█▋        | 134/809 [02:40<13:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22     0.4286     0.1481        550        640:  17%|█▋        | 136/809 [02:42<13:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22     0.4285     0.1478        234        640:  17%|█▋        | 138/809 [02:45<13:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4304     0.1473        206        640:  17%|█▋        | 140/809 [02:47<13:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4315     0.1469        148        640:  17%|█▋        | 141/809 [02:48<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4313     0.1467        234        640:  18%|█▊        | 143/809 [02:50<13:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4308     0.1464        241        640:  18%|█▊        | 145/809 [02:53<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211      0.431     0.1461        249        640:  18%|█▊        | 146/809 [02:54<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4315     0.1459        416        640:  18%|█▊        | 148/809 [02:56<13:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209     0.4323     0.1452        143        640:  19%|█▊        | 150/809 [02:59<13:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209     0.4321     0.1454        370        640:  19%|█▊        | 151/809 [03:00<13:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4315     0.1455        457        640:  19%|█▉        | 153/809 [03:02<13:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4318     0.1454        205        640:  19%|█▉        | 155/809 [03:05<13:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4323     0.1452         62        640:  19%|█▉        | 156/809 [03:06<13:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4316     0.1467        325        640:  20%|█▉        | 158/809 [03:09<13:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4333     0.1461        132        640:  20%|█▉        | 160/809 [03:11<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212      0.433     0.1459        133        640:  20%|█▉        | 161/809 [03:12<12:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4328     0.1463        201        640:  20%|██        | 163/809 [03:14<12:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211      0.433     0.1457        268        640:  20%|██        | 165/809 [03:17<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4322      0.146        500        640:  21%|██        | 166/809 [03:18<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4319     0.1462        246        640:  21%|██        | 168/809 [03:20<12:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215      0.432     0.1459        264        640:  21%|██        | 170/809 [03:23<12:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4327     0.1459        245        640:  21%|██▏       | 172/809 [03:25<12:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4324     0.1457        278        640:  21%|██▏       | 173/809 [03:26<12:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4333     0.1452         88        640:  22%|██▏       | 175/809 [03:29<12:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.208     0.4339     0.1447        242        640:  22%|██▏       | 177/809 [03:31<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209     0.4336     0.1448        209        640:  22%|██▏       | 178/809 [03:32<12:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.208      0.434     0.1446        228        640:  22%|██▏       | 180/809 [03:35<12:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.208     0.4341     0.1444        376        640:  22%|██▏       | 182/809 [03:37<12:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.208      0.434     0.1443        250        640:  23%|██▎       | 183/809 [03:38<12:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.205     0.4346     0.1438        164        640:  23%|██▎       | 185/809 [03:40<12:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4349     0.1434        248        640:  23%|██▎       | 187/809 [03:43<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4347     0.1437        279        640:  23%|██▎       | 188/809 [03:44<12:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4349     0.1438        277        640:  23%|██▎       | 190/809 [03:46<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4355     0.1436        168        640:  24%|██▎       | 192/809 [03:49<12:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.204     0.4348     0.1443        531        640:  24%|██▍       | 193/809 [03:50<12:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4349     0.1438        184        640:  24%|██▍       | 195/809 [03:52<12:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.204     0.4347     0.1438        356        640:  24%|██▍       | 197/809 [03:55<12:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4346     0.1438        310        640:  24%|██▍       | 198/809 [03:56<12:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.205     0.4338      0.144        209        640:  25%|██▍       | 200/809 [03:58<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.207     0.4333     0.1437        245        640:  25%|██▍       | 202/809 [04:01<12:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4322     0.1438        231        640:  25%|██▌       | 204/809 [04:03<12:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4324     0.1436        210        640:  25%|██▌       | 205/809 [04:04<11:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4314     0.1447        316        640:  26%|██▌       | 207/809 [04:07<12:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4308     0.1453        568        640:  26%|██▌       | 208/809 [04:08<12:20,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4312     0.1448        379        640:  26%|██▌       | 210/809 [04:10<12:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4312     0.1443        464        640:  26%|██▌       | 212/809 [04:13<12:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4309     0.1441        221        640:  26%|██▋       | 214/809 [04:15<11:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4308     0.1442        282        640:  27%|██▋       | 215/809 [04:16<11:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4298     0.1445        213        640:  27%|██▋       | 217/809 [04:19<11:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4303     0.1439        102        640:  27%|██▋       | 219/809 [04:21<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4305     0.1437        238        640:  27%|██▋       | 221/809 [04:24<11:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4305     0.1436        255        640:  27%|██▋       | 222/809 [04:25<11:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4303     0.1434        209        640:  28%|██▊       | 224/809 [04:27<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4303     0.1452        438        640:  28%|██▊       | 226/809 [04:30<11:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4305     0.1449        159        640:  28%|██▊       | 227/809 [04:31<11:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4301     0.1449        181        640:  28%|██▊       | 229/809 [04:33<11:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212       0.43      0.145        199        640:  29%|██▊       | 231/809 [04:36<11:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4298      0.145        380        640:  29%|██▊       | 232/809 [04:37<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4302     0.1455        334        640:  29%|██▉       | 234/809 [04:39<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4306     0.1451        209        640:  29%|██▉       | 236/809 [04:41<11:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209     0.4311     0.1448        173        640:  29%|██▉       | 237/809 [04:43<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.209     0.4307     0.1449        229        640:  30%|██▉       | 239/809 [04:45<11:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4296     0.1466        358        640:  30%|██▉       | 241/809 [04:48<11:46,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4292     0.1468        692        640:  30%|██▉       | 242/809 [04:49<12:05,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4293     0.1465        286        640:  30%|███       | 244/809 [04:51<11:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4299     0.1461        357        640:  30%|███       | 246/809 [04:54<11:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4301      0.146        288        640:  31%|███       | 247/809 [04:55<11:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4285     0.1474        374        640:  31%|███       | 249/809 [04:57<11:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22     0.4277     0.1479        640        640:  31%|███       | 250/809 [04:59<11:40,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.222     0.4269     0.1483        441        640:  31%|███       | 252/809 [05:01<11:34,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.223     0.4265     0.1482        198        640:  31%|███▏      | 254/809 [05:04<11:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.225     0.4259     0.1482        227        640:  32%|███▏      | 255/809 [05:05<11:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.223     0.4265     0.1477        263        640:  32%|███▏      | 257/809 [05:07<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.224     0.4261     0.1476        264        640:  32%|███▏      | 259/809 [05:09<10:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.223     0.4265     0.1473        221        640:  32%|███▏      | 261/809 [05:12<10:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.224     0.4264     0.1475        341        640:  32%|███▏      | 262/809 [05:13<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.223     0.4266     0.1472        239        640:  33%|███▎      | 264/809 [05:15<10:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.222     0.4269     0.1469        206        640:  33%|███▎      | 266/809 [05:18<10:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4272     0.1467        182        640:  33%|███▎      | 267/809 [05:19<10:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4273     0.1465        154        640:  33%|███▎      | 269/809 [05:21<10:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4275     0.1465        256        640:  33%|███▎      | 271/809 [05:24<10:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4273     0.1465        254        640:  34%|███▎      | 272/809 [05:25<10:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4277     0.1464        179        640:  34%|███▍      | 274/809 [05:27<10:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22      0.428     0.1462        251        640:  34%|███▍      | 276/809 [05:29<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4285      0.146        114        640:  34%|███▍      | 278/809 [05:32<10:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.219     0.4285     0.1459        222        640:  34%|███▍      | 279/809 [05:33<10:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4288     0.1456        233        640:  35%|███▍      | 281/809 [05:35<10:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4289     0.1455        120        640:  35%|███▍      | 283/809 [05:38<10:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.219     0.4289     0.1454        265        640:  35%|███▌      | 284/809 [05:39<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22     0.4285     0.1457        193        640:  35%|███▌      | 286/809 [05:41<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.22     0.4285     0.1455        159        640:  36%|███▌      | 288/809 [05:44<10:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4282     0.1454        401        640:  36%|███▌      | 289/809 [05:45<10:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221      0.428     0.1451        179        640:  36%|███▌      | 291/809 [05:47<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4282      0.145        199        640:  36%|███▌      | 293/809 [05:50<10:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4283     0.1449        247        640:  36%|███▋      | 294/809 [05:51<10:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.221     0.4284     0.1447        203        640:  37%|███▋      | 296/809 [05:53<10:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.219      0.429     0.1446        151        640:  37%|███▋      | 298/809 [05:56<10:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.219     0.4295     0.1445        100        640:  37%|███▋      | 300/809 [05:58<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4299     0.1443        205        640:  37%|███▋      | 301/809 [05:59<10:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4299     0.1442        192        640:  37%|███▋      | 303/809 [06:02<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4299      0.144        335        640:  38%|███▊      | 305/809 [06:04<10:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218     0.4296     0.1439        391        640:  38%|███▊      | 306/809 [06:05<10:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216       0.43     0.1435        150        640:  38%|███▊      | 308/809 [06:08<10:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4299     0.1435        472        640:  38%|███▊      | 309/809 [06:09<10:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216       0.43     0.1436        396        640:  38%|███▊      | 311/809 [06:11<10:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4303     0.1435        364        640:  39%|███▊      | 313/809 [06:14<10:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4306     0.1433        111        640:  39%|███▉      | 314/809 [06:15<09:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4305     0.1432        446        640:  39%|███▉      | 316/809 [06:17<09:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4309     0.1429        158        640:  39%|███▉      | 318/809 [06:20<09:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4313     0.1427        249        640:  40%|███▉      | 320/809 [06:22<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4311     0.1427        328        640:  40%|███▉      | 321/809 [06:23<09:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4311     0.1427        306        640:  40%|███▉      | 323/809 [06:26<09:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214      0.431      0.143        266        640:  40%|████      | 325/809 [06:28<09:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4314     0.1428        150        640:  40%|████      | 326/809 [06:29<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4312     0.1426        302        640:  41%|████      | 328/809 [06:32<09:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.214     0.4311     0.1426        257        640:  41%|████      | 330/809 [06:34<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4308     0.1425        280        640:  41%|████      | 331/809 [06:35<09:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4301     0.1427        484        640:  41%|████      | 333/809 [06:38<09:46,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.218       0.43      0.143        391        640:  41%|████▏     | 335/809 [06:40<09:50,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217     0.4303     0.1427        278        640:  42%|████▏     | 336/809 [06:41<09:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4303     0.1427        192        640:  42%|████▏     | 338/809 [06:44<09:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.217       0.43     0.1427        736        640:  42%|████▏     | 339/809 [06:45<09:58,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4301     0.1425        308        640:  42%|████▏     | 341/809 [06:48<09:35,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.216     0.4303     0.1426        622        640:  42%|████▏     | 343/809 [06:50<09:50,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4303     0.1425        254        640:  43%|████▎     | 344/809 [06:51<09:37,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4301     0.1423        275        640:  43%|████▎     | 346/809 [06:54<09:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4303     0.1426        186        640:  43%|████▎     | 348/809 [06:56<09:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.215     0.4302     0.1426        677        640:  43%|████▎     | 349/809 [06:58<09:51,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.213     0.4305     0.1424        147        640:  43%|████▎     | 351/809 [07:00<09:27,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4308     0.1423        177        640:  44%|████▎     | 352/809 [07:01<09:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.212     0.4308     0.1422        393        640:  44%|████▍     | 354/809 [07:04<09:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4312     0.1421        357        640:  44%|████▍     | 356/809 [07:06<09:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.211     0.4313      0.142        331        640:  44%|████▍     | 357/809 [07:07<08:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.21     0.4317     0.1418        229        640:  44%|████▍     | 359/809 [07:09<09:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.207     0.4327     0.1415        126        640:  45%|████▍     | 361/809 [07:12<08:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206     0.4331     0.1413        239        640:  45%|████▍     | 363/809 [07:14<08:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.206     0.4333     0.1412        389        640:  45%|████▍     | 364/809 [07:15<08:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.205      0.434      0.141        318        640:  45%|████▌     | 366/809 [07:18<08:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4345     0.1407        136        640:  45%|████▌     | 368/809 [07:20<08:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.203     0.4345     0.1406        433        640:  46%|████▌     | 369/809 [07:21<08:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.202     0.4348     0.1405        109        640:  46%|████▌     | 371/809 [07:24<08:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4354     0.1402        100        640:  46%|████▌     | 373/809 [07:26<08:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G        1.2     0.4354     0.1401        214        640:  46%|████▌     | 374/809 [07:27<08:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201      0.435     0.1402        459        640:  46%|████▋     | 376/809 [07:30<08:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4349     0.1401        288        640:  47%|████▋     | 378/809 [07:32<08:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4348     0.1401        400        640:  47%|████▋     | 379/809 [07:33<08:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.202     0.4347     0.1408        229        640:  47%|████▋     | 381/809 [07:36<08:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.202     0.4345     0.1407        238        640:  47%|████▋     | 383/809 [07:38<08:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4346     0.1406        185        640:  47%|████▋     | 384/809 [07:39<08:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.201     0.4347     0.1404        353        640:  48%|████▊     | 386/809 [07:42<08:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G        1.2      0.435     0.1404        269        640:  48%|████▊     | 388/809 [07:44<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.199     0.4354       0.14        168        640:  48%|████▊     | 390/809 [07:46<08:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.198     0.4355     0.1398        347        640:  48%|████▊     | 392/809 [07:49<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.197     0.4357     0.1396        155        640:  49%|████▊     | 393/809 [07:50<08:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.197     0.4358     0.1397        424        640:  49%|████▉     | 395/809 [07:52<08:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.195     0.4363     0.1394        220        640:  49%|████▉     | 397/809 [07:55<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.196     0.4361     0.1395        342        640:  49%|████▉     | 398/809 [07:56<08:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.195     0.4363     0.1394        203        640:  49%|████▉     | 400/809 [07:58<08:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4367     0.1391        159        640:  50%|████▉     | 402/809 [08:01<08:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4372      0.139         78        640:  50%|████▉     | 403/809 [08:02<08:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4371     0.1395        155        640:  50%|█████     | 405/809 [08:04<07:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4371     0.1394        199        640:  50%|█████     | 407/809 [08:06<08:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4368       0.14        635        640:  50%|█████     | 408/809 [08:08<08:15,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192      0.437     0.1398        232        640:  51%|█████     | 410/809 [08:10<07:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4366     0.1399        677        640:  51%|█████     | 412/809 [08:13<08:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4364     0.1398        260        640:  51%|█████     | 413/809 [08:14<08:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4367     0.1397        120        640:  51%|█████▏    | 415/809 [08:16<07:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4361       0.14        248        640:  52%|█████▏    | 417/809 [08:19<07:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4365     0.1398         61        640:  52%|█████▏    | 418/809 [08:20<07:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4366     0.1395        261        640:  52%|█████▏    | 420/809 [08:22<07:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1394        196        640:  52%|█████▏    | 422/809 [08:25<07:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4371     0.1392        204        640:  52%|█████▏    | 424/809 [08:27<07:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4371     0.1391        354        640:  53%|█████▎    | 425/809 [08:28<07:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1392        272        640:  53%|█████▎    | 427/809 [08:31<07:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4366     0.1393        400        640:  53%|█████▎    | 429/809 [08:33<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4365     0.1392        350        640:  53%|█████▎    | 430/809 [08:34<07:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4366     0.1393        296        640:  53%|█████▎    | 432/809 [08:37<07:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4369      0.139        228        640:  54%|█████▎    | 434/809 [08:39<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4369      0.139        388        640:  54%|█████▍    | 435/809 [08:40<07:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191      0.437     0.1389        342        640:  54%|█████▍    | 437/809 [08:43<07:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4369     0.1391        348        640:  54%|█████▍    | 438/809 [08:44<07:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4363     0.1401        418        640:  54%|█████▍    | 440/809 [08:46<07:41,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.195     0.4362     0.1403        285        640:  55%|█████▍    | 442/809 [08:49<07:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.195     0.4361     0.1402        241        640:  55%|█████▍    | 443/809 [08:50<07:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4363       0.14        188        640:  55%|█████▌    | 445/809 [08:52<07:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4363       0.14        167        640:  55%|█████▌    | 447/809 [08:55<07:24,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4365     0.1399        149        640:  55%|█████▌    | 448/809 [08:56<07:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191      0.437     0.1398        232        640:  56%|█████▌    | 450/809 [08:58<07:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4372     0.1399        107        640:  56%|█████▌    | 452/809 [09:01<07:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4374     0.1398         98        640:  56%|█████▌    | 453/809 [09:02<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4376     0.1399        260        640:  56%|█████▌    | 455/809 [09:04<07:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4376     0.1397        231        640:  56%|█████▋    | 457/809 [09:07<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4374     0.1397        303        640:  57%|█████▋    | 458/809 [09:08<07:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4372     0.1397        135        640:  57%|█████▋    | 460/809 [09:10<06:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4371     0.1398        114        640:  57%|█████▋    | 462/809 [09:13<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4369     0.1398        203        640:  57%|█████▋    | 463/809 [09:14<06:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4367     0.1401        240        640:  57%|█████▋    | 465/809 [09:16<06:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4367     0.1399        330        640:  58%|█████▊    | 467/809 [09:19<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191      0.437     0.1397        243        640:  58%|█████▊    | 469/809 [09:21<06:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191      0.437     0.1397        304        640:  58%|█████▊    | 470/809 [09:22<06:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4369     0.1398        439        640:  58%|█████▊    | 472/809 [09:25<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191      0.437     0.1397         84        640:  58%|█████▊    | 473/809 [09:26<06:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1398        349        640:  59%|█████▊    | 475/809 [09:28<06:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4366     0.1402        177        640:  59%|█████▉    | 477/809 [09:31<06:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4364     0.1405        438        640:  59%|█████▉    | 478/809 [09:32<06:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1403         99        640:  59%|█████▉    | 480/809 [09:34<06:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1401        220        640:  60%|█████▉    | 482/809 [09:37<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368       0.14        200        640:  60%|█████▉    | 484/809 [09:39<06:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4369     0.1399        128        640:  60%|█████▉    | 485/809 [09:40<06:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4368     0.1399        192        640:  60%|██████    | 487/809 [09:43<06:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4364     0.1405        306        640:  60%|██████    | 489/809 [09:45<06:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4365     0.1404        213        640:  61%|██████    | 490/809 [09:46<06:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4368     0.1403        232        640:  61%|██████    | 492/809 [09:49<06:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4365     0.1405        566        640:  61%|██████    | 494/809 [09:51<06:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4364     0.1406        373        640:  61%|██████    | 495/809 [09:52<06:29,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4365     0.1407        173        640:  61%|██████▏   | 497/809 [09:55<06:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4366     0.1407        274        640:  62%|██████▏   | 499/809 [09:57<06:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.194     0.4366     0.1406        289        640:  62%|██████▏   | 500/809 [09:58<06:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4368     0.1404        214        640:  62%|██████▏   | 502/809 [10:01<06:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4373     0.1403        166        640:  62%|██████▏   | 504/809 [10:03<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4373     0.1403        275        640:  62%|██████▏   | 505/809 [10:04<05:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.192     0.4372     0.1403        419        640:  63%|██████▎   | 507/809 [10:07<05:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4368     0.1402        454        640:  63%|██████▎   | 509/809 [10:09<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.193     0.4369     0.1402        138        640:  63%|██████▎   | 510/809 [10:10<05:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4372     0.1399        260        640:  63%|██████▎   | 512/809 [10:12<05:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4374     0.1397        159        640:  64%|██████▎   | 514/809 [10:15<05:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4375     0.1396        369        640:  64%|██████▍   | 516/809 [10:17<05:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4375     0.1396        405        640:  64%|██████▍   | 517/809 [10:18<05:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4377     0.1395        164        640:  64%|██████▍   | 519/809 [10:21<05:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4375     0.1394        256        640:  64%|██████▍   | 521/809 [10:23<05:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4376     0.1394        271        640:  65%|██████▍   | 522/809 [10:24<05:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4376     0.1394        183        640:  65%|██████▍   | 524/809 [10:27<05:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4374     0.1392        126        640:  65%|██████▌   | 526/809 [10:29<05:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.191     0.4375     0.1393        281        640:  65%|██████▌   | 528/809 [10:31<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4376     0.1392        255        640:  66%|██████▌   | 530/809 [10:34<05:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.19     0.4376     0.1391        159        640:  66%|██████▌   | 531/809 [10:35<05:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.189     0.4378      0.139        295        640:  66%|██████▌   | 533/809 [10:37<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.188     0.4383     0.1388         69        640:  66%|██████▌   | 535/809 [10:40<05:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4386     0.1387        179        640:  66%|██████▋   | 536/809 [10:41<05:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4387     0.1386         78        640:  67%|██████▋   | 538/809 [10:43<05:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186      0.439     0.1387        380        640:  67%|██████▋   | 540/809 [10:45<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186      0.439     0.1386        304        640:  67%|██████▋   | 541/809 [10:47<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4392     0.1385        219        640:  67%|██████▋   | 543/809 [10:49<05:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4389     0.1386        338        640:  67%|██████▋   | 545/809 [10:51<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4389     0.1385        328        640:  68%|██████▊   | 547/809 [10:54<05:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4389     0.1386        154        640:  68%|██████▊   | 548/809 [10:55<05:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4391     0.1385        126        640:  68%|██████▊   | 550/809 [10:57<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4391     0.1387        190        640:  68%|██████▊   | 552/809 [11:00<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4389     0.1388        389        640:  68%|██████▊   | 553/809 [11:01<05:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.188     0.4386     0.1392        272        640:  69%|██████▊   | 555/809 [11:03<05:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4387      0.139        245        640:  69%|██████▉   | 557/809 [11:06<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4387      0.139        184        640:  69%|██████▉   | 558/809 [11:07<04:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4387     0.1389        206        640:  69%|██████▉   | 560/809 [11:09<04:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4392     0.1388        248        640:  69%|██████▉   | 562/809 [11:12<04:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4392     0.1391        340        640:  70%|██████▉   | 563/809 [11:13<04:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4392     0.1389        222        640:  70%|██████▉   | 565/809 [11:15<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4393     0.1391        150        640:  70%|███████   | 567/809 [11:18<04:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4395     0.1391        185        640:  70%|███████   | 568/809 [11:19<04:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4398     0.1391        119        640:  70%|███████   | 570/809 [11:21<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185       0.44      0.139        417        640:  71%|███████   | 572/809 [11:24<04:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185       0.44     0.1389        189        640:  71%|███████   | 573/809 [11:25<04:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4399     0.1392        233        640:  71%|███████   | 575/809 [11:27<04:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4397     0.1392        252        640:  71%|███████▏  | 577/809 [11:30<04:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4395     0.1393        268        640:  71%|███████▏  | 578/809 [11:31<04:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.188     0.4391     0.1396        468        640:  72%|███████▏  | 580/809 [11:33<04:45,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.188     0.4391     0.1395        176        640:  72%|███████▏  | 582/809 [11:36<04:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.188     0.4391     0.1395        220        640:  72%|███████▏  | 583/809 [11:37<04:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4393     0.1394        354        640:  72%|███████▏  | 585/809 [11:39<04:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4399     0.1393        141        640:  73%|███████▎  | 587/809 [11:42<04:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4399     0.1392        178        640:  73%|███████▎  | 588/809 [11:43<04:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4399     0.1392        361        640:  73%|███████▎  | 590/809 [11:45<04:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4402     0.1391        300        640:  73%|███████▎  | 592/809 [11:48<04:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1395        429        640:  73%|███████▎  | 593/809 [11:49<04:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1393        100        640:  74%|███████▎  | 595/809 [11:51<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4402     0.1392         64        640:  74%|███████▍  | 597/809 [11:54<04:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4403     0.1391        111        640:  74%|███████▍  | 598/809 [11:55<04:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4404      0.139        261        640:  74%|███████▍  | 600/809 [11:57<04:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4404      0.139        258        640:  74%|███████▍  | 602/809 [12:00<04:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4406     0.1388        109        640:  75%|███████▍  | 604/809 [12:02<03:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4405     0.1388        178        640:  75%|███████▍  | 605/809 [12:03<03:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.183     0.4408     0.1387        171        640:  75%|███████▌  | 607/809 [12:06<04:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4407     0.1388        325        640:  75%|███████▌  | 609/809 [12:08<03:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4406     0.1387        379        640:  75%|███████▌  | 610/809 [12:09<03:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.183     0.4408     0.1387        207        640:  76%|███████▌  | 612/809 [12:11<03:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4406     0.1392        410        640:  76%|███████▌  | 614/809 [12:14<03:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4405     0.1392        339        640:  76%|███████▌  | 615/809 [12:15<03:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4405      0.139        309        640:  76%|███████▋  | 617/809 [12:18<03:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4402     0.1393        319        640:  77%|███████▋  | 619/809 [12:20<03:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4405     0.1392        207        640:  77%|███████▋  | 620/809 [12:21<03:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4407     0.1391        129        640:  77%|███████▋  | 622/809 [12:24<03:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4406     0.1392        506        640:  77%|███████▋  | 624/809 [12:26<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4404     0.1392        289        640:  77%|███████▋  | 625/809 [12:27<03:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1396        433        640:  78%|███████▊  | 627/809 [12:30<03:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1397        265        640:  78%|███████▊  | 629/809 [12:32<03:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4402     0.1397        178        640:  78%|███████▊  | 630/809 [12:33<03:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4402       0.14        731        640:  78%|███████▊  | 632/809 [12:36<03:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4403     0.1399        161        640:  78%|███████▊  | 634/809 [12:38<03:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4406     0.1399        152        640:  78%|███████▊  | 635/809 [12:39<03:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4404       0.14        325        640:  79%|███████▊  | 637/809 [12:42<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4402       0.14        334        640:  79%|███████▉  | 638/809 [12:43<03:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4405     0.1398        181        640:  79%|███████▉  | 640/809 [12:45<03:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4407     0.1398        304        640:  79%|███████▉  | 642/809 [12:48<03:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4409     0.1397        237        640:  80%|███████▉  | 644/809 [12:50<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.183      0.441     0.1397        118        640:  80%|███████▉  | 645/809 [12:51<03:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184      0.441     0.1398        406        640:  80%|███████▉  | 647/809 [12:54<03:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4408       0.14        533        640:  80%|████████  | 648/809 [12:55<03:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4403     0.1403        473        640:  80%|████████  | 650/809 [12:58<03:19,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4402     0.1406        278        640:  81%|████████  | 652/809 [13:00<03:15,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4403     0.1405        268        640:  81%|████████  | 653/809 [13:01<03:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1408        271        640:  81%|████████  | 655/809 [13:04<03:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4397      0.141        664        640:  81%|████████  | 657/809 [13:06<03:08,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4397     0.1411        264        640:  81%|████████▏ | 658/809 [13:07<03:03,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4399      0.141        174        640:  82%|████████▏ | 660/809 [13:10<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1411        276        640:  82%|████████▏ | 662/809 [13:12<02:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401      0.141        225        640:  82%|████████▏ | 663/809 [13:13<02:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1409        212        640:  82%|████████▏ | 665/809 [13:16<02:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1409        256        640:  82%|████████▏ | 667/809 [13:18<02:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1409        273        640:  83%|████████▎ | 668/809 [13:19<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1408        210        640:  83%|████████▎ | 670/809 [13:21<02:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1408        392        640:  83%|████████▎ | 672/809 [13:24<02:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1407        333        640:  83%|████████▎ | 674/809 [13:26<02:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.187     0.4398     0.1408        334        640:  83%|████████▎ | 675/809 [13:27<02:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186       0.44     0.1407        210        640:  84%|████████▎ | 677/809 [13:30<02:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4403     0.1405        153        640:  84%|████████▍ | 679/809 [13:32<02:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185       0.44     0.1407        381        640:  84%|████████▍ | 681/809 [13:35<02:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.186     0.4401     0.1407        166        640:  84%|████████▍ | 682/809 [13:36<02:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.185     0.4401     0.1406        146        640:  85%|████████▍ | 684/809 [13:38<02:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4404     0.1405        201        640:  85%|████████▍ | 686/809 [13:40<02:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4405     0.1404         98        640:  85%|████████▍ | 687/809 [13:42<02:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.184     0.4407     0.1404        419        640:  85%|████████▌ | 689/809 [13:44<02:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.183     0.4409     0.1403         84        640:  85%|████████▌ | 691/809 [13:46<02:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.182     0.4411     0.1402        279        640:  86%|████████▌ | 693/809 [13:49<02:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.182     0.4411     0.1402        263        640:  86%|████████▌ | 694/809 [13:50<02:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.182      0.441     0.1402        142        640:  86%|████████▌ | 696/809 [13:52<02:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.182     0.4409     0.1401        370        640:  86%|████████▋ | 698/809 [13:55<02:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4411       0.14         85        640:  86%|████████▋ | 699/809 [13:56<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4412     0.1399        200        640:  87%|████████▋ | 701/809 [13:58<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4411     0.1399        203        640:  87%|████████▋ | 703/809 [14:00<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4409       0.14        143        640:  87%|████████▋ | 705/809 [14:03<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181      0.441       0.14        281        640:  87%|████████▋ | 706/809 [14:04<02:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4413     0.1399        160        640:  88%|████████▊ | 708/809 [14:06<01:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4413     0.1398        266        640:  88%|████████▊ | 710/809 [14:09<01:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4415     0.1397        189        640:  88%|████████▊ | 711/809 [14:10<01:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4413     0.1396        274        640:  88%|████████▊ | 713/809 [14:12<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4411     0.1397        130        640:  88%|████████▊ | 715/809 [14:15<01:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4411     0.1397        184        640:  89%|████████▊ | 716/809 [14:16<01:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4409       0.14        305        640:  89%|████████▉ | 718/809 [14:18<01:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4409     