2024-08-06 14:23:41,788 INFO [trainer.py:870] (4/8) Training started
2024-08-06 14:23:41,789 INFO [trainer.py:889] (4/8) Device: cuda:4
2024-08-06 14:23:41,789 INFO [trainer.py:890] (4/8) {'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 100, 'reset_interval': 200, 'valid_interval': 2000, 'env_info': {'k2-version': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '279b0c87015a615b81b147251814d737a548f397', 'k2-git-date': 'Wed May 24 22:24:09 2023', 'lhotse-version': '1.26.0', 'torch-version': '2.0.1+cu118', 'torch-cuda-available': True, 'torch-cuda-version': '11.8', 'python-version': '3.10', 'icefall-git-branch': None, 'icefall-git-sha1': None, 'icefall-git-date': None, 'icefall-path': '/workspace/icefall_llm', 'k2-path': '/usr/local/lib/python3.10/dist-packages/k2/__init__.py', 'lhotse-path': '/usr/local/lib/python3.10/dist-packages/lhotse/__init__.py', 'hostname': '6867463', 'IP address': '0.104.202.7'}, 'world_size': 8, 'master_port': 12354, 'tensorboard': True, 'num_epochs': 40, 'start_epoch': 100, 'start_batch': 0, 'exp_dir': PosixPath('exp/valle'), 'optimizer_name': 'ScaledAdam', 'scheduler_name': 'Eden', 'base_lr': 0.03, 'warmup_steps': 200, 'seed': 42, 'inf_check': False, 'save_every_n': 100000, 'keep_last_k': 20, 'average_period': 0, 'accumulate_grad_steps': 2, 'dtype': 'float32', 'filter_min_duration': 0.5, 'filter_max_duration': 14.0, 'train_stage': 2, 'visualize': False, 'oom_check': False, 'model_name': 'valle', 'decoder_dim': 1024, 'nhead': 16, 'num_decoder_layers': 12, 'scale_factor': 1.0, 'norm_first': True, 'add_prenet': False, 'prefix_mode': 1, 'share_embedding': True, 'prepend_bos': False, 'num_quantizers': 8, 'scaling_xformers': False, 'manifest_dir': PosixPath('data/tokenized'), 'max_duration': 160, 'bucketing_sampler': True, 'num_buckets': 6, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 0.1, 'on_the_fly_feats': False, 'shuffle': True, 'buffer_size': 40000, 'shuffle_buffer_size': 100000, 'drop_last': False, 'return_cuts': True, 'num_workers': 8, 'enable_spec_aug': False, 'spec_aug_time_warp_factor': 80, 'input_strategy': 'PrecomputedFeatures', 'dataset': 'libritts', 'text_tokens': 'data/tokenized/unique_text_tokens.k2symbols', 'sampling_rate': 24000}
2024-08-06 14:23:41,789 INFO [trainer.py:892] (4/8) About to create model
2024-08-06 14:23:42,584 INFO [trainer.py:899] (4/8) Number of model parameters: 367386628
2024-08-06 14:23:42,585 INFO [checkpoint.py:112] (4/8) Loading checkpoint from exp/valle/epoch-99.pt
2024-08-06 14:23:47,587 INFO [trainer.py:914] (4/8) Using DDP
2024-08-06 14:23:49,643 INFO [datamodule.py:427] (4/8) About to get train cuts
2024-08-06 14:23:49,645 INFO [datamodule.py:434] (4/8) About to get dev cuts
2024-08-06 14:23:49,646 INFO [datamodule.py:292] (4/8) Disable SpecAugment
2024-08-06 14:23:49,646 INFO [datamodule.py:294] (4/8) About to create train dataset
2024-08-06 14:23:49,647 INFO [datamodule.py:323] (4/8) Using DynamicBucketingSampler
2024-08-06 14:23:50,269 INFO [datamodule.py:344] (4/8) About to create train dataloader
2024-08-06 14:23:50,270 INFO [datamodule.py:367] (4/8) About to create dev dataset
2024-08-06 14:23:50,602 INFO [datamodule.py:388] (4/8) About to create dev dataloader
2024-08-06 14:24:38,249 INFO [trainer.py:765] (4/8) Epoch 1, batch 100, train_loss[loss=106.9, NarTop10Accuracy=0.02083, over 7263.00 frames. ], tot_loss[loss=74.22, NarTop10Accuracy=0.04496, over 2352.16 frames. ], batch size: 31, lr: 2.25e-02
2024-08-06 14:25:07,518 INFO [trainer.py:765] (4/8) Epoch 1, batch 200, train_loss[loss=133.9, NarTop10Accuracy=0.01548, over 6891.00 frames. ], tot_loss[loss=97.68, NarTop10Accuracy=0.04142, over 3862.76 frames. ], batch size: 17, lr: 3.00e-02
2024-08-06 14:25:37,111 INFO [trainer.py:765] (4/8) Epoch 1, batch 300, train_loss[loss=99.71, NarTop10Accuracy=0.02296, over 7125.00 frames. ], tot_loss[loss=85.35, NarTop10Accuracy=0.04282, over 4657.50 frames. ], batch size: 22, lr: 3.00e-02
2024-08-06 14:26:07,482 INFO [trainer.py:765] (4/8) Epoch 1, batch 400, train_loss[loss=53.17, NarTop10Accuracy=0.02305, over 5133.00 frames. ], tot_loss[loss=67.96, NarTop10Accuracy=0.04728, over 5101.12 frames. ], batch size: 7, lr: 3.00e-02
2024-08-06 14:26:35,357 INFO [trainer.py:765] (4/8) Epoch 1, batch 500, train_loss[loss=14.77, NarTop10Accuracy=0.02519, over 6039.00 frames. ], tot_loss[loss=49.13, NarTop10Accuracy=0.05014, over 5373.17 frames. ], batch size: 11, lr: 2.99e-02
2024-08-06 14:27:04,000 INFO [trainer.py:765] (4/8) Epoch 1, batch 600, train_loss[loss=6.172, NarTop10Accuracy=0.1872, over 5658.00 frames. ], tot_loss[loss=33.43, NarTop10Accuracy=0.05587, over 5653.66 frames. ], batch size: 9, lr: 2.99e-02
2024-08-06 14:27:39,490 INFO [trainer.py:765] (4/8) Epoch 1, batch 700, train_loss[loss=6.809, NarTop10Accuracy=0.1082, over 5073.00 frames. ], tot_loss[loss=23.38, NarTop10Accuracy=0.06459, over 5734.64 frames. ], batch size: 6, lr: 2.99e-02
2024-08-06 14:28:08,832 INFO [trainer.py:765] (4/8) Epoch 1, batch 800, train_loss[loss=6.48, NarTop10Accuracy=0.1294, over 4425.00 frames. ], tot_loss[loss=17.14, NarTop10Accuracy=0.08496, over 5790.91 frames. ], batch size: 5, lr: 2.98e-02
2024-08-06 14:28:36,758 INFO [trainer.py:765] (4/8) Epoch 1, batch 900, train_loss[loss=5.851, NarTop10Accuracy=0.1589, over 6615.00 frames. ], tot_loss[loss=12.79, NarTop10Accuracy=0.113, over 5800.33 frames. ], batch size: 14, lr: 2.98e-02
2024-08-06 14:29:12,586 INFO [trainer.py:765] (4/8) Epoch 1, batch 1000, train_loss[loss=5.578, NarTop10Accuracy=0.2194, over 6165.00 frames. ], tot_loss[loss=10.11, NarTop10Accuracy=0.1343, over 5888.51 frames. ], batch size: 13, lr: 2.97e-02
2024-08-06 14:29:42,825 INFO [trainer.py:765] (4/8) Epoch 1, batch 1100, train_loss[loss=5.681, NarTop10Accuracy=0.196, over 6762.00 frames. ], tot_loss[loss=8.423, NarTop10Accuracy=0.1542, over 5904.54 frames. ], batch size: 17, lr: 2.96e-02
2024-08-06 14:30:11,468 INFO [trainer.py:765] (4/8) Epoch 1, batch 1200, train_loss[loss=5.958, NarTop10Accuracy=0.1501, over 7260.00 frames. ], tot_loss[loss=7.351, NarTop10Accuracy=0.1711, over 5918.74 frames. ], batch size: 31, lr: 2.96e-02
2024-08-06 14:30:48,747 INFO [trainer.py:765] (4/8) Epoch 1, batch 1300, train_loss[loss=5.341, NarTop10Accuracy=0.2774, over 5139.00 frames. ], tot_loss[loss=6.679, NarTop10Accuracy=0.1872, over 5989.25 frames. ], batch size: 6, lr: 2.95e-02
2024-08-06 14:31:18,143 INFO [trainer.py:765] (4/8) Epoch 1, batch 1400, train_loss[loss=5.591, NarTop10Accuracy=0.2127, over 6189.00 frames. ], tot_loss[loss=6.251, NarTop10Accuracy=0.198, over 6004.68 frames. ], batch size: 11, lr: 2.94e-02
2024-08-06 14:31:46,026 INFO [trainer.py:765] (4/8) Epoch 1, batch 1500, train_loss[loss=5.788, NarTop10Accuracy=0.1813, over 6327.00 frames. ], tot_loss[loss=5.965, NarTop10Accuracy=0.2105, over 5950.52 frames. ], batch size: 50, lr: 2.94e-02
2024-08-06 14:32:13,691 INFO [trainer.py:765] (4/8) Epoch 1, batch 1600, train_loss[loss=5.667, NarTop10Accuracy=0.1998, over 7182.00 frames. ], tot_loss[loss=5.791, NarTop10Accuracy=0.2177, over 5924.47 frames. ], batch size: 22, lr: 2.93e-02
2024-08-06 14:32:40,198 INFO [trainer.py:765] (4/8) Epoch 1, batch 1700, train_loss[loss=5.231, NarTop10Accuracy=0.2883, over 6195.00 frames. ], tot_loss[loss=5.667, NarTop10Accuracy=0.2255, over 5920.15 frames. ], batch size: 13, lr: 2.92e-02
2024-08-06 14:33:06,499 INFO [trainer.py:765] (4/8) Epoch 1, batch 1800, train_loss[loss=5.481, NarTop10Accuracy=0.2219, over 7254.00 frames. ], tot_loss[loss=5.575, NarTop10Accuracy=0.2331, over 5977.88 frames. ], batch size: 22, lr: 2.91e-02
2024-08-06 14:33:32,625 INFO [trainer.py:765] (4/8) Epoch 1, batch 1900, train_loss[loss=5.716, NarTop10Accuracy=0.1903, over 6150.00 frames. ], tot_loss[loss=5.51, NarTop10Accuracy=0.2405, over 6021.45 frames. ], batch size: 50, lr: 2.90e-02
2024-08-06 14:33:58,014 INFO [trainer.py:765] (4/8) Epoch 1, batch 2000, train_loss[loss=5.56, NarTop10Accuracy=0.2315, over 6456.00 frames. ], tot_loss[loss=5.451, NarTop10Accuracy=0.249, over 6014.64 frames. ], batch size: 50, lr: 2.89e-02
2024-08-06 14:33:58,015 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 14:34:06,103 INFO [trainer.py:811] (4/8) Epoch 1, validation: loss=5.397, NarTop10Accuracy=0.2581, over 1905321.00 frames. 
2024-08-06 14:34:06,104 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 26462MB
2024-08-06 14:34:06,612 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 4.749e+01 2.278e+02 7.300e+02 1.664e+04 7.177e+05, threshold=1.460e+03, percent-clipped=0.0
2024-08-06 14:34:32,061 INFO [trainer.py:765] (4/8) Epoch 1, batch 2100, train_loss[loss=5.187, NarTop10Accuracy=0.2958, over 4017.00 frames. ], tot_loss[loss=5.385, NarTop10Accuracy=0.2598, over 5979.06 frames. ], batch size: 4, lr: 2.88e-02
2024-08-06 14:34:57,303 INFO [trainer.py:765] (4/8) Epoch 1, batch 2200, train_loss[loss=5.44, NarTop10Accuracy=0.2485, over 7404.00 frames. ], tot_loss[loss=5.346, NarTop10Accuracy=0.2657, over 6006.35 frames. ], batch size: 32, lr: 2.87e-02
2024-08-06 14:35:22,455 INFO [trainer.py:765] (4/8) Epoch 1, batch 2300, train_loss[loss=5.326, NarTop10Accuracy=0.2631, over 5673.00 frames. ], tot_loss[loss=5.339, NarTop10Accuracy=0.2665, over 6024.02 frames. ], batch size: 9, lr: 2.86e-02
2024-08-06 14:35:46,815 INFO [trainer.py:765] (4/8) Epoch 1, batch 2400, train_loss[loss=5.406, NarTop10Accuracy=0.2526, over 5178.00 frames. ], tot_loss[loss=5.284, NarTop10Accuracy=0.2769, over 5766.88 frames. ], batch size: 7, lr: 2.85e-02
2024-08-06 14:36:10,408 INFO [trainer.py:765] (4/8) Epoch 1, batch 2500, train_loss[loss=5.052, NarTop10Accuracy=0.3257, over 5352.00 frames. ], tot_loss[loss=5.217, NarTop10Accuracy=0.2886, over 5482.39 frames. ], batch size: 7, lr: 2.84e-02
2024-08-06 14:36:31,007 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 14:37:29,669 INFO [trainer.py:765] (4/8) Epoch 2, batch 100, train_loss[loss=4.983, NarTop10Accuracy=0.3321, over 7494.00 frames. ], tot_loss[loss=5.187, NarTop10Accuracy=0.2957, over 2365.70 frames. ], batch size: 32, lr: 2.77e-02
2024-08-06 14:38:10,014 INFO [trainer.py:765] (4/8) Epoch 2, batch 200, train_loss[loss=5.135, NarTop10Accuracy=0.3065, over 6768.00 frames. ], tot_loss[loss=5.159, NarTop10Accuracy=0.2999, over 3860.07 frames. ], batch size: 17, lr: 2.76e-02
2024-08-06 14:38:38,297 INFO [trainer.py:765] (4/8) Epoch 2, batch 300, train_loss[loss=5.184, NarTop10Accuracy=0.2962, over 7026.00 frames. ], tot_loss[loss=5.137, NarTop10Accuracy=0.3036, over 4663.32 frames. ], batch size: 22, lr: 2.75e-02
2024-08-06 14:39:06,998 INFO [trainer.py:765] (4/8) Epoch 2, batch 400, train_loss[loss=5.036, NarTop10Accuracy=0.3214, over 5193.00 frames. ], tot_loss[loss=5.115, NarTop10Accuracy=0.3071, over 5110.12 frames. ], batch size: 7, lr: 2.74e-02
2024-08-06 14:39:46,118 INFO [trainer.py:765] (4/8) Epoch 2, batch 500, train_loss[loss=4.872, NarTop10Accuracy=0.3493, over 6105.00 frames. ], tot_loss[loss=5.076, NarTop10Accuracy=0.315, over 5391.47 frames. ], batch size: 11, lr: 2.73e-02
2024-08-06 14:40:15,082 INFO [trainer.py:765] (4/8) Epoch 2, batch 600, train_loss[loss=4.753, NarTop10Accuracy=0.3896, over 5706.00 frames. ], tot_loss[loss=5.052, NarTop10Accuracy=0.3197, over 5655.03 frames. ], batch size: 9, lr: 2.71e-02
2024-08-06 14:40:44,589 INFO [trainer.py:765] (4/8) Epoch 2, batch 700, train_loss[loss=5.025, NarTop10Accuracy=0.3314, over 5022.00 frames. ], tot_loss[loss=5.036, NarTop10Accuracy=0.3225, over 5732.88 frames. ], batch size: 6, lr: 2.70e-02
2024-08-06 14:41:24,512 INFO [trainer.py:765] (4/8) Epoch 2, batch 800, train_loss[loss=4.941, NarTop10Accuracy=0.3415, over 4290.00 frames. ], tot_loss[loss=5.021, NarTop10Accuracy=0.3249, over 5771.04 frames. ], batch size: 5, lr: 2.69e-02
2024-08-06 14:41:54,404 INFO [trainer.py:765] (4/8) Epoch 2, batch 900, train_loss[loss=4.726, NarTop10Accuracy=0.3778, over 6720.00 frames. ], tot_loss[loss=4.985, NarTop10Accuracy=0.3316, over 5794.38 frames. ], batch size: 14, lr: 2.68e-02
2024-08-06 14:42:23,901 INFO [trainer.py:765] (4/8) Epoch 2, batch 1000, train_loss[loss=4.831, NarTop10Accuracy=0.3599, over 6669.00 frames. ], tot_loss[loss=4.952, NarTop10Accuracy=0.3377, over 5893.94 frames. ], batch size: 14, lr: 2.66e-02
2024-08-06 14:42:56,254 INFO [trainer.py:765] (4/8) Epoch 2, batch 1100, train_loss[loss=4.867, NarTop10Accuracy=0.3562, over 6915.00 frames. ], tot_loss[loss=4.934, NarTop10Accuracy=0.3413, over 5922.75 frames. ], batch size: 17, lr: 2.65e-02
2024-08-06 14:43:35,186 INFO [trainer.py:765] (4/8) Epoch 2, batch 1200, train_loss[loss=4.714, NarTop10Accuracy=0.387, over 7470.00 frames. ], tot_loss[loss=4.909, NarTop10Accuracy=0.3461, over 5906.24 frames. ], batch size: 31, lr: 2.64e-02
2024-08-06 14:44:04,345 INFO [trainer.py:765] (4/8) Epoch 2, batch 1300, train_loss[loss=4.845, NarTop10Accuracy=0.3592, over 4905.00 frames. ], tot_loss[loss=4.869, NarTop10Accuracy=0.3533, over 5992.07 frames. ], batch size: 6, lr: 2.63e-02
2024-08-06 14:44:33,728 INFO [trainer.py:765] (4/8) Epoch 2, batch 1400, train_loss[loss=5.029, NarTop10Accuracy=0.3247, over 6120.00 frames. ], tot_loss[loss=4.855, NarTop10Accuracy=0.3564, over 6016.47 frames. ], batch size: 11, lr: 2.61e-02
2024-08-06 14:44:40,441 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 14:44:48,506 INFO [trainer.py:811] (4/8) Epoch 2, validation: loss=4.808, NarTop10Accuracy=0.3642, over 1905321.00 frames. 
2024-08-06 14:44:48,506 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 14:44:49,204 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 6.328e+01 1.178e+02 1.410e+02 1.789e+02 6.269e+02, threshold=2.821e+02, percent-clipped=0.0
2024-08-06 14:45:09,806 INFO [trainer.py:765] (4/8) Epoch 2, batch 1500, train_loss[loss=4.779, NarTop10Accuracy=0.372, over 6327.00 frames. ], tot_loss[loss=4.831, NarTop10Accuracy=0.3606, over 5943.92 frames. ], batch size: 51, lr: 2.60e-02
2024-08-06 14:45:37,660 INFO [trainer.py:765] (4/8) Epoch 2, batch 1600, train_loss[loss=4.765, NarTop10Accuracy=0.3742, over 7113.00 frames. ], tot_loss[loss=4.8, NarTop10Accuracy=0.3666, over 5921.48 frames. ], batch size: 22, lr: 2.59e-02
2024-08-06 14:46:04,368 INFO [trainer.py:765] (4/8) Epoch 2, batch 1700, train_loss[loss=4.756, NarTop10Accuracy=0.372, over 6198.00 frames. ], tot_loss[loss=4.789, NarTop10Accuracy=0.3691, over 5899.56 frames. ], batch size: 13, lr: 2.58e-02
2024-08-06 14:46:31,033 INFO [trainer.py:765] (4/8) Epoch 2, batch 1800, train_loss[loss=4.78, NarTop10Accuracy=0.3722, over 6996.00 frames. ], tot_loss[loss=4.773, NarTop10Accuracy=0.372, over 5988.57 frames. ], batch size: 22, lr: 2.56e-02
2024-08-06 14:46:57,531 INFO [trainer.py:765] (4/8) Epoch 2, batch 1900, train_loss[loss=4.739, NarTop10Accuracy=0.3737, over 6135.00 frames. ], tot_loss[loss=4.752, NarTop10Accuracy=0.3759, over 6027.10 frames. ], batch size: 53, lr: 2.55e-02
2024-08-06 14:47:23,233 INFO [trainer.py:765] (4/8) Epoch 2, batch 2000, train_loss[loss=4.75, NarTop10Accuracy=0.3835, over 5940.00 frames. ], tot_loss[loss=4.724, NarTop10Accuracy=0.3812, over 6007.84 frames. ], batch size: 50, lr: 2.54e-02
2024-08-06 14:47:48,588 INFO [trainer.py:765] (4/8) Epoch 2, batch 2100, train_loss[loss=4.823, NarTop10Accuracy=0.3596, over 4026.00 frames. ], tot_loss[loss=4.712, NarTop10Accuracy=0.3832, over 5972.64 frames. ], batch size: 4, lr: 2.53e-02
2024-08-06 14:48:13,764 INFO [trainer.py:765] (4/8) Epoch 2, batch 2200, train_loss[loss=4.647, NarTop10Accuracy=0.3897, over 7305.00 frames. ], tot_loss[loss=4.672, NarTop10Accuracy=0.3908, over 6015.55 frames. ], batch size: 31, lr: 2.51e-02
2024-08-06 14:48:38,951 INFO [trainer.py:765] (4/8) Epoch 2, batch 2300, train_loss[loss=4.859, NarTop10Accuracy=0.3563, over 5718.00 frames. ], tot_loss[loss=4.678, NarTop10Accuracy=0.3894, over 6013.04 frames. ], batch size: 9, lr: 2.50e-02
2024-08-06 14:49:03,319 INFO [trainer.py:765] (4/8) Epoch 2, batch 2400, train_loss[loss=4.334, NarTop10Accuracy=0.4646, over 5088.00 frames. ], tot_loss[loss=4.647, NarTop10Accuracy=0.3954, over 5779.63 frames. ], batch size: 7, lr: 2.49e-02
2024-08-06 14:49:26,867 INFO [trainer.py:765] (4/8) Epoch 2, batch 2500, train_loss[loss=4.633, NarTop10Accuracy=0.3957, over 5151.00 frames. ], tot_loss[loss=4.619, NarTop10Accuracy=0.4008, over 5484.17 frames. ], batch size: 7, lr: 2.48e-02
2024-08-06 14:49:46,802 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 14:50:51,117 INFO [trainer.py:765] (4/8) Epoch 3, batch 100, train_loss[loss=4.654, NarTop10Accuracy=0.385, over 7356.00 frames. ], tot_loss[loss=4.586, NarTop10Accuracy=0.4076, over 2369.28 frames. ], batch size: 31, lr: 2.36e-02
2024-08-06 14:51:20,387 INFO [trainer.py:765] (4/8) Epoch 3, batch 200, train_loss[loss=4.603, NarTop10Accuracy=0.3984, over 6894.00 frames. ], tot_loss[loss=4.549, NarTop10Accuracy=0.4148, over 3862.32 frames. ], batch size: 17, lr: 2.34e-02
2024-08-06 14:51:50,954 INFO [trainer.py:765] (4/8) Epoch 3, batch 300, train_loss[loss=4.699, NarTop10Accuracy=0.3826, over 7017.00 frames. ], tot_loss[loss=4.526, NarTop10Accuracy=0.4192, over 4665.53 frames. ], batch size: 22, lr: 2.33e-02
2024-08-06 14:52:32,359 INFO [trainer.py:765] (4/8) Epoch 3, batch 400, train_loss[loss=4.545, NarTop10Accuracy=0.4116, over 5163.00 frames. ], tot_loss[loss=4.497, NarTop10Accuracy=0.4252, over 5098.38 frames. ], batch size: 7, lr: 2.32e-02
2024-08-06 14:53:00,680 INFO [trainer.py:765] (4/8) Epoch 3, batch 500, train_loss[loss=4.423, NarTop10Accuracy=0.4433, over 6069.00 frames. ], tot_loss[loss=4.491, NarTop10Accuracy=0.4262, over 5373.12 frames. ], batch size: 11, lr: 2.31e-02
2024-08-06 14:53:29,551 INFO [trainer.py:765] (4/8) Epoch 3, batch 600, train_loss[loss=4.282, NarTop10Accuracy=0.4664, over 5613.00 frames. ], tot_loss[loss=4.478, NarTop10Accuracy=0.429, over 5644.85 frames. ], batch size: 9, lr: 2.30e-02
2024-08-06 14:54:12,466 INFO [trainer.py:765] (4/8) Epoch 3, batch 700, train_loss[loss=4.441, NarTop10Accuracy=0.4389, over 5016.00 frames. ], tot_loss[loss=4.45, NarTop10Accuracy=0.4347, over 5721.70 frames. ], batch size: 6, lr: 2.29e-02
2024-08-06 14:54:44,785 INFO [trainer.py:765] (4/8) Epoch 3, batch 800, train_loss[loss=4.153, NarTop10Accuracy=0.4893, over 5031.00 frames. ], tot_loss[loss=4.427, NarTop10Accuracy=0.4391, over 5774.85 frames. ], batch size: 6, lr: 2.28e-02
2024-08-06 14:54:58,684 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 14:55:06,655 INFO [trainer.py:811] (4/8) Epoch 3, validation: loss=4.276, NarTop10Accuracy=0.4689, over 1905321.00 frames. 
2024-08-06 14:55:06,656 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 14:55:07,183 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 8.443e+01 1.396e+02 1.639e+02 2.017e+02 7.124e+02, threshold=3.277e+02, percent-clipped=4.5
2024-08-06 14:55:21,052 INFO [trainer.py:765] (4/8) Epoch 3, batch 900, train_loss[loss=4.153, NarTop10Accuracy=0.4938, over 6390.00 frames. ], tot_loss[loss=4.391, NarTop10Accuracy=0.4461, over 5785.87 frames. ], batch size: 13, lr: 2.26e-02
2024-08-06 14:56:04,958 INFO [trainer.py:765] (4/8) Epoch 3, batch 1000, train_loss[loss=4.124, NarTop10Accuracy=0.5011, over 6243.00 frames. ], tot_loss[loss=4.366, NarTop10Accuracy=0.4511, over 5879.98 frames. ], batch size: 13, lr: 2.25e-02
2024-08-06 14:56:37,301 INFO [trainer.py:765] (4/8) Epoch 3, batch 1100, train_loss[loss=4.422, NarTop10Accuracy=0.4279, over 6579.00 frames. ], tot_loss[loss=4.348, NarTop10Accuracy=0.4544, over 5918.50 frames. ], batch size: 17, lr: 2.24e-02
2024-08-06 14:57:06,377 INFO [trainer.py:765] (4/8) Epoch 3, batch 1200, train_loss[loss=4.308, NarTop10Accuracy=0.4475, over 7182.00 frames. ], tot_loss[loss=4.331, NarTop10Accuracy=0.4576, over 5919.47 frames. ], batch size: 31, lr: 2.23e-02
2024-08-06 14:57:51,631 INFO [trainer.py:765] (4/8) Epoch 3, batch 1300, train_loss[loss=4.112, NarTop10Accuracy=0.5039, over 5100.00 frames. ], tot_loss[loss=4.304, NarTop10Accuracy=0.4631, over 5994.55 frames. ], batch size: 6, lr: 2.22e-02
2024-08-06 14:58:22,900 INFO [trainer.py:765] (4/8) Epoch 3, batch 1400, train_loss[loss=4.1, NarTop10Accuracy=0.5005, over 6117.00 frames. ], tot_loss[loss=4.294, NarTop10Accuracy=0.4647, over 6021.33 frames. ], batch size: 11, lr: 2.21e-02
2024-08-06 14:58:50,855 INFO [trainer.py:765] (4/8) Epoch 3, batch 1500, train_loss[loss=4.339, NarTop10Accuracy=0.4654, over 5736.00 frames. ], tot_loss[loss=4.27, NarTop10Accuracy=0.4694, over 5948.51 frames. ], batch size: 50, lr: 2.20e-02
2024-08-06 14:59:18,715 INFO [trainer.py:765] (4/8) Epoch 3, batch 1600, train_loss[loss=3.932, NarTop10Accuracy=0.5352, over 7194.00 frames. ], tot_loss[loss=4.253, NarTop10Accuracy=0.4725, over 5942.22 frames. ], batch size: 22, lr: 2.19e-02
2024-08-06 14:59:45,953 INFO [trainer.py:765] (4/8) Epoch 3, batch 1700, train_loss[loss=4.044, NarTop10Accuracy=0.5174, over 6528.00 frames. ], tot_loss[loss=4.232, NarTop10Accuracy=0.4766, over 5922.57 frames. ], batch size: 14, lr: 2.18e-02
2024-08-06 15:00:12,498 INFO [trainer.py:765] (4/8) Epoch 3, batch 1800, train_loss[loss=4.041, NarTop10Accuracy=0.5179, over 7068.00 frames. ], tot_loss[loss=4.215, NarTop10Accuracy=0.4801, over 5986.63 frames. ], batch size: 22, lr: 2.17e-02
2024-08-06 15:00:38,949 INFO [trainer.py:765] (4/8) Epoch 3, batch 1900, train_loss[loss=4.624, NarTop10Accuracy=0.3976, over 5904.00 frames. ], tot_loss[loss=4.197, NarTop10Accuracy=0.4843, over 6024.64 frames. ], batch size: 52, lr: 2.16e-02
2024-08-06 15:01:04,606 INFO [trainer.py:765] (4/8) Epoch 3, batch 2000, train_loss[loss=4.521, NarTop10Accuracy=0.4144, over 5823.00 frames. ], tot_loss[loss=4.17, NarTop10Accuracy=0.4895, over 5988.92 frames. ], batch size: 50, lr: 2.15e-02
2024-08-06 15:01:29,899 INFO [trainer.py:765] (4/8) Epoch 3, batch 2100, train_loss[loss=3.928, NarTop10Accuracy=0.5386, over 4803.00 frames. ], tot_loss[loss=4.148, NarTop10Accuracy=0.4935, over 5967.62 frames. ], batch size: 5, lr: 2.14e-02
2024-08-06 15:01:55,182 INFO [trainer.py:765] (4/8) Epoch 3, batch 2200, train_loss[loss=3.961, NarTop10Accuracy=0.5316, over 7011.00 frames. ], tot_loss[loss=4.122, NarTop10Accuracy=0.4991, over 6012.29 frames. ], batch size: 31, lr: 2.13e-02
2024-08-06 15:02:20,410 INFO [trainer.py:765] (4/8) Epoch 3, batch 2300, train_loss[loss=4.458, NarTop10Accuracy=0.4342, over 5721.00 frames. ], tot_loss[loss=4.131, NarTop10Accuracy=0.4975, over 6027.00 frames. ], batch size: 9, lr: 2.12e-02
2024-08-06 15:02:44,664 INFO [trainer.py:765] (4/8) Epoch 3, batch 2400, train_loss[loss=4.152, NarTop10Accuracy=0.4884, over 5229.00 frames. ], tot_loss[loss=4.103, NarTop10Accuracy=0.5032, over 5756.56 frames. ], batch size: 7, lr: 2.11e-02
2024-08-06 15:03:08,234 INFO [trainer.py:765] (4/8) Epoch 3, batch 2500, train_loss[loss=3.853, NarTop10Accuracy=0.5569, over 5118.00 frames. ], tot_loss[loss=4.048, NarTop10Accuracy=0.5147, over 5479.20 frames. ], batch size: 7, lr: 2.10e-02
2024-08-06 15:03:28,349 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 15:04:28,130 INFO [trainer.py:765] (4/8) Epoch 4, batch 100, train_loss[loss=3.878, NarTop10Accuracy=0.5446, over 7209.00 frames. ], tot_loss[loss=4.035, NarTop10Accuracy=0.517, over 2366.50 frames. ], batch size: 31, lr: 1.97e-02
2024-08-06 15:04:59,841 INFO [trainer.py:765] (4/8) Epoch 4, batch 200, train_loss[loss=3.821, NarTop10Accuracy=0.5554, over 6888.00 frames. ], tot_loss[loss=4.005, NarTop10Accuracy=0.5231, over 3859.55 frames. ], batch size: 17, lr: 1.96e-02
2024-08-06 15:05:27,508 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:05:35,694 INFO [trainer.py:811] (4/8) Epoch 4, validation: loss=3.804, NarTop10Accuracy=0.5644, over 1905321.00 frames. 
2024-08-06 15:05:35,695 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 15:05:36,238 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.765e+02 1.975e+02 2.270e+02 5.852e+02, threshold=3.949e+02, percent-clipped=2.8
2024-08-06 15:05:43,889 INFO [trainer.py:765] (4/8) Epoch 4, batch 300, train_loss[loss=3.868, NarTop10Accuracy=0.5582, over 7239.00 frames. ], tot_loss[loss=3.992, NarTop10Accuracy=0.5258, over 4657.70 frames. ], batch size: 22, lr: 1.95e-02
2024-08-06 15:06:16,124 INFO [trainer.py:765] (4/8) Epoch 4, batch 400, train_loss[loss=3.819, NarTop10Accuracy=0.5639, over 5349.00 frames. ], tot_loss[loss=4.005, NarTop10Accuracy=0.5234, over 5115.90 frames. ], batch size: 7, lr: 1.94e-02
2024-08-06 15:06:46,473 INFO [trainer.py:765] (4/8) Epoch 4, batch 500, train_loss[loss=4.156, NarTop10Accuracy=0.4833, over 5994.00 frames. ], tot_loss[loss=3.987, NarTop10Accuracy=0.5272, over 5401.97 frames. ], batch size: 11, lr: 1.93e-02
2024-08-06 15:07:23,818 INFO [trainer.py:765] (4/8) Epoch 4, batch 600, train_loss[loss=3.637, NarTop10Accuracy=0.5894, over 5694.00 frames. ], tot_loss[loss=3.974, NarTop10Accuracy=0.5295, over 5653.17 frames. ], batch size: 9, lr: 1.93e-02
2024-08-06 15:07:59,001 INFO [trainer.py:765] (4/8) Epoch 4, batch 700, train_loss[loss=4.274, NarTop10Accuracy=0.4614, over 5136.00 frames. ], tot_loss[loss=3.974, NarTop10Accuracy=0.5296, over 5719.17 frames. ], batch size: 6, lr: 1.92e-02
2024-08-06 15:08:32,430 INFO [trainer.py:765] (4/8) Epoch 4, batch 800, train_loss[loss=3.36, NarTop10Accuracy=0.6409, over 5175.00 frames. ], tot_loss[loss=3.96, NarTop10Accuracy=0.5323, over 5791.46 frames. ], batch size: 6, lr: 1.91e-02
2024-08-06 15:09:10,689 INFO [trainer.py:765] (4/8) Epoch 4, batch 900, train_loss[loss=3.627, NarTop10Accuracy=0.5981, over 6111.00 frames. ], tot_loss[loss=3.927, NarTop10Accuracy=0.5393, over 5786.61 frames. ], batch size: 13, lr: 1.90e-02
2024-08-06 15:09:46,075 INFO [trainer.py:765] (4/8) Epoch 4, batch 1000, train_loss[loss=3.636, NarTop10Accuracy=0.592, over 6237.00 frames. ], tot_loss[loss=3.921, NarTop10Accuracy=0.5404, over 5895.13 frames. ], batch size: 13, lr: 1.89e-02
2024-08-06 15:10:18,139 INFO [trainer.py:765] (4/8) Epoch 4, batch 1100, train_loss[loss=3.882, NarTop10Accuracy=0.5455, over 6867.00 frames. ], tot_loss[loss=3.907, NarTop10Accuracy=0.5435, over 5934.97 frames. ], batch size: 17, lr: 1.88e-02
2024-08-06 15:10:55,075 INFO [trainer.py:765] (4/8) Epoch 4, batch 1200, train_loss[loss=4.31, NarTop10Accuracy=0.4531, over 7113.00 frames. ], tot_loss[loss=3.904, NarTop10Accuracy=0.5438, over 5932.64 frames. ], batch size: 31, lr: 1.88e-02
2024-08-06 15:11:32,074 INFO [trainer.py:765] (4/8) Epoch 4, batch 1300, train_loss[loss=3.512, NarTop10Accuracy=0.6291, over 5055.00 frames. ], tot_loss[loss=3.865, NarTop10Accuracy=0.5517, over 5992.71 frames. ], batch size: 6, lr: 1.87e-02
2024-08-06 15:12:05,688 INFO [trainer.py:765] (4/8) Epoch 4, batch 1400, train_loss[loss=3.67, NarTop10Accuracy=0.5999, over 6057.00 frames. ], tot_loss[loss=3.856, NarTop10Accuracy=0.5538, over 5999.72 frames. ], batch size: 11, lr: 1.86e-02
2024-08-06 15:12:33,695 INFO [trainer.py:765] (4/8) Epoch 4, batch 1500, train_loss[loss=3.845, NarTop10Accuracy=0.5672, over 5805.00 frames. ], tot_loss[loss=3.86, NarTop10Accuracy=0.5529, over 5943.15 frames. ], batch size: 50, lr: 1.85e-02
2024-08-06 15:13:01,510 INFO [trainer.py:765] (4/8) Epoch 4, batch 1600, train_loss[loss=3.812, NarTop10Accuracy=0.5678, over 7086.00 frames. ], tot_loss[loss=3.852, NarTop10Accuracy=0.5546, over 5937.46 frames. ], batch size: 22, lr: 1.84e-02
2024-08-06 15:13:28,132 INFO [trainer.py:765] (4/8) Epoch 4, batch 1700, train_loss[loss=3.656, NarTop10Accuracy=0.595, over 6597.00 frames. ], tot_loss[loss=3.828, NarTop10Accuracy=0.5593, over 5915.58 frames. ], batch size: 14, lr: 1.84e-02
2024-08-06 15:13:54,558 INFO [trainer.py:765] (4/8) Epoch 4, batch 1800, train_loss[loss=3.932, NarTop10Accuracy=0.5465, over 7134.00 frames. ], tot_loss[loss=3.827, NarTop10Accuracy=0.5594, over 5974.57 frames. ], batch size: 22, lr: 1.83e-02
2024-08-06 15:14:20,998 INFO [trainer.py:765] (4/8) Epoch 4, batch 1900, train_loss[loss=3.787, NarTop10Accuracy=0.5686, over 5676.00 frames. ], tot_loss[loss=3.848, NarTop10Accuracy=0.5552, over 6021.45 frames. ], batch size: 50, lr: 1.82e-02
2024-08-06 15:14:46,671 INFO [trainer.py:765] (4/8) Epoch 4, batch 2000, train_loss[loss=3.72, NarTop10Accuracy=0.5847, over 5940.00 frames. ], tot_loss[loss=3.821, NarTop10Accuracy=0.5608, over 6000.89 frames. ], batch size: 51, lr: 1.81e-02
2024-08-06 15:15:11,858 INFO [trainer.py:765] (4/8) Epoch 4, batch 2100, train_loss[loss=3.593, NarTop10Accuracy=0.5992, over 4773.00 frames. ], tot_loss[loss=3.812, NarTop10Accuracy=0.562, over 5982.67 frames. ], batch size: 5, lr: 1.81e-02
2024-08-06 15:15:37,089 INFO [trainer.py:765] (4/8) Epoch 4, batch 2200, train_loss[loss=3.774, NarTop10Accuracy=0.5774, over 7401.00 frames. ], tot_loss[loss=3.802, NarTop10Accuracy=0.564, over 6016.10 frames. ], batch size: 31, lr: 1.80e-02
2024-08-06 15:15:55,089 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:16:03,243 INFO [trainer.py:811] (4/8) Epoch 4, validation: loss=3.665, NarTop10Accuracy=0.5912, over 1905321.00 frames. 