0.1399        236        640:  89%|████████▉ | 720/809 [14:21<01:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181     0.4411     0.1399         61        640:  89%|████████▉ | 721/809 [14:22<01:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.181      0.441     0.1398        163        640:  89%|████████▉ | 723/809 [14:24<01:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4411     0.1397        305        640:  90%|████████▉ | 725/809 [14:26<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4415     0.1395        139        640:  90%|████████▉ | 727/809 [14:29<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4416     0.1395        250        640:  90%|████████▉ | 728/809 [14:30<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4415     0.1399        188        640:  90%|█████████ | 730/809 [14:32<01:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4415     0.1399        364        640:  90%|█████████ | 732/809 [14:35<01:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4415     0.1398        220        640:  91%|█████████ | 733/809 [14:36<01:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4415     0.1401        153        640:  91%|█████████ | 735/809 [14:38<01:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4418     0.1399        110        640:  91%|█████████ | 737/809 [14:41<01:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1398        208        640:  91%|█████████ | 738/809 [14:42<01:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4418     0.1398        131        640:  91%|█████████▏| 740/809 [14:44<01:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1398        349        640:  92%|█████████▏| 742/809 [14:47<01:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1398        271        640:  92%|█████████▏| 743/809 [14:48<01:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4418     0.1397        231        640:  92%|█████████▏| 745/809 [14:50<01:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4418     0.1396        140        640:  92%|█████████▏| 747/809 [14:53<01:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4418     0.1397        399        640:  92%|█████████▏| 748/809 [14:54<01:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.177     0.4419     0.1396        262        640:  93%|█████████▎| 750/809 [14:56<01:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        464        640:  93%|█████████▎| 752/809 [14:59<01:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        295        640:  93%|█████████▎| 753/809 [15:00<01:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        276        640:  93%|█████████▎| 755/809 [15:02<01:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        230        640:  94%|█████████▎| 757/809 [15:05<01:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        295        640:  94%|█████████▎| 758/809 [15:06<01:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4417     0.1399        261        640:  94%|█████████▍| 760/809 [15:08<00:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4416     0.1399        262        640:  94%|█████████▍| 762/809 [15:11<00:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4416     0.1399        262        640:  94%|█████████▍| 763/809 [15:12<00:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4415     0.1398        210        640:  95%|█████████▍| 765/809 [15:14<00:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4412     0.1399        366        640:  95%|█████████▍| 767/809 [15:17<00:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4413     0.1399        151        640:  95%|█████████▍| 768/809 [15:18<00:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4414     0.1398        121        640:  95%|█████████▌| 770/809 [15:20<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4412     0.1399        233        640:  95%|█████████▌| 772/809 [15:23<00:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18      0.441     0.1399        197        640:  96%|█████████▌| 774/809 [15:25<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18      0.441     0.1399        230        640:  96%|█████████▌| 775/809 [15:26<00:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18      0.441     0.1398        356        640:  96%|█████████▌| 777/809 [15:29<00:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4408     0.1399        448        640:  96%|█████████▋| 779/809 [15:31<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4408     0.1398        380        640:  96%|█████████▋| 780/809 [15:32<00:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4408     0.1398        212        640:  97%|█████████▋| 782/809 [15:35<00:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4409     0.1397        334        640:  97%|█████████▋| 784/809 [15:37<00:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18      0.441     0.1397        340        640:  97%|█████████▋| 785/809 [15:38<00:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18      0.441     0.1397        276        640:  97%|█████████▋| 787/809 [15:41<00:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4409     0.1398        338        640:  98%|█████████▊| 789/809 [15:43<00:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G       1.18     0.4409     0.1397        300        640:  98%|█████████▊| 790/809 [15:44<00:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4409     0.1397        248        640:  98%|█████████▊| 792/809 [15:47<00:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4409     0.1396        347        640:  98%|█████████▊| 794/809 [15:49<00:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179      0.441     0.1396        210        640:  98%|█████████▊| 795/809 [15:50<00:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.179     0.4409     0.1396        294        640:  99%|█████████▊| 797/809 [15:53<00:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4411     0.1395        149        640:  99%|█████████▉| 799/809 [15:55<00:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.178     0.4412     0.1394         93        640:  99%|█████████▉| 800/809 [15:56<00:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.177     0.4413     0.1393         86        640:  99%|█████████▉| 802/809 [15:58<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.177     0.4414     0.1393        241        640:  99%|█████████▉| 804/809 [16:01<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.176     0.4417     0.1392        168        640: 100%|█████████▉| 806/809 [16:03<00:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     43/300      14.5G      1.176     0.4418     0.1391        272        640: 100%|█████████▉| 807/809 [16:04<00:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   1%|▏         | 1/69 [00:00<00:19,  3.50it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  36%|███▌      | 25/69 [00:06<00:11,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:09,  3.85it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:07,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:05,  3.53it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  3.75it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.6G     0.8912     0.4871    0.08786        207        640:   0%|          | 1/809 [00:01<15:29,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.6G       1.08     0.4507     0.1109        363        640:   0%|          | 3/809 [00:03<15:48,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.6G      1.194     0.4333     0.1709        323        640:   1%|          | 5/809 [00:05<16:03,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.249     0.4174     0.1778        381        640:   1%|          | 6/809 [00:07<16:27,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.297     0.4081     0.1748        275        640:   1%|          | 8/809 [00:09<16:13,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.304     0.4064     0.1674        355        640:   1%|          | 10/809 [00:11<15:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.311     0.4066     0.1731        361        640:   1%|▏         | 11/809 [00:13<15:51,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.284     0.4143     0.1646        245        640:   2%|▏         | 13/809 [00:15<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.258     0.4239     0.1583        216        640:   2%|▏         | 15/809 [00:17<15:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.255     0.4257     0.1587        342        640:   2%|▏         | 16/809 [00:19<15:51,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.245     0.4295      0.157        104        640:   2%|▏         | 18/809 [00:21<16:12,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.253     0.4265      0.158        413        640:   2%|▏         | 20/809 [00:24<15:58,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.248     0.4281      0.157        427        640:   3%|▎         | 21/809 [00:25<16:06,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.262     0.4231     0.1648        224        640:   3%|▎         | 23/809 [00:27<16:20,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.272     0.4188     0.1648        304        640:   3%|▎         | 25/809 [00:30<15:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.267     0.4168     0.1603        294        640:   3%|▎         | 27/809 [00:32<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.258     0.4209     0.1604         31        640:   3%|▎         | 28/809 [00:33<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.256     0.4205     0.1581        198        640:   4%|▎         | 30/809 [00:36<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.238     0.4265      0.154        180        640:   4%|▍         | 32/809 [00:38<15:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.245      0.425     0.1568        438        640:   4%|▍         | 33/809 [00:39<15:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.237      0.429     0.1537        242        640:   4%|▍         | 35/809 [00:42<15:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.246     0.4264     0.1566        761        640:   5%|▍         | 37/809 [00:44<16:22,  1.27s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.244     0.4273     0.1565        251        640:   5%|▍         | 38/809 [00:45<16:10,  1.26s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.255     0.4231     0.1562        252        640:   5%|▍         | 40/809 [00:48<15:40,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.244     0.4252     0.1534        208        640:   5%|▌         | 42/809 [00:50<15:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.254     0.4226      0.157        249        640:   5%|▌         | 43/809 [00:51<15:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.255     0.4222     0.1579        399        640:   6%|▌         | 45/809 [00:54<15:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.257     0.4218      0.157        461        640:   6%|▌         | 47/809 [00:56<15:30,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.249     0.4239     0.1558        171        640:   6%|▌         | 48/809 [00:58<15:52,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.253     0.4228      0.158        433        640:   6%|▌         | 50/809 [01:00<15:49,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.256     0.4218     0.1615        446        640:   6%|▋         | 51/809 [01:01<15:58,  1.26s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.252     0.4229     0.1599        376        640:   7%|▋         | 53/809 [01:04<15:19,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.253     0.4225     0.1614        542        640:   7%|▋         | 55/809 [01:06<15:25,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.255     0.4218     0.1615        320        640:   7%|▋         | 56/809 [01:07<15:24,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.25     0.4221     0.1601        383        640:   7%|▋         | 58/809 [01:10<15:00,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.242     0.4241     0.1576        497        640:   7%|▋         | 60/809 [01:12<15:04,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.243     0.4247     0.1578        265        640:   8%|▊         | 61/809 [01:13<14:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.244     0.4242      0.156        265        640:   8%|▊         | 63/809 [01:16<14:50,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.242     0.4249      0.154        144        640:   8%|▊         | 65/809 [01:18<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.244     0.4242     0.1535        289        640:   8%|▊         | 66/809 [01:19<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.249     0.4235     0.1544        149        640:   8%|▊         | 68/809 [01:22<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.253     0.4216     0.1544        304        640:   9%|▊         | 70/809 [01:24<14:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.254     0.4205     0.1538        281        640:   9%|▉         | 71/809 [01:25<14:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.247     0.4223     0.1524        221        640:   9%|▉         | 73/809 [01:28<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.24     0.4237     0.1508        184        640:   9%|▉         | 75/809 [01:30<14:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.231     0.4261     0.1492         93        640:  10%|▉         | 77/809 [01:32<14:18,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.231     0.4259     0.1491        253        640:  10%|▉         | 78/809 [01:34<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.233     0.4257     0.1488        222        640:  10%|▉         | 80/809 [01:36<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.234     0.4252     0.1485        190        640:  10%|█         | 82/809 [01:38<14:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.231     0.4257     0.1478        132        640:  10%|█         | 83/809 [01:39<14:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.234     0.4244     0.1482        292        640:  11%|█         | 85/809 [01:42<14:19,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.241     0.4223     0.1491        337        640:  11%|█         | 87/809 [01:44<14:29,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.24     0.4226     0.1484        174        640:  11%|█         | 88/809 [01:45<14:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.237     0.4232     0.1475        301        640:  11%|█         | 90/809 [01:48<14:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.24     0.4232     0.1513        471        640:  11%|█▏        | 92/809 [01:50<14:38,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.243     0.4221     0.1518        425        640:  11%|█▏        | 93/809 [01:52<14:37,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.245     