2024-08-06 15:16:03,243 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 15:16:03,741 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.414e+02 1.889e+02 2.096e+02 2.369e+02 1.168e+03, threshold=4.192e+02, percent-clipped=1.7
2024-08-06 15:16:10,347 INFO [trainer.py:765] (4/8) Epoch 4, batch 2300, train_loss[loss=3.645, NarTop10Accuracy=0.5987, over 5703.00 frames. ], tot_loss[loss=3.814, NarTop10Accuracy=0.5618, over 6029.16 frames. ], batch size: 9, lr: 1.79e-02
2024-08-06 15:16:34,840 INFO [trainer.py:765] (4/8) Epoch 4, batch 2400, train_loss[loss=3.313, NarTop10Accuracy=0.6544, over 5088.00 frames. ], tot_loss[loss=3.78, NarTop10Accuracy=0.5684, over 5783.26 frames. ], batch size: 7, lr: 1.79e-02
2024-08-06 15:16:58,535 INFO [trainer.py:765] (4/8) Epoch 4, batch 2500, train_loss[loss=3.431, NarTop10Accuracy=0.6503, over 5121.00 frames. ], tot_loss[loss=3.767, NarTop10Accuracy=0.5714, over 5489.80 frames. ], batch size: 7, lr: 1.78e-02
2024-08-06 15:17:18,371 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 15:18:24,100 INFO [trainer.py:765] (4/8) Epoch 5, batch 100, train_loss[loss=3.562, NarTop10Accuracy=0.6205, over 7200.00 frames. ], tot_loss[loss=3.778, NarTop10Accuracy=0.5692, over 2371.91 frames. ], batch size: 31, lr: 1.66e-02
2024-08-06 15:18:59,675 INFO [trainer.py:765] (4/8) Epoch 5, batch 200, train_loss[loss=4.033, NarTop10Accuracy=0.5126, over 7221.00 frames. ], tot_loss[loss=3.759, NarTop10Accuracy=0.5733, over 3874.25 frames. ], batch size: 18, lr: 1.65e-02
2024-08-06 15:19:32,888 INFO [trainer.py:765] (4/8) Epoch 5, batch 300, train_loss[loss=3.964, NarTop10Accuracy=0.5257, over 6888.00 frames. ], tot_loss[loss=3.722, NarTop10Accuracy=0.5811, over 4675.45 frames. ], batch size: 22, lr: 1.65e-02
2024-08-06 15:20:01,656 INFO [trainer.py:765] (4/8) Epoch 5, batch 400, train_loss[loss=3.521, NarTop10Accuracy=0.621, over 5163.00 frames. ], tot_loss[loss=3.722, NarTop10Accuracy=0.581, over 5123.18 frames. ], batch size: 7, lr: 1.64e-02
2024-08-06 15:20:38,299 INFO [trainer.py:765] (4/8) Epoch 5, batch 500, train_loss[loss=3.909, NarTop10Accuracy=0.5439, over 6174.00 frames. ], tot_loss[loss=3.734, NarTop10Accuracy=0.5779, over 5382.99 frames. ], batch size: 11, lr: 1.63e-02
2024-08-06 15:21:13,711 INFO [trainer.py:765] (4/8) Epoch 5, batch 600, train_loss[loss=3.833, NarTop10Accuracy=0.5501, over 5688.00 frames. ], tot_loss[loss=3.72, NarTop10Accuracy=0.5805, over 5653.92 frames. ], batch size: 9, lr: 1.63e-02
2024-08-06 15:21:45,881 INFO [trainer.py:765] (4/8) Epoch 5, batch 700, train_loss[loss=3.57, NarTop10Accuracy=0.6122, over 4212.00 frames. ], tot_loss[loss=3.716, NarTop10Accuracy=0.5815, over 5700.66 frames. ], batch size: 5, lr: 1.62e-02
2024-08-06 15:22:24,498 INFO [trainer.py:765] (4/8) Epoch 5, batch 800, train_loss[loss=3.965, NarTop10Accuracy=0.5297, over 5022.00 frames. ], tot_loss[loss=3.707, NarTop10Accuracy=0.5835, over 5746.15 frames. ], batch size: 6, lr: 1.62e-02
2024-08-06 15:22:56,783 INFO [trainer.py:765] (4/8) Epoch 5, batch 900, train_loss[loss=3.512, NarTop10Accuracy=0.6159, over 6609.00 frames. ], tot_loss[loss=3.696, NarTop10Accuracy=0.5856, over 5775.64 frames. ], batch size: 14, lr: 1.61e-02
2024-08-06 15:23:31,914 INFO [trainer.py:765] (4/8) Epoch 5, batch 1000, train_loss[loss=3.532, NarTop10Accuracy=0.6177, over 6321.00 frames. ], tot_loss[loss=3.68, NarTop10Accuracy=0.5886, over 5886.91 frames. ], batch size: 13, lr: 1.60e-02
2024-08-06 15:24:09,572 INFO [trainer.py:765] (4/8) Epoch 5, batch 1100, train_loss[loss=3.539, NarTop10Accuracy=0.6228, over 6765.00 frames. ], tot_loss[loss=3.682, NarTop10Accuracy=0.589, over 5931.39 frames. ], batch size: 17, lr: 1.60e-02
2024-08-06 15:24:44,529 INFO [trainer.py:765] (4/8) Epoch 5, batch 1200, train_loss[loss=3.527, NarTop10Accuracy=0.6216, over 7326.00 frames. ], tot_loss[loss=3.685, NarTop10Accuracy=0.5881, over 5933.73 frames. ], batch size: 32, lr: 1.59e-02
2024-08-06 15:25:19,380 INFO [trainer.py:765] (4/8) Epoch 5, batch 1300, train_loss[loss=3.979, NarTop10Accuracy=0.5253, over 4971.00 frames. ], tot_loss[loss=3.665, NarTop10Accuracy=0.592, over 5989.61 frames. ], batch size: 6, lr: 1.59e-02
2024-08-06 15:25:51,694 INFO [trainer.py:765] (4/8) Epoch 5, batch 1400, train_loss[loss=4.036, NarTop10Accuracy=0.5187, over 6027.00 frames. ], tot_loss[loss=3.675, NarTop10Accuracy=0.5902, over 6015.46 frames. ], batch size: 11, lr: 1.58e-02
2024-08-06 15:26:26,195 INFO [trainer.py:765] (4/8) Epoch 5, batch 1500, train_loss[loss=3.618, NarTop10Accuracy=0.6069, over 6102.00 frames. ], tot_loss[loss=3.671, NarTop10Accuracy=0.5911, over 5950.45 frames. ], batch size: 50, lr: 1.58e-02
2024-08-06 15:26:54,130 INFO [trainer.py:765] (4/8) Epoch 5, batch 1600, train_loss[loss=3.356, NarTop10Accuracy=0.6533, over 7068.00 frames. ], tot_loss[loss=3.681, NarTop10Accuracy=0.589, over 5923.98 frames. ], batch size: 22, lr: 1.57e-02
2024-08-06 15:27:19,603 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:27:27,821 INFO [trainer.py:811] (4/8) Epoch 5, validation: loss=3.552, NarTop10Accuracy=0.6147, over 1905321.00 frames. 
2024-08-06 15:27:27,822 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 15:27:28,341 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.340e+02 1.756e+02 1.962e+02 2.205e+02 5.880e+02, threshold=3.924e+02, percent-clipped=0.8
2024-08-06 15:27:29,131 INFO [trainer.py:765] (4/8) Epoch 5, batch 1700, train_loss[loss=3.677, NarTop10Accuracy=0.5881, over 6720.00 frames. ], tot_loss[loss=3.67, NarTop10Accuracy=0.5908, over 5934.18 frames. ], batch size: 14, lr: 1.56e-02
2024-08-06 15:27:55,652 INFO [trainer.py:765] (4/8) Epoch 5, batch 1800, train_loss[loss=3.956, NarTop10Accuracy=0.5262, over 7311.00 frames. ], tot_loss[loss=3.674, NarTop10Accuracy=0.5898, over 5978.16 frames. ], batch size: 22, lr: 1.56e-02
2024-08-06 15:28:22,172 INFO [trainer.py:765] (4/8) Epoch 5, batch 1900, train_loss[loss=3.709, NarTop10Accuracy=0.589, over 5844.00 frames. ], tot_loss[loss=3.679, NarTop10Accuracy=0.589, over 6018.65 frames. ], batch size: 52, lr: 1.55e-02
2024-08-06 15:28:47,893 INFO [trainer.py:765] (4/8) Epoch 5, batch 2000, train_loss[loss=3.569, NarTop10Accuracy=0.6195, over 6192.00 frames. ], tot_loss[loss=3.677, NarTop10Accuracy=0.5893, over 5996.44 frames. ], batch size: 51, lr: 1.55e-02
2024-08-06 15:29:13,769 INFO [trainer.py:765] (4/8) Epoch 5, batch 2100, train_loss[loss=3.383, NarTop10Accuracy=0.6587, over 3798.00 frames. ], tot_loss[loss=3.689, NarTop10Accuracy=0.5868, over 5973.32 frames. ], batch size: 4, lr: 1.54e-02
2024-08-06 15:29:39,177 INFO [trainer.py:765] (4/8) Epoch 5, batch 2200, train_loss[loss=4.057, NarTop10Accuracy=0.5108, over 7401.00 frames. ], tot_loss[loss=3.664, NarTop10Accuracy=0.5916, over 6007.58 frames. ], batch size: 32, lr: 1.54e-02
2024-08-06 15:30:04,430 INFO [trainer.py:765] (4/8) Epoch 5, batch 2300, train_loss[loss=3.246, NarTop10Accuracy=0.6714, over 5784.00 frames. ], tot_loss[loss=3.669, NarTop10Accuracy=0.5909, over 6014.75 frames. ], batch size: 9, lr: 1.53e-02
2024-08-06 15:30:28,862 INFO [trainer.py:765] (4/8) Epoch 5, batch 2400, train_loss[loss=3.383, NarTop10Accuracy=0.6529, over 5055.00 frames. ], tot_loss[loss=3.646, NarTop10Accuracy=0.5958, over 5773.10 frames. ], batch size: 7, lr: 1.53e-02
2024-08-06 15:30:52,503 INFO [trainer.py:765] (4/8) Epoch 5, batch 2500, train_loss[loss=3.404, NarTop10Accuracy=0.6454, over 5151.00 frames. ], tot_loss[loss=3.609, NarTop10Accuracy=0.603, over 5465.51 frames. ], batch size: 7, lr: 1.52e-02
2024-08-06 15:31:12,414 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 15:32:14,415 INFO [trainer.py:765] (4/8) Epoch 6, batch 100, train_loss[loss=3.405, NarTop10Accuracy=0.6416, over 7386.00 frames. ], tot_loss[loss=3.626, NarTop10Accuracy=0.6004, over 2386.72 frames. ], batch size: 32, lr: 1.42e-02
2024-08-06 15:32:46,015 INFO [trainer.py:765] (4/8) Epoch 6, batch 200, train_loss[loss=4.047, NarTop10Accuracy=0.5003, over 6789.00 frames. ], tot_loss[loss=3.606, NarTop10Accuracy=0.6043, over 3859.23 frames. ], batch size: 17, lr: 1.42e-02
2024-08-06 15:33:21,243 INFO [trainer.py:765] (4/8) Epoch 6, batch 300, train_loss[loss=3.503, NarTop10Accuracy=0.6318, over 7152.00 frames. ], tot_loss[loss=3.601, NarTop10Accuracy=0.605, over 4654.32 frames. ], batch size: 22, lr: 1.41e-02
2024-08-06 15:33:56,035 INFO [trainer.py:765] (4/8) Epoch 6, batch 400, train_loss[loss=3.453, NarTop10Accuracy=0.6265, over 5214.00 frames. ], tot_loss[loss=3.591, NarTop10Accuracy=0.6072, over 5089.09 frames. ], batch size: 7, lr: 1.41e-02
2024-08-06 15:34:26,759 INFO [trainer.py:765] (4/8) Epoch 6, batch 500, train_loss[loss=3.394, NarTop10Accuracy=0.6526, over 6039.00 frames. ], tot_loss[loss=3.575, NarTop10Accuracy=0.6103, over 5374.47 frames. ], batch size: 11, lr: 1.40e-02
2024-08-06 15:35:01,458 INFO [trainer.py:765] (4/8) Epoch 6, batch 600, train_loss[loss=3.298, NarTop10Accuracy=0.6721, over 5883.00 frames. ], tot_loss[loss=3.583, NarTop10Accuracy=0.6091, over 5638.02 frames. ], batch size: 9, lr: 1.40e-02
2024-08-06 15:35:32,735 INFO [trainer.py:765] (4/8) Epoch 6, batch 700, train_loss[loss=3.591, NarTop10Accuracy=0.6136, over 5121.00 frames. ], tot_loss[loss=3.58, NarTop10Accuracy=0.6099, over 5698.61 frames. ], batch size: 6, lr: 1.39e-02
2024-08-06 15:36:06,844 INFO [trainer.py:765] (4/8) Epoch 6, batch 800, train_loss[loss=3.809, NarTop10Accuracy=0.5489, over 5283.00 frames. ], tot_loss[loss=3.591, NarTop10Accuracy=0.6072, over 5763.84 frames. ], batch size: 6, lr: 1.39e-02
2024-08-06 15:36:40,385 INFO [trainer.py:765] (4/8) Epoch 6, batch 900, train_loss[loss=4.068, NarTop10Accuracy=0.5072, over 6291.00 frames. ], tot_loss[loss=3.576, NarTop10Accuracy=0.6101, over 5779.05 frames. ], batch size: 13, lr: 1.38e-02
2024-08-06 15:37:15,272 INFO [trainer.py:765] (4/8) Epoch 6, batch 1000, train_loss[loss=3.266, NarTop10Accuracy=0.6727, over 6579.00 frames. ], tot_loss[loss=3.59, NarTop10Accuracy=0.6072, over 5866.73 frames. ], batch size: 14, lr: 1.38e-02
2024-08-06 15:37:50,508 INFO [trainer.py:765] (4/8) Epoch 6, batch 1100, train_loss[loss=3.536, NarTop10Accuracy=0.622, over 6768.00 frames. ], tot_loss[loss=3.593, NarTop10Accuracy=0.6064, over 5913.92 frames. ], batch size: 17, lr: 1.38e-02
2024-08-06 15:37:55,828 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:38:04,436 INFO [trainer.py:811] (4/8) Epoch 6, validation: loss=3.421, NarTop10Accuracy=0.6418, over 1905321.00 frames. 
2024-08-06 15:38:04,437 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27356MB
2024-08-06 15:38:04,965 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.415e+02 1.809e+02 1.991e+02 2.234e+02 5.215e+02, threshold=3.983e+02, percent-clipped=0.5
2024-08-06 15:38:36,169 INFO [trainer.py:765] (4/8) Epoch 6, batch 1200, train_loss[loss=3.43, NarTop10Accuracy=0.6318, over 7320.00 frames. ], tot_loss[loss=3.58, NarTop10Accuracy=0.6092, over 5923.93 frames. ], batch size: 31, lr: 1.37e-02
2024-08-06 15:39:08,242 INFO [trainer.py:765] (4/8) Epoch 6, batch 1300, train_loss[loss=3.323, NarTop10Accuracy=0.6618, over 4398.00 frames. ], tot_loss[loss=3.575, NarTop10Accuracy=0.6101, over 5989.65 frames. ], batch size: 5, lr: 1.37e-02
2024-08-06 15:39:44,070 INFO [trainer.py:765] (4/8) Epoch 6, batch 1400, train_loss[loss=3.471, NarTop10Accuracy=0.6306, over 6135.00 frames. ], tot_loss[loss=3.574, NarTop10Accuracy=0.6109, over 6025.10 frames. ], batch size: 11, lr: 1.36e-02
2024-08-06 15:40:15,383 INFO [trainer.py:765] (4/8) Epoch 6, batch 1500, train_loss[loss=3.963, NarTop10Accuracy=0.5223, over 6033.00 frames. ], tot_loss[loss=3.573, NarTop10Accuracy=0.6104, over 5955.89 frames. ], batch size: 50, lr: 1.36e-02
2024-08-06 15:40:43,106 INFO [trainer.py:765] (4/8) Epoch 6, batch 1600, train_loss[loss=3.409, NarTop10Accuracy=0.6413, over 7359.00 frames. ], tot_loss[loss=3.57, NarTop10Accuracy=0.6113, over 5930.43 frames. ], batch size: 23, lr: 1.35e-02
2024-08-06 15:41:09,789 INFO [trainer.py:765] (4/8) Epoch 6, batch 1700, train_loss[loss=3.395, NarTop10Accuracy=0.65, over 6255.00 frames. ], tot_loss[loss=3.556, NarTop10Accuracy=0.6142, over 5928.90 frames. ], batch size: 13, lr: 1.35e-02
2024-08-06 15:41:36,317 INFO [trainer.py:765] (4/8) Epoch 6, batch 1800, train_loss[loss=3.367, NarTop10Accuracy=0.6516, over 7140.00 frames. ], tot_loss[loss=3.563, NarTop10Accuracy=0.6123, over 5983.66 frames. ], batch size: 22, lr: 1.35e-02
2024-08-06 15:42:02,720 INFO [trainer.py:765] (4/8) Epoch 6, batch 1900, train_loss[loss=3.833, NarTop10Accuracy=0.5565, over 6012.00 frames. ], tot_loss[loss=3.583, NarTop10Accuracy=0.6083, over 6026.36 frames. ], batch size: 51, lr: 1.34e-02
2024-08-06 15:42:28,319 INFO [trainer.py:765] (4/8) Epoch 6, batch 2000, train_loss[loss=3.526, NarTop10Accuracy=0.6214, over 5970.00 frames. ], tot_loss[loss=3.57, NarTop10Accuracy=0.6108, over 6005.95 frames. ], batch size: 50, lr: 1.34e-02
2024-08-06 15:42:53,668 INFO [trainer.py:765] (4/8) Epoch 6, batch 2100, train_loss[loss=3.385, NarTop10Accuracy=0.6564, over 4812.00 frames. ], tot_loss[loss=3.56, NarTop10Accuracy=0.6132, over 5974.84 frames. ], batch size: 5, lr: 1.33e-02
2024-08-06 15:43:18,977 INFO [trainer.py:765] (4/8) Epoch 6, batch 2200, train_loss[loss=3.667, NarTop10Accuracy=0.5933, over 7131.00 frames. ], tot_loss[loss=3.568, NarTop10Accuracy=0.6116, over 6017.77 frames. ], batch size: 31, lr: 1.33e-02
2024-08-06 15:43:44,105 INFO [trainer.py:765] (4/8) Epoch 6, batch 2300, train_loss[loss=3.224, NarTop10Accuracy=0.6811, over 5682.00 frames. ], tot_loss[loss=3.574, NarTop10Accuracy=0.6107, over 6024.96 frames. ], batch size: 9, lr: 1.33e-02
2024-08-06 15:44:08,620 INFO [trainer.py:765] (4/8) Epoch 6, batch 2400, train_loss[loss=3.364, NarTop10Accuracy=0.6586, over 5031.00 frames. ], tot_loss[loss=3.544, NarTop10Accuracy=0.6167, over 5776.84 frames. ], batch size: 7, lr: 1.32e-02
2024-08-06 15:44:32,132 INFO [trainer.py:765] (4/8) Epoch 6, batch 2500, train_loss[loss=3.39, NarTop10Accuracy=0.6428, over 5307.00 frames. ], tot_loss[loss=3.525, NarTop10Accuracy=0.6206, over 5479.01 frames. ], batch size: 7, lr: 1.32e-02
2024-08-06 15:44:51,677 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 15:45:58,043 INFO [trainer.py:765] (4/8) Epoch 7, batch 100, train_loss[loss=3.302, NarTop10Accuracy=0.6689, over 7098.00 frames. ], tot_loss[loss=3.534, NarTop10Accuracy=0.6189, over 2369.42 frames. ], batch size: 31, lr: 1.24e-02
2024-08-06 15:46:33,614 INFO [trainer.py:765] (4/8) Epoch 7, batch 200, train_loss[loss=3.454, NarTop10Accuracy=0.6284, over 6894.00 frames. ], tot_loss[loss=3.531, NarTop10Accuracy=0.6191, over 3864.21 frames. ], batch size: 17, lr: 1.23e-02
2024-08-06 15:47:03,246 INFO [trainer.py:765] (4/8) Epoch 7, batch 300, train_loss[loss=3.745, NarTop10Accuracy=0.5787, over 7056.00 frames. ], tot_loss[loss=3.541, NarTop10Accuracy=0.6172, over 4659.61 frames. ], batch size: 22, lr: 1.23e-02
2024-08-06 15:47:34,496 INFO [trainer.py:765] (4/8) Epoch 7, batch 400, train_loss[loss=3.526, NarTop10Accuracy=0.6278, over 5037.00 frames. ], tot_loss[loss=3.531, NarTop10Accuracy=0.6191, over 5103.77 frames. ], batch size: 7, lr: 1.23e-02
2024-08-06 15:48:13,730 INFO [trainer.py:765] (4/8) Epoch 7, batch 500, train_loss[loss=3.589, NarTop10Accuracy=0.6064, over 6093.00 frames. ], tot_loss[loss=3.525, NarTop10Accuracy=0.6201, over 5389.87 frames. ], batch size: 11, lr: 1.22e-02
2024-08-06 15:48:26,370 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:48:34,533 INFO [trainer.py:811] (4/8) Epoch 7, validation: loss=3.326, NarTop10Accuracy=0.6612, over 1905321.00 frames. 
2024-08-06 15:48:34,534 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27360MB
2024-08-06 15:48:35,078 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.466e+02 1.860e+02 2.018e+02 2.241e+02 5.111e+02, threshold=4.035e+02, percent-clipped=0.3
2024-08-06 15:48:52,720 INFO [trainer.py:765] (4/8) Epoch 7, batch 600, train_loss[loss=3.163, NarTop10Accuracy=0.6997, over 5622.00 frames. ], tot_loss[loss=3.533, NarTop10Accuracy=0.6188, over 5647.99 frames. ], batch size: 9, lr: 1.22e-02
2024-08-06 15:49:24,912 INFO [trainer.py:765] (4/8) Epoch 7, batch 700, train_loss[loss=3.813, NarTop10Accuracy=0.5544, over 5055.00 frames. ], tot_loss[loss=3.515, NarTop10Accuracy=0.6226, over 5718.13 frames. ], batch size: 6, lr: 1.21e-02
2024-08-06 15:50:04,381 INFO [trainer.py:765] (4/8) Epoch 7, batch 800, train_loss[loss=3.242, NarTop10Accuracy=0.685, over 4950.00 frames. ], tot_loss[loss=3.5, NarTop10Accuracy=0.6255, over 5787.78 frames. ], batch size: 6, lr: 1.21e-02
2024-08-06 15:50:34,549 INFO [trainer.py:765] (4/8) Epoch 7, batch 900, train_loss[loss=3.236, NarTop10Accuracy=0.6882, over 6261.00 frames. ], tot_loss[loss=3.49, NarTop10Accuracy=0.6272, over 5805.94 frames. ], batch size: 13, lr: 1.21e-02
2024-08-06 15:51:07,155 INFO [trainer.py:765] (4/8) Epoch 7, batch 1000, train_loss[loss=3.243, NarTop10Accuracy=0.6864, over 6252.00 frames. ], tot_loss[loss=3.489, NarTop10Accuracy=0.6274, over 5929.27 frames. ], batch size: 13, lr: 1.20e-02
2024-08-06 15:51:51,758 INFO [trainer.py:765] (4/8) Epoch 7, batch 1100, train_loss[loss=3.343, NarTop10Accuracy=0.6605, over 6816.00 frames. ], tot_loss[loss=3.496, NarTop10Accuracy=0.6264, over 5943.36 frames. ], batch size: 17, lr: 1.20e-02
2024-08-06 15:52:22,699 INFO [trainer.py:765] (4/8) Epoch 7, batch 1200, train_loss[loss=3.318, NarTop10Accuracy=0.6643, over 7389.00 frames. ], tot_loss[loss=3.487, NarTop10Accuracy=0.6281, over 5920.44 frames. ], batch size: 31, lr: 1.20e-02
2024-08-06 15:52:52,008 INFO [trainer.py:765] (4/8) Epoch 7, batch 1300, train_loss[loss=3.427, NarTop10Accuracy=0.6299, over 4902.00 frames. ], tot_loss[loss=3.493, NarTop10Accuracy=0.6269, over 5993.35 frames. ], batch size: 6, lr: 1.19e-02
2024-08-06 15:53:33,842 INFO [trainer.py:765] (4/8) Epoch 7, batch 1400, train_loss[loss=3.328, NarTop10Accuracy=0.6573, over 6117.00 frames. ], tot_loss[loss=3.497, NarTop10Accuracy=0.6258, over 5998.92 frames. ], batch size: 11, lr: 1.19e-02
2024-08-06 15:54:04,600 INFO [trainer.py:765] (4/8) Epoch 7, batch 1500, train_loss[loss=3.73, NarTop10Accuracy=0.5772, over 5700.00 frames. ], tot_loss[loss=3.479, NarTop10Accuracy=0.6297, over 5949.36 frames. ], batch size: 50, lr: 1.19e-02
2024-08-06 15:54:32,385 INFO [trainer.py:765] (4/8) Epoch 7, batch 1600, train_loss[loss=3.723, NarTop10Accuracy=0.5778, over 7029.00 frames. ], tot_loss[loss=3.479, NarTop10Accuracy=0.6295, over 5930.55 frames. ], batch size: 22, lr: 1.19e-02
2024-08-06 15:54:59,055 INFO [trainer.py:765] (4/8) Epoch 7, batch 1700, train_loss[loss=3.526, NarTop10Accuracy=0.6198, over 6612.00 frames. ], tot_loss[loss=3.498, NarTop10Accuracy=0.6258, over 5921.29 frames. ], batch size: 14, lr: 1.18e-02
2024-08-06 15:55:25,512 INFO [trainer.py:765] (4/8) Epoch 7, batch 1800, train_loss[loss=3.878, NarTop10Accuracy=0.5465, over 7113.00 frames. ], tot_loss[loss=3.494, NarTop10Accuracy=0.6263, over 5980.11 frames. ], batch size: 22, lr: 1.18e-02
2024-08-06 15:55:52,083 INFO [trainer.py:765] (4/8) Epoch 7, batch 1900, train_loss[loss=3.332, NarTop10Accuracy=0.6628, over 6033.00 frames. ], tot_loss[loss=3.516, NarTop10Accuracy=0.622, over 6024.69 frames. ], batch size: 50, lr: 1.18e-02
2024-08-06 15:56:17,592 INFO [trainer.py:765] (4/8) Epoch 7, batch 2000, train_loss[loss=3.721, NarTop10Accuracy=0.5771, over 5994.00 frames. ], tot_loss[loss=3.502, NarTop10Accuracy=0.6247, over 5997.76 frames. ], batch size: 50, lr: 1.17e-02
2024-08-06 15:56:42,858 INFO [trainer.py:765] (4/8) Epoch 7, batch 2100, train_loss[loss=3.533, NarTop10Accuracy=0.5997, over 3960.00 frames. ], tot_loss[loss=3.483, NarTop10Accuracy=0.6285, over 5976.95 frames. ], batch size: 4, lr: 1.17e-02
2024-08-06 15:57:08,079 INFO [trainer.py:765] (4/8) Epoch 7, batch 2200, train_loss[loss=3.483, NarTop10Accuracy=0.6271, over 7224.00 frames. ], tot_loss[loss=3.505, NarTop10Accuracy=0.6242, over 6011.70 frames. ], batch size: 31, lr: 1.17e-02
2024-08-06 15:57:33,178 INFO [trainer.py:765] (4/8) Epoch 7, batch 2300, train_loss[loss=3.352, NarTop10Accuracy=0.6582, over 5577.00 frames. ], tot_loss[loss=3.512, NarTop10Accuracy=0.6228, over 6022.51 frames. ], batch size: 9, lr: 1.16e-02
2024-08-06 15:57:57,619 INFO [trainer.py:765] (4/8) Epoch 7, batch 2400, train_loss[loss=3.189, NarTop10Accuracy=0.6828, over 5229.00 frames. ], tot_loss[loss=3.49, NarTop10Accuracy=0.6271, over 5773.06 frames. ], batch size: 7, lr: 1.16e-02
2024-08-06 15:58:21,090 INFO [trainer.py:765] (4/8) Epoch 7, batch 2500, train_loss[loss=3.557, NarTop10Accuracy=0.6066, over 5730.00 frames. ], tot_loss[loss=3.469, NarTop10Accuracy=0.6315, over 5479.67 frames. ], batch size: 8, lr: 1.16e-02
2024-08-06 15:58:31,565 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 15:58:39,769 INFO [trainer.py:811] (4/8) Epoch 7, validation: loss=3.381, NarTop10Accuracy=0.6488, over 1905321.00 frames. 
2024-08-06 15:58:39,770 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27360MB
2024-08-06 15:58:40,221 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.471e+02 1.831e+02 1.996e+02 2.207e+02 5.229e+02, threshold=3.992e+02, percent-clipped=0.2
2024-08-06 15:58:49,111 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 15:59:52,877 INFO [trainer.py:765] (4/8) Epoch 8, batch 100, train_loss[loss=3.608, NarTop10Accuracy=0.5939, over 7539.00 frames. ], tot_loss[loss=3.452, NarTop10Accuracy=0.6358, over 2360.25 frames. ], batch size: 32, lr: 1.09e-02
2024-08-06 16:00:27,881 INFO [trainer.py:765] (4/8) Epoch 8, batch 200, train_loss[loss=3.348, NarTop10Accuracy=0.6592, over 6516.00 frames. ], tot_loss[loss=3.475, NarTop10Accuracy=0.6302, over 3852.35 frames. ], batch size: 17, lr: 1.09e-02
2024-08-06 16:00:58,563 INFO [trainer.py:765] (4/8) Epoch 8, batch 300, train_loss[loss=3.295, NarTop10Accuracy=0.6712, over 7278.00 frames. ], tot_loss[loss=3.461, NarTop10Accuracy=0.6335, over 4659.40 frames. ], batch size: 23, lr: 1.08e-02
2024-08-06 16:01:29,760 INFO [trainer.py:765] (4/8) Epoch 8, batch 400, train_loss[loss=3.597, NarTop10Accuracy=0.5971, over 5214.00 frames. ], tot_loss[loss=3.47, NarTop10Accuracy=0.6313, over 5122.70 frames. ], batch size: 7, lr: 1.08e-02
2024-08-06 16:02:04,066 INFO [trainer.py:765] (4/8) Epoch 8, batch 500, train_loss[loss=3.768, NarTop10Accuracy=0.5623, over 6168.00 frames. ], tot_loss[loss=3.458, NarTop10Accuracy=0.6337, over 5385.37 frames. ], batch size: 11, lr: 1.08e-02
2024-08-06 16:02:41,836 INFO [trainer.py:765] (4/8) Epoch 8, batch 600, train_loss[loss=3.129, NarTop10Accuracy=0.7053, over 5760.00 frames. ], tot_loss[loss=3.474, NarTop10Accuracy=0.6303, over 5642.61 frames. ], batch size: 9, lr: 1.08e-02
2024-08-06 16:03:11,500 INFO [trainer.py:765] (4/8) Epoch 8, batch 700, train_loss[loss=3.667, NarTop10Accuracy=0.5858, over 4266.00 frames. ], tot_loss[loss=3.483, NarTop10Accuracy=0.629, over 5702.94 frames. ], batch size: 5, lr: 1.07e-02
2024-08-06 16:03:50,084 INFO [trainer.py:765] (4/8) Epoch 8, batch 800, train_loss[loss=3.336, NarTop10Accuracy=0.6514, over 5154.00 frames. ], tot_loss[loss=3.468, NarTop10Accuracy=0.6318, over 5768.77 frames. ], batch size: 6, lr: 1.07e-02
2024-08-06 16:04:27,588 INFO [trainer.py:765] (4/8) Epoch 8, batch 900, train_loss[loss=3.365, NarTop10Accuracy=0.6597, over 6528.00 frames. ], tot_loss[loss=3.453, NarTop10Accuracy=0.6345, over 5812.51 frames. ], batch size: 14, lr: 1.07e-02
2024-08-06 16:04:57,466 INFO [trainer.py:765] (4/8) Epoch 8, batch 1000, train_loss[loss=3.683, NarTop10Accuracy=0.5814, over 6645.00 frames. ], tot_loss[loss=3.442, NarTop10Accuracy=0.6369, over 5902.06 frames. ], batch size: 14, lr: 1.07e-02
2024-08-06 16:05:37,294 INFO [trainer.py:765] (4/8) Epoch 8, batch 1100, train_loss[loss=3.667, NarTop10Accuracy=0.5928, over 6807.00 frames. ], tot_loss[loss=3.439, NarTop10Accuracy=0.6377, over 5961.87 frames. ], batch size: 17, lr: 1.06e-02
2024-08-06 16:06:15,858 INFO [trainer.py:765] (4/8) Epoch 8, batch 1200, train_loss[loss=3.492, NarTop10Accuracy=0.6243, over 7110.00 frames. ], tot_loss[loss=3.448, NarTop10Accuracy=0.6356, over 5949.20 frames. ], batch size: 31, lr: 1.06e-02
2024-08-06 16:06:45,187 INFO [trainer.py:765] (4/8) Epoch 8, batch 1300, train_loss[loss=3.046, NarTop10Accuracy=0.7258, over 4989.00 frames. ], tot_loss[loss=3.438, NarTop10Accuracy=0.6376, over 6020.72 frames. ], batch size: 6, lr: 1.06e-02
2024-08-06 16:07:24,235 INFO [trainer.py:765] (4/8) Epoch 8, batch 1400, train_loss[loss=3.48, NarTop10Accuracy=0.6204, over 6129.00 frames. ], tot_loss[loss=3.439, NarTop10Accuracy=0.6371, over 6046.38 frames. ], batch size: 11, lr: 1.05e-02
2024-08-06 16:07:52,169 INFO [trainer.py:765] (4/8) Epoch 8, batch 1500, train_loss[loss=3.446, NarTop10Accuracy=0.64, over 5610.00 frames. ], tot_loss[loss=3.418, NarTop10Accuracy=0.6414, over 5981.61 frames. ], batch size: 51, lr: 1.05e-02
2024-08-06 16:08:19,948 INFO [trainer.py:765] (4/8) Epoch 8, batch 1600, train_loss[loss=3.142, NarTop10Accuracy=0.6966, over 7089.00 frames. ], tot_loss[loss=3.417, NarTop10Accuracy=0.6421, over 5949.01 frames. ], batch size: 22, lr: 1.05e-02
2024-08-06 16:08:46,617 INFO [trainer.py:765] (4/8) Epoch 8, batch 1700, train_loss[loss=3.275, NarTop10Accuracy=0.6789, over 6354.00 frames. ], tot_loss[loss=3.421, NarTop10Accuracy=0.6413, over 5941.09 frames. ], batch size: 13, lr: 1.05e-02
2024-08-06 16:09:13,106 INFO [trainer.py:765] (4/8) Epoch 8, batch 1800, train_loss[loss=3.279, NarTop10Accuracy=0.6754, over 7173.00 frames. ], tot_loss[loss=3.416, NarTop10Accuracy=0.6424, over 5987.29 frames. ], batch size: 22, lr: 1.04e-02
2024-08-06 16:09:39,636 INFO [trainer.py:765] (4/8) Epoch 8, batch 1900, train_loss[loss=3.818, NarTop10Accuracy=0.5566, over 6060.00 frames. ], tot_loss[loss=3.404, NarTop10Accuracy=0.6448, over 6032.80 frames. ], batch size: 50, lr: 1.04e-02
2024-08-06 16:09:56,940 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 16:10:04,970 INFO [trainer.py:811] (4/8) Epoch 8, validation: loss=3.282, NarTop10Accuracy=0.6699, over 1905321.00 frames. 