0.4211     0.1519        284        640:  12%|█▏        | 95/809 [01:54<14:28,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.242      0.423      0.152        109        640:  12%|█▏        | 97/809 [01:56<14:13,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.242     0.4227     0.1516        276        640:  12%|█▏        | 98/809 [01:58<14:07,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.243     0.4223     0.1522        240        640:  12%|█▏        | 100/809 [02:00<14:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.242     0.4226     0.1519        163        640:  12%|█▏        | 101/809 [02:01<14:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.238      0.424     0.1511        191        640:  13%|█▎        | 103/809 [02:04<14:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.233      0.425     0.1504        270        640:  13%|█▎        | 105/809 [02:06<13:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.228     0.4262     0.1499        140        640:  13%|█▎        | 107/809 [02:08<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.225     0.4269     0.1493        241        640:  13%|█▎        | 108/809 [02:09<13:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.217     0.4293     0.1479        145        640:  14%|█▎        | 110/809 [02:12<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.214     0.4297     0.1472        236        640:  14%|█▍        | 112/809 [02:14<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.215     0.4296     0.1476        296        640:  14%|█▍        | 113/809 [02:15<13:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.213     0.4299      0.147        388        640:  14%|█▍        | 115/809 [02:18<14:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.209     0.4302     0.1463        395        640:  14%|█▍        | 117/809 [02:20<13:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.204     0.4314     0.1459        115        640:  15%|█▍        | 118/809 [02:21<13:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.201     0.4322      0.145        239        640:  15%|█▍        | 120/809 [02:24<13:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.199     0.4321     0.1445        320        640:  15%|█▌        | 122/809 [02:26<13:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4319     0.1445        343        640:  15%|█▌        | 123/809 [02:27<13:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4327     0.1437        193        640:  15%|█▌        | 125/809 [02:30<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.199     0.4324     0.1435        405        640:  16%|█▌        | 127/809 [02:32<13:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.195     0.4329      0.143        122        640:  16%|█▌        | 129/809 [02:35<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.194     0.4332     0.1426        258        640:  16%|█▌        | 130/809 [02:36<13:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.192     0.4341     0.1423         93        640:  16%|█▋        | 132/809 [02:38<13:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.192     0.4339      0.142        357        640:  17%|█▋        | 134/809 [02:40<13:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.194     0.4331     0.1428        433        640:  17%|█▋        | 135/809 [02:42<13:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.201     0.4312     0.1446        232        640:  17%|█▋        | 137/809 [02:44<14:17,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4317     0.1443        179        640:  17%|█▋        | 138/809 [02:46<13:54,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.208       0.43      0.149        473        640:  17%|█▋        | 140/809 [02:48<14:08,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.204     0.4313     0.1485        127        640:  17%|█▋        | 141/809 [02:49<13:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4325     0.1479        222        640:  18%|█▊        | 143/809 [02:52<13:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.201     0.4327     0.1481        351        640:  18%|█▊        | 145/809 [02:54<13:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.208     0.4313     0.1558        682        640:  18%|█▊        | 146/809 [02:56<14:07,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.213     0.4302     0.1569        554        640:  18%|█▊        | 148/809 [02:58<13:46,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.212     0.4302     0.1562        132        640:  19%|█▊        | 150/809 [03:01<13:35,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.211     0.4303     0.1558        270        640:  19%|█▊        | 151/809 [03:02<13:30,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.21     0.4305     0.1551        243        640:  19%|█▉        | 153/809 [03:04<13:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.208     0.4305     0.1543        302        640:  19%|█▉        | 155/809 [03:07<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.208     0.4304      0.154        239        640:  19%|█▉        | 156/809 [03:08<13:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.209     0.4307     0.1552        517        640:  20%|█▉        | 158/809 [03:10<13:30,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.208     0.4309     0.1548        122        640:  20%|█▉        | 159/809 [03:12<13:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.212     0.4296     0.1547        361        640:  20%|█▉        | 161/809 [03:14<13:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.212     0.4294     0.1543        157        640:  20%|██        | 163/809 [03:16<12:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.207      0.431     0.1535        127        640:  20%|██        | 165/809 [03:19<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.206     0.4311     0.1531        212        640:  21%|██        | 166/809 [03:20<12:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.201     0.4325     0.1522        258        640:  21%|██        | 168/809 [03:22<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.199     0.4328     0.1519        336        640:  21%|██        | 170/809 [03:25<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4336     0.1516        113        640:  21%|██        | 171/809 [03:26<12:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4326     0.1518        246        640:  21%|██▏       | 173/809 [03:28<12:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.196     0.4339      0.151        171        640:  22%|██▏       | 175/809 [03:30<12:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.195     0.4341     0.1507        245        640:  22%|██▏       | 176/809 [03:32<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197     0.4335     0.1508        347        640:  22%|██▏       | 178/809 [03:34<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4331     0.1512        242        640:  22%|██▏       | 180/809 [03:36<12:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4334     0.1511        192        640:  22%|██▏       | 181/809 [03:38<12:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.199     0.4334     0.1515        343        640:  23%|██▎       | 183/809 [03:40<12:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2     0.4332     0.1528        509        640:  23%|██▎       | 185/809 [03:43<12:53,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G        1.2      0.433     0.1525        323        640:  23%|██▎       | 186/809 [03:44<12:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197     0.4337      0.152        303        640:  23%|██▎       | 188/809 [03:46<12:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197     0.4335     0.1519        305        640:  23%|██▎       | 190/809 [03:49<12:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.196      0.434     0.1518        163        640:  24%|██▎       | 191/809 [03:50<12:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.196     0.4344      0.152        328        640:  24%|██▍       | 193/809 [03:52<12:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197      0.434     0.1544        260        640:  24%|██▍       | 195/809 [03:55<12:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4336     0.1541        356        640:  24%|██▍       | 196/809 [03:56<12:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4336      0.154        311        640:  24%|██▍       | 198/809 [03:58<12:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.199     0.4332     0.1536        359        640:  25%|██▍       | 200/809 [04:01<12:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.198     0.4334     0.1533        257        640:  25%|██▍       | 201/809 [04:02<12:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197     0.4338     0.1528        216        640:  25%|██▌       | 203/809 [04:04<12:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.197     0.4336     0.1523        190        640:  25%|██▌       | 205/809 [04:07<12:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.196     0.4343     0.1521        132        640:  25%|██▌       | 206/809 [04:08<12:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.194     0.4345     0.1517        223        640:  26%|██▌       | 208/809 [04:10<12:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.194     0.4342     0.1515        380        640:  26%|██▌       | 210/809 [04:13<12:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.195      0.434     0.1513        276        640:  26%|██▌       | 211/809 [04:14<12:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.195     0.4339     0.1509        312        640:  26%|██▋       | 213/809 [04:16<11:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.192      0.435     0.1505         83        640:  27%|██▋       | 215/809 [04:19<11:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.189     0.4362       0.15        138        640:  27%|██▋       | 217/809 [04:21<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.186      0.437     0.1494        108        640:  27%|██▋       | 219/809 [04:23<11:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184     0.4374     0.1491         74        640:  27%|██▋       | 220/809 [04:24<11:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.185     0.4374     0.1494        361        640:  27%|██▋       | 222/809 [04:27<11:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.183      0.438      0.149        176        640:  28%|██▊       | 224/809 [04:29<11:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184      0.438     0.1493        154        640:  28%|██▊       | 225/809 [04:31<11:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184      0.438     0.1491        125        640:  28%|██▊       | 227/809 [04:33<11:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184     0.4381     0.1489        287        640:  28%|██▊       | 229/809 [04:35<11:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184     0.4381     0.1487        310        640:  28%|██▊       | 230/809 [04:36<11:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.183     0.4387     0.1484        152        640:  29%|██▊       | 232/809 [04:39<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.184     0.4387     0.1484        232        640:  29%|██▉       | 234/809 [04:41<11:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.182     0.4393     0.1483        105        640:  29%|██▉       | 235/809 [04:42<11:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.182     0.4394     0.1489        343        640:  29%|██▉       | 237/809 [04:45<11:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181     0.4399     0.1485        199        640:  30%|██▉       | 239/809 [04:47<11:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.182     0.4397     0.1486        228        640:  30%|██▉       | 240/809 [04:48<11:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181     0.4401     0.1485        213        640:  30%|██▉       | 242/809 [04:51<11:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181     0.4402     0.1486        317        640:  30%|███       | 244/809 [04:53<11:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.18     0.4403     0.1483        316        640:  30%|███       | 245/809 [04:54<11:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.179     0.4408     0.1482        193        640:  31%|███       | 247/809 [04:57<11:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.18     0.4404     0.1481        297        640:  31%|███       | 249/809 [04:59<11:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181       0.44     0.1483        402        640:  31%|███       | 250/809 [05:00<11:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181       0.44      0.148        100        640:  31%|███       | 252/809 [05:03<11:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.181       0.44     0.1477        257        640:  31%|███▏      | 254/809 [05:05<11:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.18     0.4404     0.1473        188        640:  32%|███▏      | 256/809 [05:08<10:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.179     0.4405     0.1471        321        640:  32%|███▏      | 257/809 [05:09<10:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.177     0.4409     0.1468        320        640:  32%|███▏      | 259/809 [05:11<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.18     0.4402     0.1473        261        640:  32%|███▏      | 261/809 [05:14<10:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.178     0.4405     0.1471        123        640:  32%|███▏      | 262/809 [05:15<10:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.178     0.4408     0.1468        165        640:  33%|███▎      | 264/809 [05:17<10:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4415     0.1464        211        640:  33%|███▎      | 266/809 [05:19<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4412     0.1462        191        640:  33%|███▎      | 267/809 [05:21<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4415     0.1458        168        640:  33%|███▎      | 269/809 [05:23<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4419     0.1455        214        640:  33%|███▎      | 271/809 [05:25<10:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4419     0.1454        439        640:  34%|███▎      | 272/809 [05:27<10:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4427      0.145        212        640:  34%|███▍      | 274/809 [05:29<10:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4431     0.1448        198        640:  34%|███▍      | 276/809 [05:31<10:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4434     0.1446        173        640:  34%|███▍      | 277/809 [05:32<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4434     0.1444        215        640:  34%|███▍      | 279/809 [05:35<10:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4433     0.1441        228        640:  35%|███▍      | 281/809 [05:37<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4427     0.1445        185        640:  35%|███▍      | 283/809 [05:40<10:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4425     0.1444        277        640:  35%|███▌      | 284/809 [05:41<10:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4421     0.1449        521        640:  35%|███▌      | 286/809 [05:43<11:01,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4423     0.1448        213        640:  35%|███▌      | 287/809 [05:45<10:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1445         88        640:  36%|███▌      | 289/809 [05:47<10:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4423     0.1443        349        640:  36%|███▌      | 291/809 [05:49<10:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174      0.442     0.1443        428        640:  36%|███▌      | 293/809 [05:52<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4415     0.1452        332        640:  36%|███▋      | 294/809 [05:53<10:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4415     0.1451        299        640:  37%|███▋      | 296/809 [05:55<10:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4418      0.145        216        640:  37%|███▋      | 297/809 [05:57<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4417     0.1448        119        640:  37%|███▋      | 299/809 [05:59<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4415     0.1446        173        640:  37%|███▋      | 301/809 [06:01<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4417     0.1444         63        640:  37%|███▋      | 303/809 [06:04<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4416     0.1443        143        640:  38%|███▊      | 304/809 [06:05<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1442        265        640:  38%|███▊      | 306/809 [06:07<09:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173      0.442     0.1441        119        640:  38%|███▊      | 308/809 [06:09<09:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4418     0.1441        351        640:  38%|███▊      | 309/809 [06:11<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4419     0.1439        226        640:  38%|███▊      | 311/809 [06:13<09:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4417     0.1441        475        640:  39%|███▊      | 313/809 [06:16<10:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4417     0.1441        325        640:  39%|███▉      | 314/809 [06:17<10:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4413     0.1446        507        640:  39%|███▉      | 316/809 [06:19<09:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4416     0.1444        313        640:  39%|███▉      | 318/809 [06:22<10:01,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4419     0.1442         63        640:  39%|███▉      | 319/809 [06:23<09:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4422     0.1448        168        640:  40%|███▉      | 321/809 [06:25<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4422     0.1447        160        640:  40%|███▉      | 323/809 [06:28<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4424     0.1446        291        640:  40%|████      | 324/809 [06:29<09:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4424     0.1451        378        640:  40%|████      | 326/809 [06:31<09:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4428     0.1449        202        640:  41%|████      | 328/809 [06:33<09:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4428     0.1447        157        640:  41%|████      | 329/809 [06:35<09:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173      0.443     0.1446        202        640:  41%|████      | 331/809 [06:37<09:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4428     0.1451        427        640:  41%|████      | 333/809 [06:40<09:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4429      0.145        215        640:  41%|████▏     | 334/809 [06:41<09:42,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4431     0.1447        149        640:  42%|████▏     | 336/809 [06:43<09:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174      0.443     0.1447        127        640:  42%|████▏     | 338/809 [06:46<09:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4429     0.1445        201        640:  42%|████▏     | 339/809 [06:47<09:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4426     0.1444        243        640:  42%|████▏     | 341/809 [06:49<09:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4433     0.1442         61        640:  42%|████▏     | 343/809 [06:52<09:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4433     0.1441        332        640:  43%|████▎     | 344/809 [06:53<09:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4435     0.1439        136        640:  43%|████▎     | 346/809 [06:55<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4438     0.1435        115        640:  43%|████▎     | 348/809 [06:57<08:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4439     0.1434        190        640:  43%|████▎     | 349/809 [06:59<09:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171      0.444     0.1433        196        640:  43%|████▎     | 351/809 [07:01<09:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4438     0.1432        319        640:  44%|████▎     | 353/809 [07:04<09:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4438     0.1431        371        640:  44%|████▍     | 354/809 [07:05<09:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4438     0.1429        310        640:  44%|████▍     | 356/809 [07:07<08:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4438     0.1427        201        640:  44%|████▍     | 358/809 [07:09<09:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4444     0.1425        228        640:  44%|████▍     | 360/809 [07:12<08:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4444     0.1423        356        640:  45%|████▍     | 361/809 [07:13<08:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4446     0.1421        242        640:  45%|████▍     | 363/809 [07:15<08:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4443     0.1419        281        640:  45%|████▌     | 365/809 [07:18<08:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442      0.142        228        640:  45%|████▌     | 366/809 [07:19<08:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4445     0.1418         73        640:  45%|████▌     | 368/809 [07:21<08:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4444     0.1417        155        640:  46%|████▌     | 370/809 [07:24<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4445     0.1417        326        640:  46%|████▌     | 372/809 [07:26<08:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4439     0.1418        250        640:  46%|████▌     | 374/809 [07:28<08:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4441     0.1417        251        640:  46%|████▋     | 375/809 [07:30<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4441     0.1415        280        640:  47%|████▋     | 377/809 [07:32<08:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1415        180        640:  47%|████▋     | 379/809 [07:34<08:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1415        448        640:  47%|████▋     | 380/809 [07:36<08:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4447     0.1412        299        640:  47%|████▋     | 382/809 [07:38<08:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4448      0.141        399        640:  47%|████▋     | 384/809 [07:40<08:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4449     0.1409        171        640:  48%|████▊     | 385/809 [07:42<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4446     0.1408        246        640:  48%|████▊     | 387/809 [07:44<08:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4444     0.1406        231        640:  48%|████▊     | 389/809 [07:46<08:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1407        403        640:  48%|████▊     | 390/809 [07:48<08:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4442     0.1405        128        640:  48%|████▊     | 392/809 [07:50<08:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1405        213        640:  49%|████▊     | 394/809 [07:52<08:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4446     0.1404         92        640:  49%|████▉     | 395/809 [07:54<08:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4445     0.1403        290        640:  49%|████▉     | 397/809 [07:56<08:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4445     0.1404        119        640:  49%|████▉     | 399/809 [07:58<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4444     0.1404        324        640:  49%|████▉     | 400/809 [07:59<08:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4443     0.1402        242        640:  50%|████▉     | 402/809 [08:02<08:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4439     0.1405        354        640:  50%|████▉     | 404/809 [08:04<08:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4439     0.1404        223        640:  50%|█████     | 405/809 [08:06<08:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4439     0.1403        286        640:  50%|█████     | 407/809 [08:08<07:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4436     0.1404        409        640:  51%|█████     | 409/809 [08:10<08:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4436     0.1405        336        640:  51%|█████     | 410/809 [08:12<07:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174      0.444     0.1403        201        640:  51%|█████     | 412/809 [08:14<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4439     0.1403        247        640:  51%|█████     | 414/809 [08:16<07:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4438     0.1403        395        640:  51%|█████▏    | 415/809 [08:18<07:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174      0.444     0.1401        258        640:  52%|█████▏    | 417/809 [08:20<07:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4439     0.1407        248        640:  52%|█████▏    | 419/809 [08:23<08:03,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4441     0.1406        223        640:  52%|█████▏    | 420/809 [08:24<07:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175      0.444     0.1407        287        640:  52%|█████▏    | 422/809 [08:26<08:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4437     0.1406        213        640:  52%|█████▏    | 424/809 [08:29<07:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.177     0.4434     0.1407        271        640:  53%|█████▎    | 425/809 [08:30<07:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4433     0.1406        223        640:  53%|█████▎    | 427/809 [08:32<07:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.177     0.4432     0.1406        256        640:  53%|█████▎    | 429/809 [08:34<07:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.177      0.443     0.1406        251        640:  53%|█████▎    | 430/809 [08:36<07:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.176     0.4435     0.1404        139        640:  53%|█████▎    | 432/809 [08:38<07:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4436     0.1403        286        640:  54%|█████▎    | 434/809 [08:40<07:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4435     0.1402        182        640:  54%|█████▍    | 436/809 [08:43<07:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4437     0.1401        262        640:  54%|█████▍    | 437/809 [08:44<07:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4439     0.1399        255        640:  54%|█████▍    | 439/809 [08:46<07:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4441     0.1397        286        640:  55%|█████▍    | 441/809 [08:49<07:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1396        103        640:  55%|█████▍    | 442/809 [08:50<07:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4446     0.1396        247        640:  55%|█████▍    | 444/809 [08:52<07:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4448     0.1397        170        640:  55%|█████▌    | 446/809 [08:55<07:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4447     0.1396        198        640:  55%|█████▌    | 447/809 [08:56<07:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4447     0.1395        341        640:  56%|█████▌    | 449/809 [08:58<07:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4445     0.1395        355        640:  56%|█████▌    | 451/809 [09:01<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4446     0.1395        248        640:  56%|█████▌    | 452/809 [09:02<06:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4445     0.1395        124        640:  56%|█████▌    | 454/809 [09:04<07:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4448     0.1395        180        640:  56%|█████▋    | 456/809 [09:06<06:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4448     0.1393        228        640:  57%|█████▋    | 458/809 [09:09<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4447     0.1394        372        640:  57%|█████▋    | 459/809 [09:10<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4449     0.1392        265        640:  57%|█████▋    | 461/809 [09:12<06:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4452     0.1393        289        640:  57%|█████▋    | 463/809 [09:15<06:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4453     0.1392        268        640:  57%|█████▋    | 464/809 [09:16<06:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4453     0.1391        311        640:  58%|█████▊    | 466/809 [09:18<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4452      0.139        256        640:  58%|█████▊    | 468/809 [09:20<06:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4453      0.139         98        640:  58%|█████▊    | 469/809 [09:22<06:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4453      0.139        203        640:  58%|█████▊    | 471/809 [09:24<06:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4451     0.1389        117        640:  58%|█████▊    | 473/809 [09:26<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4448     0.1389        200        640:  59%|██��██▊    | 475/809 [09:29<06:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4448     0.1388        158        640:  59%|█████▉    | 476/809 [09:30<06:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4444     0.1393        139        640:  59%|█████▉    | 478/809 [09:32<06:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4442     0.1394        405        640:  59%|█████▉    | 479/809 [09:34<06:45,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4441     0.1394        403        640:  59%|█████▉    | 481/809 [09:36<06:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173      0.444     0.1395        379        640:  60%|█████▉    | 483/809 [09:38<06:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4441     0.1394        169        640:  60%|█████▉    | 485/809 [09:41<06:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4442     0.1393        142        640:  60%|██████    | 486/809 [09:42<06:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4444     0.1391        115        640:  60%|██████    | 488/809 [09:44<06:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4449     0.1389        152        640:  61%|██████    | 490/809 [09:47<06:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4451     0.1388        250        640:  61%|██████    | 491/809 [09:48<06:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4451     0.1389        267        640:  61%|██████    | 493/809 [09:50<06:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4451     0.1389        437        640:  61%|██████    | 495/809 [09:53<06:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4451     0.1388        328        640:  61%|██████▏   | 496/809 [09:54<06:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171      0.445     0.1389        185        640:  62%|██████▏   | 498/809 [09:56<06:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4454     0.1388        253        640:  62%|██████▏   | 500/809 [09:59<06:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4456     0.1387        126        640:  62%|██████▏   | 501/809 [10:00<06:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4459     0.1385        249        640:  62%|██████▏   | 503/809 [10:02<06:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4463     0.1383        133        640:  62%|██████▏   | 505/809 [10:04<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4461     0.1383        353        640:  63%|██████▎   | 507/809 [10:07<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4458     0.1384        317        640:  63%|██████▎   | 508/809 [10:08<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4459     0.1382        312        640:  63%|██████▎   | 510/809 [10:10<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4463     0.1382        115        640:  63%|██████▎   | 512/809 [10:13<06:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4464     0.1382        183        640:  63%|██████▎   | 513/809 [10:14<05:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4463     0.1381        460        640:  64%|██████▎   | 515/809 [10:16<05:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4464     0.1381        191        640:  64%|██████▍   | 517/809 [10:19<05:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4463     0.1383        395        640:  64%|██████▍   | 519/809 [10:21<05:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4462     0.1382        150        640:  64%|██████▍   | 520/809 [10:22<05:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4461     0.1381        292        640:  65%|██████▍   | 522/809 [10:25<05:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4462     0.1379        183        640:  65%|██████▍   | 524/809 [10:27<05:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165      0.446     0.1379        268        640:  65%|██████▍   | 525/809 [10:28<05:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4461     0.1379        156        640:  65%|██████▌   | 527/809 [10:31<05:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.164     0.4465     0.1377        315        640:  65%|██████▌   | 529/809 [10:33<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.163     0.4466     0.1376        264        640:  66%|██████▌   | 530/809 [10:34<05:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.163     0.4468     0.1376        278        640:  66%|██████▌   | 532/809 [10:37<05:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.162     0.4471     0.1374        323        640:  66%|██████▌   | 534/809 [10:39<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.161     0.4473     0.1372        180        640:  66%|██████▋   | 536/809 [10:41<05:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4473     0.1371        161        640:  66%|██████▋   | 537/809 [10:43<05:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4474     0.1369        162        640:  67%|██████▋   | 539/809 [10:45<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4473      0.137        301        640:  67%|██████▋   | 541/809 [10:47<05:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4474      0.137        137        640:  67%|██████▋   | 542/809 [10:49<05:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4476     0.1368        130        640:  67%|██████▋   | 544/809 [10:51<05:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4476     0.1367        243        640:  67%|██████▋   | 546/809 [10:53<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4476     0.1366        162        640:  68%|██████▊   | 547/809 [10:54<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4474     0.1366        326        640:  68%|██████▊   | 549/809 [10:57<05:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4472     0.1366        318        640:  68%|██████▊   | 551/809 [10:59<05:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4473     0.1365        162        640:  68%|██████▊   | 552/809 [11:00<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4473     0.1364        236        640:  68%|██████▊   | 554/809 [11:03<05:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4471     0.1364        380        640:  69%|██████▊   | 556/809 [11:05<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.158     0.4473     0.1363        183        640:  69%|██████▉   | 557/809 [11:06<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4469     0.1368        802        640:  69%|██████▉   | 559/809 [11:09<05:14,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16      0.447     0.1367        226        640:  69%|██████▉   | 561/809 [11:11<05:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4471     0.1367        256        640:  69%|██████▉   | 562/809 [11:13<04:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4471      0.137        109        640:  70%|██████▉   | 564/809 [11:15<04:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4471     0.1368        156        640:  70%|██████▉   | 566/809 [11:17<04:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4472     0.1368        257        640:  70%|███████   | 567/809 [11:19<04:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.158     0.4475     0.1367        218        640:  70%|███████   | 569/809 [11:21<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.158     0.4474     0.1367        300        640:  71%|███████   | 571/809 [11:23<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.158     0.4475     0.1366        284        640:  71%|███████   | 572/809 [11:24<04:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4475     0.1366        209        640:  71%|███████   | 574/809 [11:27<04:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4476     0.1365        120        640:  71%|███████   | 576/809 [11:29<04:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4475     0.1365        557        640:  71%|███���███▏  | 577/809 [11:31<04:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4476     0.1364        204        640:  72%|███████▏  | 579/809 [11:33<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4474     0.1367        512        640:  72%|███████▏  | 581/809 [11:35<04:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4474     0.1366        228        640:  72%|███████▏  | 582/809 [11:37<04:38,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.156     0.4476     0.1365        183        640:  72%|███████▏  | 584/809 [11:39<04:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4475     0.1367        361        640:  72%|███████▏  | 586/809 [11:41<04:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4474     0.1367        321        640:  73%|███████▎  | 587/809 [11:43<04:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4476     0.1366        192        640:  73%|███████▎  | 589/809 [11:45<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.156     0.4476     0.1365        335        640:  73%|███████▎  | 591/809 [11:47<04:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.156     0.4479     0.1365        175        640:  73%|███████▎  | 592/809 [11:49<04:26,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.156     0.4478     0.1365        105        640:  73%|███████▎  | 594/809 [11:51<04:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.157     0.4477     0.1366        350        640:  74%|███████▎  | 596/809 [11:53<04:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.158     0.4474      0.137        270        640:  74%|███████▍  | 597/809 [11:55<04:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159     0.4471      0.137         93        640:  74%|███████▍  | 599/809 [11:57<04:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.159      0.447      0.137        149        640:  74%|███████▍  | 601/809 [12:00<04:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4467     0.1371        445        640:  74%|███████▍  | 602/809 [12:01<04:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.16     0.4466     0.1371        305        640:  75%|███████▍  | 604/809 [12:03<04:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.162     0.4462     0.1371        269        640:  75%|███████▍  | 606/809 [12:06<04:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.162     0.4462     0.1372        306        640:  75%|███████▌  | 607/809 [12:07<04:08,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.161     0.4464     0.1371        146        640:  75%|███████▌  | 609/809 [12:09<04:08,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.161     0.4464      0.137        169        640:  76%|███████▌  | 611/809 [12:12<03:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.162     0.4462     0.1374        477        640:  76%|███████▌  | 612/809 [12:13<04:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.163      0.446     0.1377        436        640:  76%|███████▌  | 614/809 [12:15<04:03,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.163     0.4462     0.1377         77        640:  76%|███████▌  | 615/809 [12:17<03:59,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.162     0.4463     0.1376        131        640:  76%|███████▋  | 617/809 [12:19<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.164     0.4459     0.1378        216        640:  77%|███████▋  | 619/809 [12:21<03:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.164     0.4458     0.1379        319        640:  77%|███████▋  | 620/809 [12:23<03:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4455     0.1379        281        640:  77%|███████▋  | 622/809 [12:25<03:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4451     0.1389        346        640:  77%|███████▋  | 624/809 [12:28<03:52,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167      0.445     0.1389        316        640:  77%|███████▋  | 625/809 [12:29<03:46,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4452     0.1388        290        640:  78%|███████▊  | 627/809 [12:31<03:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4453     0.1387        235        640:  78%|███████▊  | 629/809 [12:34<03:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.165     0.4455     0.1386        253        640:  78%|███████▊  | 630/809 [12:35<03:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4452     0.1388        394        640:  78%|███████▊  | 632/809 [12:37<03:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4452     0.1387        201        640:  78%|███████▊  | 634/809 [12:39<03:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.166     0.4452     0.1387        372        640:  78%|███████▊  | 635/809 [12:41<03:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167      0.445     0.1386        345        640:  79%|███████▊  | 637/809 [12:43<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167      0.445     0.1386        214        640:  79%|███████▉  | 639/809 [12:46<03:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.167     0.4448     0.1388        634        640:  79%|███████▉  | 640/809 [12:47<03:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4447     0.1387        148        640:  79%|███████▉  | 642/809 [12:49<03:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4445     0.1387        392        640:  80%|███████▉  | 644/809 [12:52<03:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4443     0.1391        438        640:  80%|███████▉  | 645/809 [12:53<03:21,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4441     0.1391        362        640:  80%|███████▉  | 647/809 [12:55<03:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4436     0.1392        251        640:  80%|████████  | 649/809 [12:58<03:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4433     0.1393        305        640:  80%|████████  | 650/809 [12:59<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4428     0.1393        249        640:  81%|████████  | 652/809 [13:01<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4422     0.1395        478        640:  81%|████████  | 654/809 [13:04<03:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4423     0.1394        173        640:  81%|████████  | 656/809 [13:06<03:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4427     0.1393        228        640:  81%|████████▏ | 658/809 [13:08<02:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4427     0.1393        269        640:  81%|████████▏ | 659/809 [13:10<02:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4425     0.1394        365        640:  82%|████████▏ | 661/809 [13:12<02:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4423     0.1394        221        640:  82%|████████▏ | 662/809 [13:13<02:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4424     0.1394        140        640:  82%|████████▏ | 664/809 [13:16<02:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4425     0.1393        362        640:  82%|████████▏ | 666/809 [13:18<02:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4424     0.1392        253        640:  83%|████████▎ | 668/809 [13:20<02:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4424     0.1392        176        640:  83%|████████▎ | 669/809 [13:21<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4427     0.1391        142        640:  83%|████████▎ | 671/809 [13:24<02:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4428     