2024-08-06 16:10:04,970 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27360MB
2024-08-06 16:10:05,469 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.411e+02 1.814e+02 1.981e+02 2.158e+02 5.862e+02, threshold=3.962e+02, percent-clipped=0.1
2024-08-06 16:10:13,202 INFO [trainer.py:765] (4/8) Epoch 8, batch 2000, train_loss[loss=3.933, NarTop10Accuracy=0.535, over 6426.00 frames. ], tot_loss[loss=3.413, NarTop10Accuracy=0.643, over 6001.70 frames. ], batch size: 50, lr: 1.04e-02
2024-08-06 16:10:38,513 INFO [trainer.py:765] (4/8) Epoch 8, batch 2100, train_loss[loss=3.379, NarTop10Accuracy=0.6465, over 4743.00 frames. ], tot_loss[loss=3.406, NarTop10Accuracy=0.6443, over 5971.79 frames. ], batch size: 5, lr: 1.04e-02
2024-08-06 16:11:03,746 INFO [trainer.py:765] (4/8) Epoch 8, batch 2200, train_loss[loss=3.579, NarTop10Accuracy=0.6122, over 7386.00 frames. ], tot_loss[loss=3.418, NarTop10Accuracy=0.6419, over 6010.62 frames. ], batch size: 31, lr: 1.04e-02
2024-08-06 16:11:28,903 INFO [trainer.py:765] (4/8) Epoch 8, batch 2300, train_loss[loss=3.611, NarTop10Accuracy=0.6017, over 5751.00 frames. ], tot_loss[loss=3.434, NarTop10Accuracy=0.6385, over 6015.58 frames. ], batch size: 9, lr: 1.03e-02
2024-08-06 16:11:53,091 INFO [trainer.py:765] (4/8) Epoch 8, batch 2400, train_loss[loss=3.578, NarTop10Accuracy=0.6012, over 5319.00 frames. ], tot_loss[loss=3.422, NarTop10Accuracy=0.6408, over 5756.32 frames. ], batch size: 7, lr: 1.03e-02
2024-08-06 16:12:16,443 INFO [trainer.py:765] (4/8) Epoch 8, batch 2500, train_loss[loss=3.224, NarTop10Accuracy=0.6737, over 5007.00 frames. ], tot_loss[loss=3.415, NarTop10Accuracy=0.6417, over 5453.29 frames. ], batch size: 7, lr: 1.03e-02
2024-08-06 16:12:36,389 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 16:13:37,514 INFO [trainer.py:765] (4/8) Epoch 9, batch 100, train_loss[loss=3.26, NarTop10Accuracy=0.6813, over 7260.00 frames. ], tot_loss[loss=3.372, NarTop10Accuracy=0.6521, over 2365.02 frames. ], batch size: 31, lr: 9.72e-03
2024-08-06 16:14:14,440 INFO [trainer.py:765] (4/8) Epoch 9, batch 200, train_loss[loss=3.649, NarTop10Accuracy=0.5968, over 6801.00 frames. ], tot_loss[loss=3.364, NarTop10Accuracy=0.6536, over 3838.75 frames. ], batch size: 17, lr: 9.70e-03
2024-08-06 16:14:44,507 INFO [trainer.py:765] (4/8) Epoch 9, batch 300, train_loss[loss=3.378, NarTop10Accuracy=0.6491, over 7074.00 frames. ], tot_loss[loss=3.377, NarTop10Accuracy=0.6506, over 4641.76 frames. ], batch size: 22, lr: 9.68e-03
2024-08-06 16:15:14,914 INFO [trainer.py:765] (4/8) Epoch 9, batch 400, train_loss[loss=3.22, NarTop10Accuracy=0.6892, over 5211.00 frames. ], tot_loss[loss=3.361, NarTop10Accuracy=0.6539, over 5091.87 frames. ], batch size: 7, lr: 9.65e-03
2024-08-06 16:15:50,336 INFO [trainer.py:765] (4/8) Epoch 9, batch 500, train_loss[loss=3.362, NarTop10Accuracy=0.6591, over 6057.00 frames. ], tot_loss[loss=3.346, NarTop10Accuracy=0.6569, over 5387.15 frames. ], batch size: 11, lr: 9.63e-03
2024-08-06 16:16:23,972 INFO [trainer.py:765] (4/8) Epoch 9, batch 600, train_loss[loss=3.592, NarTop10Accuracy=0.6058, over 5691.00 frames. ], tot_loss[loss=3.345, NarTop10Accuracy=0.6574, over 5652.09 frames. ], batch size: 9, lr: 9.61e-03
2024-08-06 16:16:57,145 INFO [trainer.py:765] (4/8) Epoch 9, batch 700, train_loss[loss=3.293, NarTop10Accuracy=0.6705, over 4230.00 frames. ], tot_loss[loss=3.355, NarTop10Accuracy=0.6547, over 5695.00 frames. ], batch size: 5, lr: 9.59e-03
2024-08-06 16:17:32,052 INFO [trainer.py:765] (4/8) Epoch 9, batch 800, train_loss[loss=3.211, NarTop10Accuracy=0.6788, over 5052.00 frames. ], tot_loss[loss=3.39, NarTop10Accuracy=0.6475, over 5757.28 frames. ], batch size: 6, lr: 9.57e-03
2024-08-06 16:18:07,816 INFO [trainer.py:765] (4/8) Epoch 9, batch 900, train_loss[loss=3.16, NarTop10Accuracy=0.7036, over 6189.00 frames. ], tot_loss[loss=3.382, NarTop10Accuracy=0.6492, over 5799.33 frames. ], batch size: 13, lr: 9.55e-03
2024-08-06 16:18:39,345 INFO [trainer.py:765] (4/8) Epoch 9, batch 1000, train_loss[loss=3.084, NarTop10Accuracy=0.706, over 6255.00 frames. ], tot_loss[loss=3.391, NarTop10Accuracy=0.6473, over 5906.37 frames. ], batch size: 13, lr: 9.53e-03
2024-08-06 16:19:15,382 INFO [trainer.py:765] (4/8) Epoch 9, batch 1100, train_loss[loss=3.382, NarTop10Accuracy=0.6534, over 6846.00 frames. ], tot_loss[loss=3.393, NarTop10Accuracy=0.6466, over 5945.57 frames. ], batch size: 17, lr: 9.50e-03
2024-08-06 16:19:53,877 INFO [trainer.py:765] (4/8) Epoch 9, batch 1200, train_loss[loss=3.756, NarTop10Accuracy=0.564, over 7413.00 frames. ], tot_loss[loss=3.4, NarTop10Accuracy=0.6451, over 5961.66 frames. ], batch size: 32, lr: 9.48e-03
2024-08-06 16:20:24,906 INFO [trainer.py:765] (4/8) Epoch 9, batch 1300, train_loss[loss=3.098, NarTop10Accuracy=0.7073, over 4299.00 frames. ], tot_loss[loss=3.399, NarTop10Accuracy=0.6455, over 6020.70 frames. ], batch size: 5, lr: 9.46e-03
2024-08-06 16:20:56,580 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 16:21:04,483 INFO [trainer.py:811] (4/8) Epoch 9, validation: loss=3.266, NarTop10Accuracy=0.6725, over 1905321.00 frames. 
2024-08-06 16:21:04,484 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 27360MB
2024-08-06 16:21:05,035 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.473e+02 1.808e+02 1.967e+02 2.142e+02 6.126e+02, threshold=3.935e+02, percent-clipped=0.5
2024-08-06 16:21:06,691 INFO [trainer.py:765] (4/8) Epoch 9, batch 1400, train_loss[loss=3.685, NarTop10Accuracy=0.5834, over 6177.00 frames. ], tot_loss[loss=3.41, NarTop10Accuracy=0.643, over 6021.58 frames. ], batch size: 11, lr: 9.44e-03
2024-08-06 16:21:38,897 INFO [trainer.py:765] (4/8) Epoch 9, batch 1500, train_loss[loss=3.371, NarTop10Accuracy=0.6492, over 6327.00 frames. ], tot_loss[loss=3.387, NarTop10Accuracy=0.6477, over 5953.27 frames. ], batch size: 51, lr: 9.42e-03
2024-08-06 16:22:06,721 INFO [trainer.py:765] (4/8) Epoch 9, batch 1600, train_loss[loss=3.284, NarTop10Accuracy=0.6712, over 7152.00 frames. ], tot_loss[loss=3.378, NarTop10Accuracy=0.6503, over 5929.63 frames. ], batch size: 22, lr: 9.40e-03
2024-08-06 16:22:33,470 INFO [trainer.py:765] (4/8) Epoch 9, batch 1700, train_loss[loss=3.533, NarTop10Accuracy=0.6268, over 6123.00 frames. ], tot_loss[loss=3.393, NarTop10Accuracy=0.6473, over 5908.94 frames. ], batch size: 13, lr: 9.38e-03
2024-08-06 16:23:00,063 INFO [trainer.py:765] (4/8) Epoch 9, batch 1800, train_loss[loss=3.19, NarTop10Accuracy=0.6883, over 7350.00 frames. ], tot_loss[loss=3.377, NarTop10Accuracy=0.6506, over 5980.15 frames. ], batch size: 23, lr: 9.36e-03
2024-08-06 16:23:26,782 INFO [trainer.py:765] (4/8) Epoch 9, batch 1900, train_loss[loss=3.494, NarTop10Accuracy=0.6373, over 6183.00 frames. ], tot_loss[loss=3.387, NarTop10Accuracy=0.6485, over 6030.45 frames. ], batch size: 50, lr: 9.34e-03
2024-08-06 16:23:52,485 INFO [trainer.py:765] (4/8) Epoch 9, batch 2000, train_loss[loss=3.976, NarTop10Accuracy=0.5262, over 5928.00 frames. ], tot_loss[loss=3.385, NarTop10Accuracy=0.6485, over 5997.29 frames. ], batch size: 50, lr: 9.32e-03
2024-08-06 16:24:17,962 INFO [trainer.py:765] (4/8) Epoch 9, batch 2100, train_loss[loss=2.885, NarTop10Accuracy=0.7434, over 3963.00 frames. ], tot_loss[loss=3.386, NarTop10Accuracy=0.6487, over 5981.65 frames. ], batch size: 4, lr: 9.30e-03
2024-08-06 16:24:43,421 INFO [trainer.py:765] (4/8) Epoch 9, batch 2200, train_loss[loss=3.751, NarTop10Accuracy=0.576, over 7326.00 frames. ], tot_loss[loss=3.399, NarTop10Accuracy=0.6462, over 6015.27 frames. ], batch size: 32, lr: 9.28e-03
2024-08-06 16:25:08,720 INFO [trainer.py:765] (4/8) Epoch 9, batch 2300, train_loss[loss=3.36, NarTop10Accuracy=0.6486, over 5706.00 frames. ], tot_loss[loss=3.409, NarTop10Accuracy=0.6437, over 6013.74 frames. ], batch size: 9, lr: 9.26e-03
2024-08-06 16:25:33,164 INFO [trainer.py:765] (4/8) Epoch 9, batch 2400, train_loss[loss=3.195, NarTop10Accuracy=0.6859, over 5118.00 frames. ], tot_loss[loss=3.403, NarTop10Accuracy=0.6444, over 5762.98 frames. ], batch size: 7, lr: 9.25e-03
2024-08-06 16:25:56,768 INFO [trainer.py:765] (4/8) Epoch 9, batch 2500, train_loss[loss=3.033, NarTop10Accuracy=0.7284, over 5091.00 frames. ], tot_loss[loss=3.366, NarTop10Accuracy=0.652, over 5463.80 frames. ], batch size: 7, lr: 9.23e-03
2024-08-06 16:26:16,448 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 16:27:19,583 INFO [trainer.py:765] (4/8) Epoch 10, batch 100, train_loss[loss=3.274, NarTop10Accuracy=0.6762, over 7248.00 frames. ], tot_loss[loss=3.36, NarTop10Accuracy=0.6545, over 2364.57 frames. ], batch size: 31, lr: 8.76e-03
2024-08-06 16:27:52,627 INFO [trainer.py:765] (4/8) Epoch 10, batch 200, train_loss[loss=3.159, NarTop10Accuracy=0.6931, over 6849.00 frames. ], tot_loss[loss=3.35, NarTop10Accuracy=0.6565, over 3851.29 frames. ], batch size: 17, lr: 8.74e-03
2024-08-06 16:28:23,057 INFO [trainer.py:765] (4/8) Epoch 10, batch 300, train_loss[loss=3.138, NarTop10Accuracy=0.6978, over 7173.00 frames. ], tot_loss[loss=3.352, NarTop10Accuracy=0.6559, over 4655.95 frames. ], batch size: 22, lr: 8.72e-03
2024-08-06 16:28:59,200 INFO [trainer.py:765] (4/8) Epoch 10, batch 400, train_loss[loss=3.14, NarTop10Accuracy=0.6962, over 5199.00 frames. ], tot_loss[loss=3.345, NarTop10Accuracy=0.6573, over 5095.18 frames. ], batch size: 7, lr: 8.71e-03
2024-08-06 16:29:29,217 INFO [trainer.py:765] (4/8) Epoch 10, batch 500, train_loss[loss=3.018, NarTop10Accuracy=0.7181, over 6111.00 frames. ], tot_loss[loss=3.335, NarTop10Accuracy=0.6591, over 5382.80 frames. ], batch size: 11, lr: 8.69e-03
2024-08-06 16:30:02,765 INFO [trainer.py:765] (4/8) Epoch 10, batch 600, train_loss[loss=3.515, NarTop10Accuracy=0.6181, over 5859.00 frames. ], tot_loss[loss=3.349, NarTop10Accuracy=0.6559, over 5641.27 frames. ], batch size: 9, lr: 8.67e-03
2024-08-06 16:30:34,264 INFO [trainer.py:765] (4/8) Epoch 10, batch 700, train_loss[loss=3.394, NarTop10Accuracy=0.6501, over 5268.00 frames. ], tot_loss[loss=3.353, NarTop10Accuracy=0.655, over 5707.09 frames. ], batch size: 6, lr: 8.65e-03
2024-08-06 16:31:09,842 INFO [trainer.py:765] (4/8) Epoch 10, batch 800, train_loss[loss=3.52, NarTop10Accuracy=0.6095, over 5091.00 frames. ], tot_loss[loss=3.358, NarTop10Accuracy=0.6541, over 5778.37 frames. ], batch size: 6, lr: 8.64e-03
2024-08-06 16:31:16,257 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 16:31:24,565 INFO [trainer.py:811] (4/8) Epoch 10, validation: loss=3.184, NarTop10Accuracy=0.6898, over 1905321.00 frames. 
2024-08-06 16:31:24,566 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 16:31:25,154 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.434e+02 1.851e+02 2.012e+02 2.196e+02 4.599e+02, threshold=4.024e+02, percent-clipped=0.1
2024-08-06 16:31:50,345 INFO [trainer.py:765] (4/8) Epoch 10, batch 900, train_loss[loss=3.236, NarTop10Accuracy=0.6742, over 6612.00 frames. ], tot_loss[loss=3.334, NarTop10Accuracy=0.6589, over 5809.42 frames. ], batch size: 14, lr: 8.62e-03
2024-08-06 16:32:28,589 INFO [trainer.py:765] (4/8) Epoch 10, batch 1000, train_loss[loss=3.08, NarTop10Accuracy=0.7201, over 6165.00 frames. ], tot_loss[loss=3.341, NarTop10Accuracy=0.6576, over 5913.22 frames. ], batch size: 13, lr: 8.60e-03
2024-08-06 16:33:06,376 INFO [trainer.py:765] (4/8) Epoch 10, batch 1100, train_loss[loss=3.086, NarTop10Accuracy=0.7075, over 6756.00 frames. ], tot_loss[loss=3.348, NarTop10Accuracy=0.6563, over 5936.98 frames. ], batch size: 17, lr: 8.59e-03
2024-08-06 16:33:40,960 INFO [trainer.py:765] (4/8) Epoch 10, batch 1200, train_loss[loss=3.236, NarTop10Accuracy=0.6684, over 7071.00 frames. ], tot_loss[loss=3.336, NarTop10Accuracy=0.6584, over 5930.66 frames. ], batch size: 31, lr: 8.57e-03
2024-08-06 16:34:16,170 INFO [trainer.py:765] (4/8) Epoch 10, batch 1300, train_loss[loss=3.128, NarTop10Accuracy=0.7012, over 5175.00 frames. ], tot_loss[loss=3.342, NarTop10Accuracy=0.6571, over 6019.15 frames. ], batch size: 6, lr: 8.55e-03
2024-08-06 16:34:51,201 INFO [trainer.py:765] (4/8) Epoch 10, batch 1400, train_loss[loss=3.442, NarTop10Accuracy=0.6376, over 6297.00 frames. ], tot_loss[loss=3.366, NarTop10Accuracy=0.6521, over 6044.74 frames. ], batch size: 11, lr: 8.54e-03
2024-08-06 16:35:22,159 INFO [trainer.py:765] (4/8) Epoch 10, batch 1500, train_loss[loss=3.599, NarTop10Accuracy=0.6024, over 6129.00 frames. ], tot_loss[loss=3.349, NarTop10Accuracy=0.6559, over 5964.94 frames. ], batch size: 50, lr: 8.52e-03
2024-08-06 16:35:50,137 INFO [trainer.py:765] (4/8) Epoch 10, batch 1600, train_loss[loss=3.667, NarTop10Accuracy=0.5881, over 7062.00 frames. ], tot_loss[loss=3.334, NarTop10Accuracy=0.6592, over 5921.30 frames. ], batch size: 22, lr: 8.50e-03
2024-08-06 16:36:16,976 INFO [trainer.py:765] (4/8) Epoch 10, batch 1700, train_loss[loss=3.392, NarTop10Accuracy=0.646, over 6639.00 frames. ], tot_loss[loss=3.34, NarTop10Accuracy=0.6577, over 5912.08 frames. ], batch size: 14, lr: 8.49e-03
2024-08-06 16:36:43,648 INFO [trainer.py:765] (4/8) Epoch 10, batch 1800, train_loss[loss=3.248, NarTop10Accuracy=0.6774, over 7161.00 frames. ], tot_loss[loss=3.328, NarTop10Accuracy=0.6603, over 5971.53 frames. ], batch size: 22, lr: 8.47e-03
2024-08-06 16:37:10,290 INFO [trainer.py:765] (4/8) Epoch 10, batch 1900, train_loss[loss=3.306, NarTop10Accuracy=0.6713, over 6192.00 frames. ], tot_loss[loss=3.331, NarTop10Accuracy=0.6602, over 6018.67 frames. ], batch size: 52, lr: 8.45e-03
2024-08-06 16:37:36,089 INFO [trainer.py:765] (4/8) Epoch 10, batch 2000, train_loss[loss=3.248, NarTop10Accuracy=0.6819, over 6879.00 frames. ], tot_loss[loss=3.324, NarTop10Accuracy=0.6611, over 5993.33 frames. ], batch size: 50, lr: 8.44e-03
2024-08-06 16:38:01,650 INFO [trainer.py:765] (4/8) Epoch 10, batch 2100, train_loss[loss=3.326, NarTop10Accuracy=0.6524, over 3909.00 frames. ], tot_loss[loss=3.337, NarTop10Accuracy=0.6581, over 5978.90 frames. ], batch size: 4, lr: 8.42e-03
2024-08-06 16:38:27,120 INFO [trainer.py:765] (4/8) Epoch 10, batch 2200, train_loss[loss=3.812, NarTop10Accuracy=0.5626, over 7455.00 frames. ], tot_loss[loss=3.343, NarTop10Accuracy=0.657, over 6009.80 frames. ], batch size: 31, lr: 8.41e-03
2024-08-06 16:38:52,447 INFO [trainer.py:765] (4/8) Epoch 10, batch 2300, train_loss[loss=3.212, NarTop10Accuracy=0.6822, over 5628.00 frames. ], tot_loss[loss=3.346, NarTop10Accuracy=0.6567, over 6020.22 frames. ], batch size: 9, lr: 8.39e-03
2024-08-06 16:39:17,006 INFO [trainer.py:765] (4/8) Epoch 10, batch 2400, train_loss[loss=3.156, NarTop10Accuracy=0.6912, over 5199.00 frames. ], tot_loss[loss=3.318, NarTop10Accuracy=0.6624, over 5779.30 frames. ], batch size: 7, lr: 8.37e-03
2024-08-06 16:39:40,802 INFO [trainer.py:765] (4/8) Epoch 10, batch 2500, train_loss[loss=3.592, NarTop10Accuracy=0.6009, over 5238.00 frames. ], tot_loss[loss=3.293, NarTop10Accuracy=0.6672, over 5470.90 frames. ], batch size: 7, lr: 8.36e-03
2024-08-06 16:40:00,620 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 16:41:06,234 INFO [trainer.py:765] (4/8) Epoch 11, batch 100, train_loss[loss=3.661, NarTop10Accuracy=0.5899, over 7182.00 frames. ], tot_loss[loss=3.383, NarTop10Accuracy=0.65, over 2369.52 frames. ], batch size: 31, lr: 7.97e-03
2024-08-06 16:41:39,021 INFO [trainer.py:765] (4/8) Epoch 11, batch 200, train_loss[loss=3.563, NarTop10Accuracy=0.6075, over 6864.00 frames. ], tot_loss[loss=3.337, NarTop10Accuracy=0.6587, over 3851.77 frames. ], batch size: 17, lr: 7.95e-03
2024-08-06 16:41:53,189 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 16:42:01,355 INFO [trainer.py:811] (4/8) Epoch 11, validation: loss=3.116, NarTop10Accuracy=0.7034, over 1905321.00 frames. 
2024-08-06 16:42:01,356 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 16:42:01,879 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.526e+02 1.889e+02 2.046e+02 2.249e+02 5.417e+02, threshold=4.093e+02, percent-clipped=0.2
2024-08-06 16:42:17,974 INFO [trainer.py:765] (4/8) Epoch 11, batch 300, train_loss[loss=2.994, NarTop10Accuracy=0.7331, over 7140.00 frames. ], tot_loss[loss=3.303, NarTop10Accuracy=0.6657, over 4651.77 frames. ], batch size: 22, lr: 7.94e-03
2024-08-06 16:42:55,153 INFO [trainer.py:765] (4/8) Epoch 11, batch 400, train_loss[loss=3.435, NarTop10Accuracy=0.6376, over 5148.00 frames. ], tot_loss[loss=3.298, NarTop10Accuracy=0.6667, over 5093.41 frames. ], batch size: 7, lr: 7.92e-03
2024-08-06 16:43:25,718 INFO [trainer.py:765] (4/8) Epoch 11, batch 500, train_loss[loss=3.124, NarTop10Accuracy=0.6988, over 6051.00 frames. ], tot_loss[loss=3.285, NarTop10Accuracy=0.669, over 5380.48 frames. ], batch size: 11, lr: 7.91e-03
2024-08-06 16:44:02,241 INFO [trainer.py:765] (4/8) Epoch 11, batch 600, train_loss[loss=3.499, NarTop10Accuracy=0.6359, over 5751.00 frames. ], tot_loss[loss=3.297, NarTop10Accuracy=0.666, over 5646.10 frames. ], batch size: 9, lr: 7.89e-03
2024-08-06 16:44:35,715 INFO [trainer.py:765] (4/8) Epoch 11, batch 700, train_loss[loss=3.41, NarTop10Accuracy=0.6252, over 5235.00 frames. ], tot_loss[loss=3.291, NarTop10Accuracy=0.6673, over 5718.85 frames. ], batch size: 6, lr: 7.88e-03
2024-08-06 16:45:10,467 INFO [trainer.py:765] (4/8) Epoch 11, batch 800, train_loss[loss=2.929, NarTop10Accuracy=0.745, over 5178.00 frames. ], tot_loss[loss=3.309, NarTop10Accuracy=0.6635, over 5790.47 frames. ], batch size: 6, lr: 7.86e-03
2024-08-06 16:45:46,457 INFO [trainer.py:765] (4/8) Epoch 11, batch 900, train_loss[loss=3.728, NarTop10Accuracy=0.5812, over 6297.00 frames. ], tot_loss[loss=3.306, NarTop10Accuracy=0.664, over 5823.55 frames. ], batch size: 13, lr: 7.85e-03
2024-08-06 16:46:20,309 INFO [trainer.py:765] (4/8) Epoch 11, batch 1000, train_loss[loss=3.328, NarTop10Accuracy=0.6506, over 6273.00 frames. ], tot_loss[loss=3.306, NarTop10Accuracy=0.6643, over 5928.39 frames. ], batch size: 13, lr: 7.84e-03
2024-08-06 16:46:53,457 INFO [trainer.py:765] (4/8) Epoch 11, batch 1100, train_loss[loss=3.125, NarTop10Accuracy=0.7089, over 7140.00 frames. ], tot_loss[loss=3.296, NarTop10Accuracy=0.6667, over 5949.22 frames. ], batch size: 18, lr: 7.82e-03
2024-08-06 16:47:33,029 INFO [trainer.py:765] (4/8) Epoch 11, batch 1200, train_loss[loss=3.452, NarTop10Accuracy=0.6337, over 7434.00 frames. ], tot_loss[loss=3.305, NarTop10Accuracy=0.6651, over 5923.05 frames. ], batch size: 33, lr: 7.81e-03
2024-08-06 16:48:06,481 INFO [trainer.py:765] (4/8) Epoch 11, batch 1300, train_loss[loss=2.845, NarTop10Accuracy=0.7432, over 4299.00 frames. ], tot_loss[loss=3.31, NarTop10Accuracy=0.6636, over 5977.42 frames. ], batch size: 5, lr: 7.79e-03
2024-08-06 16:48:41,353 INFO [trainer.py:765] (4/8) Epoch 11, batch 1400, train_loss[loss=3.614, NarTop10Accuracy=0.594, over 6051.00 frames. ], tot_loss[loss=3.335, NarTop10Accuracy=0.6582, over 5990.14 frames. ], batch size: 11, lr: 7.78e-03
2024-08-06 16:49:09,344 INFO [trainer.py:765] (4/8) Epoch 11, batch 1500, train_loss[loss=3.295, NarTop10Accuracy=0.6654, over 5799.00 frames. ], tot_loss[loss=3.331, NarTop10Accuracy=0.6589, over 5935.73 frames. ], batch size: 50, lr: 7.77e-03
2024-08-06 16:49:37,102 INFO [trainer.py:765] (4/8) Epoch 11, batch 1600, train_loss[loss=3.311, NarTop10Accuracy=0.6613, over 7098.00 frames. ], tot_loss[loss=3.31, NarTop10Accuracy=0.6631, over 5897.57 frames. ], batch size: 22, lr: 7.75e-03
2024-08-06 16:50:03,791 INFO [trainer.py:765] (4/8) Epoch 11, batch 1700, train_loss[loss=3.411, NarTop10Accuracy=0.6398, over 6594.00 frames. ], tot_loss[loss=3.301, NarTop10Accuracy=0.665, over 5900.04 frames. ], batch size: 14, lr: 7.74e-03
2024-08-06 16:50:30,352 INFO [trainer.py:765] (4/8) Epoch 11, batch 1800, train_loss[loss=3.333, NarTop10Accuracy=0.6513, over 7035.00 frames. ], tot_loss[loss=3.324, NarTop10Accuracy=0.6608, over 5970.09 frames. ], batch size: 22, lr: 7.72e-03
2024-08-06 16:50:56,821 INFO [trainer.py:765] (4/8) Epoch 11, batch 1900, train_loss[loss=3.792, NarTop10Accuracy=0.5626, over 5595.00 frames. ], tot_loss[loss=3.327, NarTop10Accuracy=0.66, over 6015.78 frames. ], batch size: 50, lr: 7.71e-03
2024-08-06 16:51:22,404 INFO [trainer.py:765] (4/8) Epoch 11, batch 2000, train_loss[loss=3.835, NarTop10Accuracy=0.5573, over 5982.00 frames. ], tot_loss[loss=3.318, NarTop10Accuracy=0.6617, over 5992.19 frames. ], batch size: 50, lr: 7.70e-03
2024-08-06 16:51:47,793 INFO [trainer.py:765] (4/8) Epoch 11, batch 2100, train_loss[loss=3.036, NarTop10Accuracy=0.7207, over 3807.00 frames. ], tot_loss[loss=3.314, NarTop10Accuracy=0.6625, over 5968.70 frames. ], batch size: 4, lr: 7.68e-03
2024-08-06 16:52:13,117 INFO [trainer.py:765] (4/8) Epoch 11, batch 2200, train_loss[loss=3.298, NarTop10Accuracy=0.6622, over 7029.00 frames. ], tot_loss[loss=3.313, NarTop10Accuracy=0.6629, over 6008.66 frames. ], batch size: 31, lr: 7.67e-03
2024-08-06 16:52:23,898 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 16:52:32,079 INFO [trainer.py:811] (4/8) Epoch 11, validation: loss=3.101, NarTop10Accuracy=0.7058, over 1905321.00 frames. 
2024-08-06 16:52:32,080 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 16:52:32,593 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.491e+02 1.920e+02 2.088e+02 2.244e+02 3.599e+02, threshold=4.177e+02, percent-clipped=0.0
2024-08-06 16:52:46,444 INFO [trainer.py:765] (4/8) Epoch 11, batch 2300, train_loss[loss=3.197, NarTop10Accuracy=0.6923, over 5691.00 frames. ], tot_loss[loss=3.326, NarTop10Accuracy=0.6606, over 6015.51 frames. ], batch size: 9, lr: 7.66e-03
2024-08-06 16:53:10,886 INFO [trainer.py:765] (4/8) Epoch 11, batch 2400, train_loss[loss=3.427, NarTop10Accuracy=0.6447, over 5118.00 frames. ], tot_loss[loss=3.308, NarTop10Accuracy=0.6643, over 5777.54 frames. ], batch size: 7, lr: 7.64e-03
2024-08-06 16:53:34,371 INFO [trainer.py:765] (4/8) Epoch 11, batch 2500, train_loss[loss=3.611, NarTop10Accuracy=0.5951, over 5166.00 frames. ], tot_loss[loss=3.295, NarTop10Accuracy=0.6663, over 5480.54 frames. ], batch size: 7, lr: 7.63e-03
2024-08-06 16:53:54,150 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 16:54:58,524 INFO [trainer.py:765] (4/8) Epoch 12, batch 100, train_loss[loss=3.605, NarTop10Accuracy=0.5995, over 7194.00 frames. ], tot_loss[loss=3.302, NarTop10Accuracy=0.6656, over 2362.90 frames. ], batch size: 31, lr: 7.30e-03
2024-08-06 16:55:32,431 INFO [trainer.py:765] (4/8) Epoch 12, batch 200, train_loss[loss=3.145, NarTop10Accuracy=0.6986, over 6855.00 frames. ], tot_loss[loss=3.277, NarTop10Accuracy=0.6709, over 3849.18 frames. ], batch size: 17, lr: 7.29e-03
2024-08-06 16:56:05,095 INFO [trainer.py:765] (4/8) Epoch 12, batch 300, train_loss[loss=3.089, NarTop10Accuracy=0.7063, over 6996.00 frames. ], tot_loss[loss=3.252, NarTop10Accuracy=0.6762, over 4657.96 frames. ], batch size: 22, lr: 7.27e-03
2024-08-06 16:56:36,425 INFO [trainer.py:765] (4/8) Epoch 12, batch 400, train_loss[loss=3.164, NarTop10Accuracy=0.6936, over 5052.00 frames. ], tot_loss[loss=3.26, NarTop10Accuracy=0.6742, over 5118.92 frames. ], batch size: 7, lr: 7.26e-03
2024-08-06 16:57:10,502 INFO [trainer.py:765] (4/8) Epoch 12, batch 500, train_loss[loss=3.64, NarTop10Accuracy=0.5963, over 6174.00 frames. ], tot_loss[loss=3.274, NarTop10Accuracy=0.6713, over 5419.69 frames. ], batch size: 11, lr: 7.25e-03
2024-08-06 16:57:45,483 INFO [trainer.py:765] (4/8) Epoch 12, batch 600, train_loss[loss=2.874, NarTop10Accuracy=0.7543, over 5610.00 frames. ], tot_loss[loss=3.272, NarTop10Accuracy=0.6719, over 5667.99 frames. ], batch size: 9, lr: 7.24e-03
2024-08-06 16:58:17,004 INFO [trainer.py:765] (4/8) Epoch 12, batch 700, train_loss[loss=3.461, NarTop10Accuracy=0.626, over 4350.00 frames. ], tot_loss[loss=3.288, NarTop10Accuracy=0.6686, over 5729.80 frames. ], batch size: 5, lr: 7.22e-03
2024-08-06 16:58:53,468 INFO [trainer.py:765] (4/8) Epoch 12, batch 800, train_loss[loss=3.303, NarTop10Accuracy=0.6664, over 4995.00 frames. ], tot_loss[loss=3.286, NarTop10Accuracy=0.6682, over 5770.36 frames. ], batch size: 6, lr: 7.21e-03
2024-08-06 16:59:27,205 INFO [trainer.py:765] (4/8) Epoch 12, batch 900, train_loss[loss=3.212, NarTop10Accuracy=0.6876, over 6150.00 frames. ], tot_loss[loss=3.267, NarTop10Accuracy=0.6724, over 5794.17 frames. ], batch size: 13, lr: 7.20e-03
2024-08-06 17:00:01,573 INFO [trainer.py:765] (4/8) Epoch 12, batch 1000, train_loss[loss=3.007, NarTop10Accuracy=0.7182, over 6666.00 frames. ], tot_loss[loss=3.282, NarTop10Accuracy=0.6694, over 5890.75 frames. ], batch size: 14, lr: 7.19e-03
2024-08-06 17:00:39,188 INFO [trainer.py:765] (4/8) Epoch 12, batch 1100, train_loss[loss=3.496, NarTop10Accuracy=0.63, over 7035.00 frames. ], tot_loss[loss=3.301, NarTop10Accuracy=0.6653, over 5916.01 frames. ], batch size: 17, lr: 7.18e-03
2024-08-06 17:01:13,963 INFO [trainer.py:765] (4/8) Epoch 12, batch 1200, train_loss[loss=3.184, NarTop10Accuracy=0.6957, over 7302.00 frames. ], tot_loss[loss=3.27, NarTop10Accuracy=0.6715, over 5921.75 frames. ], batch size: 31, lr: 7.17e-03
2024-08-06 17:01:48,107 INFO [trainer.py:765] (4/8) Epoch 12, batch 1300, train_loss[loss=3.398, NarTop10Accuracy=0.6379, over 4179.00 frames. ], tot_loss[loss=3.283, NarTop10Accuracy=0.6688, over 5980.03 frames. ], batch size: 5, lr: 7.15e-03
2024-08-06 17:02:22,322 INFO [trainer.py:765] (4/8) Epoch 12, batch 1400, train_loss[loss=3.649, NarTop10Accuracy=0.6013, over 6081.00 frames. ], tot_loss[loss=3.288, NarTop10Accuracy=0.6683, over 6018.24 frames. ], batch size: 11, lr: 7.14e-03
2024-08-06 17:02:52,877 INFO [trainer.py:765] (4/8) Epoch 12, batch 1500, train_loss[loss=3.392, NarTop10Accuracy=0.6556, over 5757.00 frames. ], tot_loss[loss=3.264, NarTop10Accuracy=0.673, over 5952.44 frames. ], batch size: 50, lr: 7.13e-03
2024-08-06 17:03:20,690 INFO [trainer.py:765] (4/8) Epoch 12, batch 1600, train_loss[loss=3.228, NarTop10Accuracy=0.6778, over 7095.00 frames. ], tot_loss[loss=3.283, NarTop10Accuracy=0.6692, over 5933.14 frames. ], batch size: 22, lr: 7.12e-03
2024-08-06 17:03:38,296 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:03:46,474 INFO [trainer.py:811] (4/8) Epoch 12, validation: loss=3.054, NarTop10Accuracy=0.7153, over 1905321.00 frames. 