0.1389        147        640:  83%|████████▎ | 673/809 [13:26<02:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4428     0.1389        155        640:  83%|████████▎ | 674/809 [13:27<02:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4426      0.139        468        640:  84%|████████▎ | 676/809 [13:30<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4428     0.1388        180        640:  84%|████████▍ | 678/809 [13:32<02:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1387        197        640:  84%|████████▍ | 680/809 [13:34<02:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1387        187        640:  84%|████████▍ | 681/809 [13:36<02:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4425     0.1389        450        640:  84%|████████▍ | 683/809 [13:38<02:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4424     0.1389        348        640:  85%|████████▍ | 685/809 [13:40<02:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4423     0.1389        194        640:  85%|████████▍ | 686/809 [13:42<02:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4426     0.1389         90        640:  85%|████████▌ | 688/809 [13:44<02:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4426     0.1389        166        640:  85%|████████▌ | 690/809 [13:46<02:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1389        288        640:  85%|████████▌ | 691/809 [13:47<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4428     0.1389        105        640:  86%|████████▌ | 693/809 [13:50<02:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4428      0.139         89        640:  86%|████████▌ | 695/809 [13:52<02:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4427      0.139        327        640:  86%|████████▌ | 696/809 [13:53<02:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4429     0.1389        319        640:  86%|████████▋ | 698/809 [13:56<02:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4432     0.1388        206        640:  87%|████████▋ | 700/809 [13:58<02:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4432     0.1388        248        640:  87%|████████▋ | 702/809 [14:01<02:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4433     0.1387        152        640:  87%|████████▋ | 703/809 [14:02<02:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4432     0.1387        323        640:  87%|████████▋ | 705/809 [14:04<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4432     0.1387        423        640:  87%|████████▋ | 707/809 [14:06<02:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4432     0.1387        325        640:  88%|████████▊ | 708/809 [14:08<02:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4429     0.1387        343        640:  88%|████████▊ | 710/809 [14:10<02:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169      0.443     0.1386        188        640:  88%|████████▊ | 712/809 [14:13<01:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4431     0.1386        174        640:  88%|████████▊ | 713/809 [14:14<01:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4428     0.1387        257        640:  88%|████████▊ | 715/809 [14:16<01:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4427     0.1386        127        640:  89%|████████▊ | 717/809 [14:19<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4428     0.1387        339        640:  89%|████████▉ | 718/809 [14:20<01:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4428     0.1386        271        640:  89%|████████▉ | 720/809 [14:22<01:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169      0.443     0.1385        193        640:  89%|████████▉ | 722/809 [14:24<01:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4432     0.1384        101        640:  89%|████████▉ | 723/809 [14:26<01:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169      0.443     0.1387        422        640:  90%|████████▉ | 725/809 [14:28<01:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4431     0.1386        154        640:  90%|████████▉ | 727/809 [14:30<01:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.168     0.4432     0.1385        140        640:  90%|████████▉ | 728/809 [14:32<01:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4428     0.1387        423        640:  90%|█████████ | 730/809 [14:34<01:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4426     0.1387        253        640:  90%|█████████ | 732/809 [14:36<01:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4426     0.1386        147        640:  91%|█████████ | 734/809 [14:39<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4425     0.1386        192        640:  91%|█████████ | 735/809 [14:40<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4427     0.1385        197        640:  91%|█████████ | 737/809 [14:42<01:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4426     0.1386        477        640:  91%|█████████▏| 739/809 [14:45<01:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4428     0.1386        389        640:  91%|█████████▏| 740/809 [14:46<01:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169      0.443     0.1385        121        640:  92%|█████████▏| 742/809 [14:48<01:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.169     0.4429     0.1384        419        640:  92%|█████████▏| 744/809 [14:51<01:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4427     0.1386        573        640:  92%|█████████▏| 745/809 [14:52<01:19,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4428     0.1385        304        640:  92%|█████████▏| 747/809 [14:54<01:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4426     0.1385        473        640:  92%|█████████▏| 748/809 [14:56<01:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4423     0.1386        341        640:  93%|█████████▎| 750/809 [14:58<01:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4424     0.1385        160        640:  93%|█████████▎| 752/809 [15:00<01:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G       1.17     0.4427     0.1384        206        640:  93%|█████████▎| 754/809 [15:03<01:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.171     0.4425     0.1386        498        640:  93%|█████████▎| 755/809 [15:04<01:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4424     0.1386        291        640:  94%|█████████▎| 757/809 [15:06<01:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4422     0.1388        468        640:  94%|█████████▎| 758/809 [15:08<01:03,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1391        519        640:  94%|█████████▍| 760/809 [15:10<01:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1391        176        640:  94%|█████████▍| 762/809 [15:12<00:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1391        215        640:  94%|██��██████▍| 764/809 [15:15<00:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4416     0.1392        295        640:  95%|█████████▍| 765/809 [15:16<00:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418      0.139        221        640:  95%|█████████▍| 767/809 [15:18<00:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418      0.139        180        640:  95%|█████████▌| 769/809 [15:21<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418      0.139        253        640:  95%|█████████▌| 770/809 [15:22<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1389        195        640:  95%|█████████▌| 772/809 [15:24<00:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1389        202        640:  96%|█████████▌| 774/809 [15:27<00:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418      0.139        206        640:  96%|█████████▌| 775/809 [15:28<00:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4418     0.1389        210        640:  96%|█████████▌| 777/809 [15:30<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4417     0.1391        444        640:  96%|█████████▋| 779/809 [15:33<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4416     0.1391        278        640:  96%|█████████���| 780/809 [15:34<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4416      0.139        104        640:  97%|█████████▋| 782/809 [15:36<00:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4416      0.139        162        640:  97%|█████████▋| 784/809 [15:38<00:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4416     0.1389        284        640:  97%|█████████▋| 786/809 [15:41<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4417     0.1389        163        640:  97%|█████████▋| 787/809 [15:42<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4418     0.1388         84        640:  98%|█████████▊| 789/809 [15:44<00:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4418     0.1389        163        640:  98%|█████████▊| 791/809 [15:47<00:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4418      0.139        318        640:  98%|█████████▊| 792/809 [15:48<00:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.175     0.4419     0.1389        391        640:  98%|█████████▊| 794/809 [15:50<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4422     0.1388        211        640:  98%|█████████▊| 796/809 [15:53<00:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.174     0.4423     0.1387        117        640:  99%|█████████▊| 797/809 [15:54<00:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4425     0.1386        101        640:  99%|█████████▉| 799/809 [15:56<00:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4424     0.1386        348        640:  99%|█████████▉| 801/809 [15:58<00:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1384        141        640:  99%|█████████▉| 803/809 [16:01<00:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4427     0.1384        322        640: 100%|█████████▉| 805/809 [16:03<00:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.172     0.4428     0.1384        368        640: 100%|█████████▉| 806/809 [16:04<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     44/300      13.7G      1.173     0.4426     0.1386        388        640: 100%|█████████▉| 808/809 [16:07<00:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:16,  3.99it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:14,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:12,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  54%|█████▎    | 37/69 [00:09<00:08,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:11<00:07,  3.56it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:13<00:04,  3.77it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:15<00:02,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.7G      1.175     0.4156     0.1119        332        640:   0%|          | 2/809 [00:02<15:43,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.7G        1.3      0.392     0.1483        552        640:   0%|          | 3/809 [00:03<16:06,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.058     0.4548     0.1182        196        640:   1%|          | 5/809 [00:06<16:44,  1.25s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.066      0.475     0.1125        363        640:   1%|          | 7/809 [00:08<16:12,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.099     0.4634     0.1196        322        640:   1%|          | 8/809 [00:09<16:08,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.114     0.4648     0.1191        299        640:   1%|          | 10/809 [00:12<16:04,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.102     0.4707     0.1147        177        640:   1%|▏         | 12/809 [00:14<15:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.108     0.4711     0.1144        269        640:   2%|▏         | 13/809 [00:15<16:25,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.099     0.4758     0.1138         75        640:   2%|▏         | 15/809 [00:18<16:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.108      0.474     0.1179        288        640:   2%|▏         | 17/809 [00:20<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.111     0.4727     0.1199        225        640:   2%|▏         | 18/809 [00:21<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.122     0.4702     0.1239        241        640:   2%|▏         | 20/809 [00:24<15:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.124     0.4703     0.1226        311        640:   3%|▎         | 22/809 [00:26<15:41,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.137     0.4665     0.1246        352        640:   3%|▎         | 23/809 [00:27<15:41,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G       1.16     0.4591     0.1251        281        640:   3%|▎         | 25/809 [00:30<15:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.172     0.4594      0.126        209        640:   3%|▎         | 27/809 [00:32<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.169     0.4619     0.1257        247        640:   3%|▎         | 28/809 [00:33<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.167     0.4623     0.1264        384        640:   4%|▎         | 30/809 [00:36<15:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.161     0.4641     0.1269        308        640:   4%|▍         | 32/809 [00:38<15:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G       1.16     0.4648     0.1278        265        640:   4%|▍         | 34/809 [00:40<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.155     0.4665     0.1268        181        640:   4%|▍         | 35/809 [00:41<15:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.136     0.4725     0.1257        104        640:   5%|▍         | 37/809 [00:44<15:30,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.133     0.4745     0.1244        271        640:   5%|▍         | 39/809 [00:46<15:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.129     0.4749     0.1237        214        640:   5%|▍         | 40/809 [00:47<15:37,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.135     0.4732     0.1231        405        640:   5%|▌         | 42/809 [00:50<15:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.128     0.4752     0.1237        189        640:   5%|▌         | 44/809 [00:52<14:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.126     0.4766     0.1229        283        640:   6%|▌         | 45/809 [00:53<15:16,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G       1.12     0.4783     0.1223        233        640:   6%|▌         | 47/809 [00:56<15:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.113     0.4819      0.123        333        640:   6%|▌         | 49/809 [00:58<14:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     45/300      13.8G      1.112     0.4812     0.1226        299        640:   6%|▌         | 50/809 [00:59<14:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)