2024-08-06 17:03:46,474 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:03:46,988 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.507e+02 1.899e+02 2.078e+02 2.276e+02 5.455e+02, threshold=4.157e+02, percent-clipped=0.1
2024-08-06 17:03:55,603 INFO [trainer.py:765] (4/8) Epoch 12, batch 1700, train_loss[loss=3.432, NarTop10Accuracy=0.644, over 6117.00 frames. ], tot_loss[loss=3.287, NarTop10Accuracy=0.6683, over 5925.82 frames. ], batch size: 13, lr: 7.11e-03
2024-08-06 17:04:22,121 INFO [trainer.py:765] (4/8) Epoch 12, batch 1800, train_loss[loss=3.607, NarTop10Accuracy=0.6025, over 7110.00 frames. ], tot_loss[loss=3.281, NarTop10Accuracy=0.6695, over 5997.31 frames. ], batch size: 22, lr: 7.10e-03
2024-08-06 17:04:48,591 INFO [trainer.py:765] (4/8) Epoch 12, batch 1900, train_loss[loss=3.275, NarTop10Accuracy=0.6757, over 6273.00 frames. ], tot_loss[loss=3.284, NarTop10Accuracy=0.6695, over 6043.43 frames. ], batch size: 50, lr: 7.08e-03
2024-08-06 17:05:14,197 INFO [trainer.py:765] (4/8) Epoch 12, batch 2000, train_loss[loss=3.541, NarTop10Accuracy=0.6095, over 6198.00 frames. ], tot_loss[loss=3.275, NarTop10Accuracy=0.6715, over 6013.67 frames. ], batch size: 51, lr: 7.07e-03
2024-08-06 17:05:39,467 INFO [trainer.py:765] (4/8) Epoch 12, batch 2100, train_loss[loss=3.402, NarTop10Accuracy=0.6477, over 4863.00 frames. ], tot_loss[loss=3.281, NarTop10Accuracy=0.6701, over 5979.23 frames. ], batch size: 5, lr: 7.06e-03
2024-08-06 17:06:04,691 INFO [trainer.py:765] (4/8) Epoch 12, batch 2200, train_loss[loss=3.516, NarTop10Accuracy=0.6209, over 7209.00 frames. ], tot_loss[loss=3.291, NarTop10Accuracy=0.6674, over 6012.35 frames. ], batch size: 31, lr: 7.05e-03
2024-08-06 17:06:29,847 INFO [trainer.py:765] (4/8) Epoch 12, batch 2300, train_loss[loss=3.261, NarTop10Accuracy=0.6715, over 5763.00 frames. ], tot_loss[loss=3.29, NarTop10Accuracy=0.6679, over 6015.04 frames. ], batch size: 9, lr: 7.04e-03
2024-08-06 17:06:54,200 INFO [trainer.py:765] (4/8) Epoch 12, batch 2400, train_loss[loss=3.143, NarTop10Accuracy=0.6941, over 5100.00 frames. ], tot_loss[loss=3.279, NarTop10Accuracy=0.6698, over 5784.91 frames. ], batch size: 7, lr: 7.03e-03
2024-08-06 17:07:17,645 INFO [trainer.py:765] (4/8) Epoch 12, batch 2500, train_loss[loss=3.253, NarTop10Accuracy=0.6701, over 5115.00 frames. ], tot_loss[loss=3.252, NarTop10Accuracy=0.6749, over 5476.30 frames. ], batch size: 7, lr: 7.02e-03
2024-08-06 17:07:37,805 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 17:08:40,079 INFO [trainer.py:765] (4/8) Epoch 13, batch 100, train_loss[loss=3.003, NarTop10Accuracy=0.7316, over 7326.00 frames. ], tot_loss[loss=3.298, NarTop10Accuracy=0.6663, over 2369.18 frames. ], batch size: 31, lr: 6.73e-03
2024-08-06 17:09:14,120 INFO [trainer.py:765] (4/8) Epoch 13, batch 200, train_loss[loss=3.007, NarTop10Accuracy=0.7242, over 6750.00 frames. ], tot_loss[loss=3.294, NarTop10Accuracy=0.6673, over 3868.52 frames. ], batch size: 17, lr: 6.72e-03
2024-08-06 17:09:46,277 INFO [trainer.py:765] (4/8) Epoch 13, batch 300, train_loss[loss=3.507, NarTop10Accuracy=0.6221, over 7257.00 frames. ], tot_loss[loss=3.263, NarTop10Accuracy=0.6736, over 4672.02 frames. ], batch size: 23, lr: 6.71e-03
2024-08-06 17:10:19,164 INFO [trainer.py:765] (4/8) Epoch 13, batch 400, train_loss[loss=2.906, NarTop10Accuracy=0.7415, over 5085.00 frames. ], tot_loss[loss=3.245, NarTop10Accuracy=0.6771, over 5122.41 frames. ], batch size: 7, lr: 6.70e-03
2024-08-06 17:10:49,335 INFO [trainer.py:765] (4/8) Epoch 13, batch 500, train_loss[loss=3.207, NarTop10Accuracy=0.6817, over 5934.00 frames. ], tot_loss[loss=3.243, NarTop10Accuracy=0.6776, over 5410.04 frames. ], batch size: 11, lr: 6.69e-03
2024-08-06 17:11:26,245 INFO [trainer.py:765] (4/8) Epoch 13, batch 600, train_loss[loss=3, NarTop10Accuracy=0.7232, over 5748.00 frames. ], tot_loss[loss=3.241, NarTop10Accuracy=0.6783, over 5658.19 frames. ], batch size: 9, lr: 6.68e-03
2024-08-06 17:11:57,381 INFO [trainer.py:765] (4/8) Epoch 13, batch 700, train_loss[loss=3.154, NarTop10Accuracy=0.6898, over 4941.00 frames. ], tot_loss[loss=3.245, NarTop10Accuracy=0.6774, over 5722.69 frames. ], batch size: 6, lr: 6.67e-03
2024-08-06 17:12:33,442 INFO [trainer.py:765] (4/8) Epoch 13, batch 800, train_loss[loss=2.949, NarTop10Accuracy=0.7427, over 4401.00 frames. ], tot_loss[loss=3.25, NarTop10Accuracy=0.6761, over 5782.43 frames. ], batch size: 5, lr: 6.66e-03
2024-08-06 17:13:10,031 INFO [trainer.py:765] (4/8) Epoch 13, batch 900, train_loss[loss=3.278, NarTop10Accuracy=0.6694, over 6741.00 frames. ], tot_loss[loss=3.244, NarTop10Accuracy=0.6774, over 5792.31 frames. ], batch size: 14, lr: 6.65e-03
2024-08-06 17:13:41,443 INFO [trainer.py:765] (4/8) Epoch 13, batch 1000, train_loss[loss=3.543, NarTop10Accuracy=0.6139, over 6687.00 frames. ], tot_loss[loss=3.25, NarTop10Accuracy=0.6763, over 5892.48 frames. ], batch size: 14, lr: 6.64e-03
2024-08-06 17:14:15,537 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:14:23,644 INFO [trainer.py:811] (4/8) Epoch 13, validation: loss=3.099, NarTop10Accuracy=0.7062, over 1905321.00 frames. 
2024-08-06 17:14:23,645 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:14:24,470 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.548e+02 1.948e+02 2.091e+02 2.295e+02 3.353e+02, threshold=4.181e+02, percent-clipped=0.0
2024-08-06 17:14:26,697 INFO [trainer.py:765] (4/8) Epoch 13, batch 1100, train_loss[loss=3.426, NarTop10Accuracy=0.6408, over 6681.00 frames. ], tot_loss[loss=3.256, NarTop10Accuracy=0.6749, over 5920.68 frames. ], batch size: 17, lr: 6.63e-03
2024-08-06 17:15:03,475 INFO [trainer.py:765] (4/8) Epoch 13, batch 1200, train_loss[loss=3.47, NarTop10Accuracy=0.6322, over 7119.00 frames. ], tot_loss[loss=3.257, NarTop10Accuracy=0.6746, over 5918.74 frames. ], batch size: 31, lr: 6.62e-03
2024-08-06 17:15:35,514 INFO [trainer.py:765] (4/8) Epoch 13, batch 1300, train_loss[loss=3.024, NarTop10Accuracy=0.7153, over 4311.00 frames. ], tot_loss[loss=3.258, NarTop10Accuracy=0.6744, over 5995.26 frames. ], batch size: 5, lr: 6.61e-03
2024-08-06 17:16:11,783 INFO [trainer.py:765] (4/8) Epoch 13, batch 1400, train_loss[loss=3.059, NarTop10Accuracy=0.713, over 6138.00 frames. ], tot_loss[loss=3.259, NarTop10Accuracy=0.6742, over 6005.76 frames. ], batch size: 11, lr: 6.60e-03
2024-08-06 17:16:39,788 INFO [trainer.py:765] (4/8) Epoch 13, batch 1500, train_loss[loss=3.583, NarTop10Accuracy=0.6047, over 5718.00 frames. ], tot_loss[loss=3.255, NarTop10Accuracy=0.675, over 5935.15 frames. ], batch size: 51, lr: 6.59e-03
2024-08-06 17:17:07,603 INFO [trainer.py:765] (4/8) Epoch 13, batch 1600, train_loss[loss=3.039, NarTop10Accuracy=0.7277, over 7035.00 frames. ], tot_loss[loss=3.268, NarTop10Accuracy=0.6726, over 5929.74 frames. ], batch size: 22, lr: 6.58e-03
2024-08-06 17:17:34,259 INFO [trainer.py:765] (4/8) Epoch 13, batch 1700, train_loss[loss=3.138, NarTop10Accuracy=0.6805, over 6636.00 frames. ], tot_loss[loss=3.262, NarTop10Accuracy=0.6737, over 5901.16 frames. ], batch size: 14, lr: 6.57e-03
2024-08-06 17:18:00,762 INFO [trainer.py:765] (4/8) Epoch 13, batch 1800, train_loss[loss=3.086, NarTop10Accuracy=0.7071, over 6867.00 frames. ], tot_loss[loss=3.256, NarTop10Accuracy=0.6752, over 5980.53 frames. ], batch size: 22, lr: 6.56e-03
2024-08-06 17:18:27,244 INFO [trainer.py:765] (4/8) Epoch 13, batch 1900, train_loss[loss=3.547, NarTop10Accuracy=0.6149, over 5835.00 frames. ], tot_loss[loss=3.251, NarTop10Accuracy=0.676, over 6033.26 frames. ], batch size: 50, lr: 6.55e-03
2024-08-06 17:18:52,777 INFO [trainer.py:765] (4/8) Epoch 13, batch 2000, train_loss[loss=3.574, NarTop10Accuracy=0.6157, over 6267.00 frames. ], tot_loss[loss=3.237, NarTop10Accuracy=0.6787, over 6001.07 frames. ], batch size: 50, lr: 6.54e-03
2024-08-06 17:19:18,147 INFO [trainer.py:765] (4/8) Epoch 13, batch 2100, train_loss[loss=2.968, NarTop10Accuracy=0.7355, over 3879.00 frames. ], tot_loss[loss=3.236, NarTop10Accuracy=0.6788, over 5958.88 frames. ], batch size: 4, lr: 6.53e-03
2024-08-06 17:19:43,412 INFO [trainer.py:765] (4/8) Epoch 13, batch 2200, train_loss[loss=3.433, NarTop10Accuracy=0.6338, over 7071.00 frames. ], tot_loss[loss=3.243, NarTop10Accuracy=0.6768, over 5997.10 frames. ], batch size: 31, lr: 6.52e-03
2024-08-06 17:20:08,543 INFO [trainer.py:765] (4/8) Epoch 13, batch 2300, train_loss[loss=3.712, NarTop10Accuracy=0.575, over 5706.00 frames. ], tot_loss[loss=3.26, NarTop10Accuracy=0.6732, over 6007.78 frames. ], batch size: 9, lr: 6.51e-03
2024-08-06 17:20:32,939 INFO [trainer.py:765] (4/8) Epoch 13, batch 2400, train_loss[loss=3.584, NarTop10Accuracy=0.6043, over 5247.00 frames. ], tot_loss[loss=3.231, NarTop10Accuracy=0.6788, over 5760.10 frames. ], batch size: 7, lr: 6.50e-03
2024-08-06 17:20:56,410 INFO [trainer.py:765] (4/8) Epoch 13, batch 2500, train_loss[loss=3.482, NarTop10Accuracy=0.6257, over 5235.00 frames. ], tot_loss[loss=3.218, NarTop10Accuracy=0.6816, over 5473.60 frames. ], batch size: 7, lr: 6.49e-03
2024-08-06 17:21:16,322 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 17:22:19,315 INFO [trainer.py:765] (4/8) Epoch 14, batch 100, train_loss[loss=2.96, NarTop10Accuracy=0.735, over 7314.00 frames. ], tot_loss[loss=3.209, NarTop10Accuracy=0.6844, over 2392.00 frames. ], batch size: 31, lr: 6.24e-03
2024-08-06 17:22:50,377 INFO [trainer.py:765] (4/8) Epoch 14, batch 200, train_loss[loss=3.282, NarTop10Accuracy=0.6663, over 6843.00 frames. ], tot_loss[loss=3.229, NarTop10Accuracy=0.6796, over 3864.48 frames. ], batch size: 17, lr: 6.23e-03
2024-08-06 17:23:23,879 INFO [trainer.py:765] (4/8) Epoch 14, batch 300, train_loss[loss=3.181, NarTop10Accuracy=0.6922, over 7065.00 frames. ], tot_loss[loss=3.205, NarTop10Accuracy=0.6849, over 4664.16 frames. ], batch size: 22, lr: 6.22e-03
2024-08-06 17:23:57,485 INFO [trainer.py:765] (4/8) Epoch 14, batch 400, train_loss[loss=3.003, NarTop10Accuracy=0.7231, over 5169.00 frames. ], tot_loss[loss=3.223, NarTop10Accuracy=0.6809, over 5110.03 frames. ], batch size: 7, lr: 6.22e-03
2024-08-06 17:24:32,113 INFO [trainer.py:765] (4/8) Epoch 14, batch 500, train_loss[loss=3.362, NarTop10Accuracy=0.6493, over 5994.00 frames. ], tot_loss[loss=3.231, NarTop10Accuracy=0.6793, over 5390.55 frames. ], batch size: 11, lr: 6.21e-03
2024-08-06 17:24:36,213 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:24:44,275 INFO [trainer.py:811] (4/8) Epoch 14, validation: loss=3.004, NarTop10Accuracy=0.726, over 1905321.00 frames. 
2024-08-06 17:24:44,275 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:24:44,822 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.601e+02 1.969e+02 2.114e+02 2.287e+02 4.406e+02, threshold=4.227e+02, percent-clipped=0.1
2024-08-06 17:25:12,913 INFO [trainer.py:765] (4/8) Epoch 14, batch 600, train_loss[loss=2.879, NarTop10Accuracy=0.7499, over 5652.00 frames. ], tot_loss[loss=3.231, NarTop10Accuracy=0.6791, over 5644.01 frames. ], batch size: 9, lr: 6.20e-03
2024-08-06 17:25:48,547 INFO [trainer.py:765] (4/8) Epoch 14, batch 700, train_loss[loss=3.463, NarTop10Accuracy=0.6359, over 4311.00 frames. ], tot_loss[loss=3.223, NarTop10Accuracy=0.6809, over 5724.56 frames. ], batch size: 5, lr: 6.19e-03
2024-08-06 17:26:25,278 INFO [trainer.py:765] (4/8) Epoch 14, batch 800, train_loss[loss=2.93, NarTop10Accuracy=0.741, over 4287.00 frames. ], tot_loss[loss=3.212, NarTop10Accuracy=0.6833, over 5780.08 frames. ], batch size: 5, lr: 6.18e-03
2024-08-06 17:26:57,658 INFO [trainer.py:765] (4/8) Epoch 14, batch 900, train_loss[loss=3.283, NarTop10Accuracy=0.6646, over 6093.00 frames. ], tot_loss[loss=3.208, NarTop10Accuracy=0.6841, over 5801.50 frames. ], batch size: 13, lr: 6.17e-03
2024-08-06 17:27:31,715 INFO [trainer.py:765] (4/8) Epoch 14, batch 1000, train_loss[loss=3.49, NarTop10Accuracy=0.6236, over 6243.00 frames. ], tot_loss[loss=3.221, NarTop10Accuracy=0.681, over 5915.19 frames. ], batch size: 13, lr: 6.16e-03
2024-08-06 17:28:11,596 INFO [trainer.py:765] (4/8) Epoch 14, batch 1100, train_loss[loss=3.023, NarTop10Accuracy=0.7261, over 6834.00 frames. ], tot_loss[loss=3.219, NarTop10Accuracy=0.6818, over 5946.71 frames. ], batch size: 17, lr: 6.15e-03
2024-08-06 17:28:40,732 INFO [trainer.py:765] (4/8) Epoch 14, batch 1200, train_loss[loss=3.466, NarTop10Accuracy=0.6328, over 7056.00 frames. ], tot_loss[loss=3.224, NarTop10Accuracy=0.6809, over 5907.99 frames. ], batch size: 31, lr: 6.15e-03
2024-08-06 17:29:16,213 INFO [trainer.py:765] (4/8) Epoch 14, batch 1300, train_loss[loss=3.473, NarTop10Accuracy=0.6219, over 5079.00 frames. ], tot_loss[loss=3.22, NarTop10Accuracy=0.6816, over 5974.44 frames. ], batch size: 6, lr: 6.14e-03
2024-08-06 17:29:54,601 INFO [trainer.py:765] (4/8) Epoch 14, batch 1400, train_loss[loss=3.296, NarTop10Accuracy=0.6698, over 6015.00 frames. ], tot_loss[loss=3.23, NarTop10Accuracy=0.6798, over 6004.61 frames. ], batch size: 11, lr: 6.13e-03
2024-08-06 17:30:25,314 INFO [trainer.py:765] (4/8) Epoch 14, batch 1500, train_loss[loss=3.804, NarTop10Accuracy=0.5648, over 6189.00 frames. ], tot_loss[loss=3.241, NarTop10Accuracy=0.6775, over 5952.03 frames. ], batch size: 50, lr: 6.12e-03
2024-08-06 17:30:53,041 INFO [trainer.py:765] (4/8) Epoch 14, batch 1600, train_loss[loss=2.976, NarTop10Accuracy=0.7259, over 6987.00 frames. ], tot_loss[loss=3.231, NarTop10Accuracy=0.6795, over 5945.95 frames. ], batch size: 22, lr: 6.11e-03
2024-08-06 17:31:19,727 INFO [trainer.py:765] (4/8) Epoch 14, batch 1700, train_loss[loss=3.061, NarTop10Accuracy=0.7205, over 6663.00 frames. ], tot_loss[loss=3.213, NarTop10Accuracy=0.6837, over 5940.04 frames. ], batch size: 14, lr: 6.10e-03
2024-08-06 17:31:46,288 INFO [trainer.py:765] (4/8) Epoch 14, batch 1800, train_loss[loss=3.04, NarTop10Accuracy=0.7166, over 7041.00 frames. ], tot_loss[loss=3.192, NarTop10Accuracy=0.6879, over 5985.90 frames. ], batch size: 22, lr: 6.09e-03
2024-08-06 17:32:12,726 INFO [trainer.py:765] (4/8) Epoch 14, batch 1900, train_loss[loss=3.729, NarTop10Accuracy=0.5711, over 5985.00 frames. ], tot_loss[loss=3.204, NarTop10Accuracy=0.6857, over 6044.73 frames. ], batch size: 51, lr: 6.09e-03
2024-08-06 17:32:38,281 INFO [trainer.py:765] (4/8) Epoch 14, batch 2000, train_loss[loss=3.257, NarTop10Accuracy=0.6796, over 6087.00 frames. ], tot_loss[loss=3.215, NarTop10Accuracy=0.6831, over 6027.79 frames. ], batch size: 50, lr: 6.08e-03
2024-08-06 17:33:03,645 INFO [trainer.py:765] (4/8) Epoch 14, batch 2100, train_loss[loss=2.974, NarTop10Accuracy=0.7372, over 4854.00 frames. ], tot_loss[loss=3.218, NarTop10Accuracy=0.6824, over 6000.53 frames. ], batch size: 5, lr: 6.07e-03
2024-08-06 17:33:28,997 INFO [trainer.py:765] (4/8) Epoch 14, batch 2200, train_loss[loss=3.249, NarTop10Accuracy=0.6774, over 7470.00 frames. ], tot_loss[loss=3.219, NarTop10Accuracy=0.6823, over 6018.33 frames. ], batch size: 31, lr: 6.06e-03
2024-08-06 17:33:54,085 INFO [trainer.py:765] (4/8) Epoch 14, batch 2300, train_loss[loss=2.834, NarTop10Accuracy=0.7621, over 5703.00 frames. ], tot_loss[loss=3.234, NarTop10Accuracy=0.6794, over 6021.89 frames. ], batch size: 9, lr: 6.05e-03
2024-08-06 17:34:18,533 INFO [trainer.py:765] (4/8) Epoch 14, batch 2400, train_loss[loss=2.869, NarTop10Accuracy=0.7596, over 5223.00 frames. ], tot_loss[loss=3.234, NarTop10Accuracy=0.6788, over 5761.86 frames. ], batch size: 7, lr: 6.04e-03
2024-08-06 17:34:42,115 INFO [trainer.py:765] (4/8) Epoch 14, batch 2500, train_loss[loss=2.956, NarTop10Accuracy=0.7338, over 5073.00 frames. ], tot_loss[loss=3.2, NarTop10Accuracy=0.6857, over 5479.95 frames. ], batch size: 7, lr: 6.04e-03
2024-08-06 17:34:45,394 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:34:53,209 INFO [trainer.py:811] (4/8) Epoch 14, validation: loss=3.062, NarTop10Accuracy=0.7136, over 1905321.00 frames. 
2024-08-06 17:34:53,209 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:34:53,679 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.574e+02 1.975e+02 2.132e+02 2.304e+02 3.875e+02, threshold=4.265e+02, percent-clipped=0.0
2024-08-06 17:35:09,645 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 17:36:11,738 INFO [trainer.py:765] (4/8) Epoch 15, batch 100, train_loss[loss=3.109, NarTop10Accuracy=0.7027, over 7182.00 frames. ], tot_loss[loss=3.212, NarTop10Accuracy=0.6836, over 2350.11 frames. ], batch size: 31, lr: 5.82e-03
2024-08-06 17:36:44,334 INFO [trainer.py:765] (4/8) Epoch 15, batch 200, train_loss[loss=3.546, NarTop10Accuracy=0.6146, over 6660.00 frames. ], tot_loss[loss=3.188, NarTop10Accuracy=0.6882, over 3829.15 frames. ], batch size: 17, lr: 5.81e-03
2024-08-06 17:37:17,714 INFO [trainer.py:765] (4/8) Epoch 15, batch 300, train_loss[loss=3.314, NarTop10Accuracy=0.667, over 6972.00 frames. ], tot_loss[loss=3.19, NarTop10Accuracy=0.6879, over 4658.30 frames. ], batch size: 22, lr: 5.80e-03
2024-08-06 17:37:48,904 INFO [trainer.py:765] (4/8) Epoch 15, batch 400, train_loss[loss=2.985, NarTop10Accuracy=0.7338, over 5790.00 frames. ], tot_loss[loss=3.187, NarTop10Accuracy=0.6891, over 5121.25 frames. ], batch size: 8, lr: 5.80e-03
2024-08-06 17:38:22,354 INFO [trainer.py:765] (4/8) Epoch 15, batch 500, train_loss[loss=2.865, NarTop10Accuracy=0.7507, over 5985.00 frames. ], tot_loss[loss=3.187, NarTop10Accuracy=0.689, over 5372.96 frames. ], batch size: 11, lr: 5.79e-03
2024-08-06 17:38:53,093 INFO [trainer.py:765] (4/8) Epoch 15, batch 600, train_loss[loss=2.899, NarTop10Accuracy=0.7491, over 5664.00 frames. ], tot_loss[loss=3.194, NarTop10Accuracy=0.687, over 5623.54 frames. ], batch size: 9, lr: 5.78e-03
2024-08-06 17:39:27,922 INFO [trainer.py:765] (4/8) Epoch 15, batch 700, train_loss[loss=2.89, NarTop10Accuracy=0.7521, over 5118.00 frames. ], tot_loss[loss=3.202, NarTop10Accuracy=0.6857, over 5723.01 frames. ], batch size: 6, lr: 5.77e-03
2024-08-06 17:40:05,564 INFO [trainer.py:765] (4/8) Epoch 15, batch 800, train_loss[loss=3.359, NarTop10Accuracy=0.6472, over 5040.00 frames. ], tot_loss[loss=3.228, NarTop10Accuracy=0.6801, over 5765.60 frames. ], batch size: 6, lr: 5.76e-03
2024-08-06 17:40:35,791 INFO [trainer.py:765] (4/8) Epoch 15, batch 900, train_loss[loss=3.354, NarTop10Accuracy=0.6533, over 6159.00 frames. ], tot_loss[loss=3.206, NarTop10Accuracy=0.6844, over 5793.28 frames. ], batch size: 13, lr: 5.76e-03
2024-08-06 17:41:11,251 INFO [trainer.py:765] (4/8) Epoch 15, batch 1000, train_loss[loss=3.041, NarTop10Accuracy=0.7115, over 6567.00 frames. ], tot_loss[loss=3.192, NarTop10Accuracy=0.6872, over 5896.83 frames. ], batch size: 14, lr: 5.75e-03
2024-08-06 17:41:46,452 INFO [trainer.py:765] (4/8) Epoch 15, batch 1100, train_loss[loss=3.078, NarTop10Accuracy=0.7015, over 6807.00 frames. ], tot_loss[loss=3.193, NarTop10Accuracy=0.6871, over 5945.70 frames. ], batch size: 17, lr: 5.74e-03
2024-08-06 17:42:19,456 INFO [trainer.py:765] (4/8) Epoch 15, batch 1200, train_loss[loss=3.334, NarTop10Accuracy=0.6545, over 6978.00 frames. ], tot_loss[loss=3.227, NarTop10Accuracy=0.6799, over 5938.76 frames. ], batch size: 31, lr: 5.73e-03
2024-08-06 17:42:54,427 INFO [trainer.py:765] (4/8) Epoch 15, batch 1300, train_loss[loss=2.965, NarTop10Accuracy=0.7255, over 5157.00 frames. ], tot_loss[loss=3.21, NarTop10Accuracy=0.6834, over 6002.21 frames. ], batch size: 6, lr: 5.73e-03
2024-08-06 17:43:26,607 INFO [trainer.py:765] (4/8) Epoch 15, batch 1400, train_loss[loss=3.471, NarTop10Accuracy=0.6253, over 6078.00 frames. ], tot_loss[loss=3.215, NarTop10Accuracy=0.6824, over 6016.80 frames. ], batch size: 11, lr: 5.72e-03
2024-08-06 17:43:56,558 INFO [trainer.py:765] (4/8) Epoch 15, batch 1500, train_loss[loss=3.077, NarTop10Accuracy=0.709, over 6159.00 frames. ], tot_loss[loss=3.217, NarTop10Accuracy=0.682, over 5961.03 frames. ], batch size: 50, lr: 5.71e-03
2024-08-06 17:44:24,241 INFO [trainer.py:765] (4/8) Epoch 15, batch 1600, train_loss[loss=3.616, NarTop10Accuracy=0.5989, over 7356.00 frames. ], tot_loss[loss=3.198, NarTop10Accuracy=0.686, over 5936.46 frames. ], batch size: 23, lr: 5.70e-03
2024-08-06 17:44:50,856 INFO [trainer.py:765] (4/8) Epoch 15, batch 1700, train_loss[loss=3.015, NarTop10Accuracy=0.7106, over 6120.00 frames. ], tot_loss[loss=3.193, NarTop10Accuracy=0.6869, over 5918.77 frames. ], batch size: 13, lr: 5.70e-03
2024-08-06 17:45:17,293 INFO [trainer.py:765] (4/8) Epoch 15, batch 1800, train_loss[loss=3.138, NarTop10Accuracy=0.6962, over 7518.00 frames. ], tot_loss[loss=3.189, NarTop10Accuracy=0.6883, over 5989.08 frames. ], batch size: 23, lr: 5.69e-03
2024-08-06 17:45:43,679 INFO [trainer.py:765] (4/8) Epoch 15, batch 1900, train_loss[loss=3.186, NarTop10Accuracy=0.6899, over 6339.00 frames. ], tot_loss[loss=3.216, NarTop10Accuracy=0.6826, over 6031.20 frames. ], batch size: 50, lr: 5.68e-03
2024-08-06 17:45:53,540 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:46:01,743 INFO [trainer.py:811] (4/8) Epoch 15, validation: loss=3.006, NarTop10Accuracy=0.725, over 1905321.00 frames. 
2024-08-06 17:46:01,743 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:46:02,216 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.631e+02 2.004e+02 2.149e+02 2.324e+02 3.721e+02, threshold=4.298e+02, percent-clipped=0.0
2024-08-06 17:46:17,371 INFO [trainer.py:765] (4/8) Epoch 15, batch 2000, train_loss[loss=3.258, NarTop10Accuracy=0.6744, over 5493.00 frames. ], tot_loss[loss=3.2, NarTop10Accuracy=0.6857, over 6005.19 frames. ], batch size: 50, lr: 5.67e-03
2024-08-06 17:46:42,773 INFO [trainer.py:765] (4/8) Epoch 15, batch 2100, train_loss[loss=3.194, NarTop10Accuracy=0.6856, over 4023.00 frames. ], tot_loss[loss=3.204, NarTop10Accuracy=0.6848, over 5971.54 frames. ], batch size: 4, lr: 5.67e-03
2024-08-06 17:47:08,032 INFO [trainer.py:765] (4/8) Epoch 15, batch 2200, train_loss[loss=3.056, NarTop10Accuracy=0.7148, over 7140.00 frames. ], tot_loss[loss=3.204, NarTop10Accuracy=0.6851, over 6005.32 frames. ], batch size: 31, lr: 5.66e-03
2024-08-06 17:47:33,291 INFO [trainer.py:765] (4/8) Epoch 15, batch 2300, train_loss[loss=3.515, NarTop10Accuracy=0.6157, over 5727.00 frames. ], tot_loss[loss=3.205, NarTop10Accuracy=0.6849, over 6024.94 frames. ], batch size: 9, lr: 5.65e-03
2024-08-06 17:47:57,640 INFO [trainer.py:765] (4/8) Epoch 15, batch 2400, train_loss[loss=3.48, NarTop10Accuracy=0.6346, over 5070.00 frames. ], tot_loss[loss=3.187, NarTop10Accuracy=0.6887, over 5792.62 frames. ], batch size: 7, lr: 5.65e-03
2024-08-06 17:48:21,161 INFO [trainer.py:765] (4/8) Epoch 15, batch 2500, train_loss[loss=2.972, NarTop10Accuracy=0.7382, over 5082.00 frames. ], tot_loss[loss=3.17, NarTop10Accuracy=0.6918, over 5482.78 frames. ], batch size: 7, lr: 5.64e-03
2024-08-06 17:48:41,525 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 17:49:41,220 INFO [trainer.py:765] (4/8) Epoch 16, batch 100, train_loss[loss=3.469, NarTop10Accuracy=0.6372, over 7191.00 frames. ], tot_loss[loss=3.151, NarTop10Accuracy=0.6961, over 2359.93 frames. ], batch size: 31, lr: 5.45e-03
2024-08-06 17:50:12,156 INFO [trainer.py:765] (4/8) Epoch 16, batch 200, train_loss[loss=2.98, NarTop10Accuracy=0.7327, over 6840.00 frames. ], tot_loss[loss=3.204, NarTop10Accuracy=0.6845, over 3845.81 frames. ], batch size: 17, lr: 5.44e-03
2024-08-06 17:50:45,158 INFO [trainer.py:765] (4/8) Epoch 16, batch 300, train_loss[loss=3.127, NarTop10Accuracy=0.702, over 7092.00 frames. ], tot_loss[loss=3.198, NarTop10Accuracy=0.6859, over 4652.18 frames. ], batch size: 22, lr: 5.43e-03
2024-08-06 17:51:15,975 INFO [trainer.py:765] (4/8) Epoch 16, batch 400, train_loss[loss=3.433, NarTop10Accuracy=0.6259, over 5136.00 frames. ], tot_loss[loss=3.195, NarTop10Accuracy=0.6865, over 5111.50 frames. ], batch size: 7, lr: 5.43e-03
2024-08-06 17:51:50,322 INFO [trainer.py:765] (4/8) Epoch 16, batch 500, train_loss[loss=3.036, NarTop10Accuracy=0.7213, over 5973.00 frames. ], tot_loss[loss=3.186, NarTop10Accuracy=0.6884, over 5385.51 frames. ], batch size: 11, lr: 5.42e-03
2024-08-06 17:52:24,250 INFO [trainer.py:765] (4/8) Epoch 16, batch 600, train_loss[loss=3.005, NarTop10Accuracy=0.729, over 5733.00 frames. ], tot_loss[loss=3.195, NarTop10Accuracy=0.6864, over 5655.47 frames. ], batch size: 9, lr: 5.41e-03
2024-08-06 17:52:55,385 INFO [trainer.py:765] (4/8) Epoch 16, batch 700, train_loss[loss=2.857, NarTop10Accuracy=0.7487, over 4935.00 frames. ], tot_loss[loss=3.192, NarTop10Accuracy=0.6874, over 5732.71 frames. ], batch size: 6, lr: 5.41e-03
2024-08-06 17:53:33,814 INFO [trainer.py:765] (4/8) Epoch 16, batch 800, train_loss[loss=2.93, NarTop10Accuracy=0.7286, over 5232.00 frames. ], tot_loss[loss=3.182, NarTop10Accuracy=0.6896, over 5798.94 frames. ], batch size: 6, lr: 5.40e-03
2024-08-06 17:54:03,922 INFO [trainer.py:765] (4/8) Epoch 16, batch 900, train_loss[loss=3.519, NarTop10Accuracy=0.6184, over 6234.00 frames. ], tot_loss[loss=3.174, NarTop10Accuracy=0.6914, over 5813.10 frames. ], batch size: 13, lr: 5.39e-03
2024-08-06 17:54:37,606 INFO [trainer.py:765] (4/8) Epoch 16, batch 1000, train_loss[loss=2.989, NarTop10Accuracy=0.7289, over 6255.00 frames. ], tot_loss[loss=3.163, NarTop10Accuracy=0.6933, over 5892.64 frames. ], batch size: 13, lr: 5.39e-03
2024-08-06 17:55:17,195 INFO [trainer.py:765] (4/8) Epoch 16, batch 1100, train_loss[loss=3.203, NarTop10Accuracy=0.6821, over 6933.00 frames. ], tot_loss[loss=3.194, NarTop10Accuracy=0.6867, over 5934.88 frames. ], batch size: 17, lr: 5.38e-03
2024-08-06 17:55:46,208 INFO [trainer.py:765] (4/8) Epoch 16, batch 1200, train_loss[loss=3.491, NarTop10Accuracy=0.6208, over 7206.00 frames. ], tot_loss[loss=3.205, NarTop10Accuracy=0.6846, over 5912.06 frames. ], batch size: 31, lr: 5.37e-03
2024-08-06 17:56:22,774 INFO [trainer.py:765] (4/8) Epoch 16, batch 1300, train_loss[loss=3.245, NarTop10Accuracy=0.6746, over 4989.00 frames. ], tot_loss[loss=3.197, NarTop10Accuracy=0.686, over 5985.80 frames. ], batch size: 6, lr: 5.37e-03
2024-08-06 17:56:44,647 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 17:56:53,428 INFO [trainer.py:811] (4/8) Epoch 16, validation: loss=3.112, NarTop10Accuracy=0.703, over 1905321.00 frames. 
2024-08-06 17:56:53,429 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 17:56:54,007 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.620e+02 1.974e+02 2.136e+02 2.310e+02 5.351e+02, threshold=4.271e+02, percent-clipped=0.2
2024-08-06 17:57:06,171 INFO [trainer.py:765] (4/8) Epoch 16, batch 1400, train_loss[loss=3.151, NarTop10Accuracy=0.6955, over 6183.00 frames. ], tot_loss[loss=3.189, NarTop10Accuracy=0.6879, over 6017.14 frames. ], batch size: 11, lr: 5.36e-03
2024-08-06 17:57:34,033 INFO [trainer.py:765] (4/8) Epoch 16, batch 1500, train_loss[loss=3.386, NarTop10Accuracy=0.6482, over 5556.00 frames. ], tot_loss[loss=3.185, NarTop10Accuracy=0.6887, over 5948.60 frames. ], batch size: 50, lr: 5.35e-03
2024-08-06 17:58:01,775 INFO [trainer.py:765] (4/8) Epoch 16, batch 1600, train_loss[loss=2.981, NarTop10Accuracy=0.7252, over 7095.00 frames. ], tot_loss[loss=3.182, NarTop10Accuracy=0.6893, over 5917.08 frames. ], batch size: 22, lr: 5.35e-03
2024-08-06 17:58:28,475 INFO [trainer.py:765] (4/8) Epoch 16, batch 1700, train_loss[loss=2.97, NarTop10Accuracy=0.7414, over 6540.00 frames. ], tot_loss[loss=3.193, NarTop10Accuracy=0.687, over 5912.40 frames. ], batch size: 14, lr: 5.34e-03
2024-08-06 17:58:54,976 INFO [trainer.py:765] (4/8) Epoch 16, batch 1800, train_loss[loss=3.04, NarTop10Accuracy=0.714, over 7164.00 frames. ], tot_loss[loss=3.179, NarTop10Accuracy=0.6902, over 5981.10 frames. ], batch size: 22, lr: 5.33e-03
2024-08-06 17:59:21,360 INFO [trainer.py:765] (4/8) Epoch 16, batch 1900, train_loss[loss=3.469, NarTop10Accuracy=0.6325, over 6132.00 frames. ], tot_loss[loss=3.206, NarTop10Accuracy=0.6849, over 6004.84 frames. ], batch size: 50, lr: 5.33e-03
2024-08-06 17:59:46,857 INFO [trainer.py:765] (4/8) Epoch 16, batch 2000, train_loss[loss=3.091, NarTop10Accuracy=0.7086, over 6429.00 frames. ], tot_loss[loss=3.181, NarTop10Accuracy=0.6901, over 5985.90 frames. ], batch size: 52, lr: 5.32e-03
2024-08-06 18:00:12,117 INFO [trainer.py:765] (4/8) Epoch 16, batch 2100, train_loss[loss=3.382, NarTop10Accuracy=0.6419, over 4926.00 frames. ], tot_loss[loss=3.204, NarTop10Accuracy=0.6851, over 5970.91 frames. ], batch size: 5, lr: 5.32e-03
2024-08-06 18:00:37,334 INFO [trainer.py:765] (4/8) Epoch 16, batch 2200, train_loss[loss=3.201, NarTop10Accuracy=0.686, over 7395.00 frames. ], tot_loss[loss=3.22, NarTop10Accuracy=0.6812, over 5996.05 frames. ], batch size: 31, lr: 5.31e-03
2024-08-06 18:01:02,502 INFO [trainer.py:765] (4/8) Epoch 16, batch 2300, train_loss[loss=2.957, NarTop10Accuracy=0.7312, over 5682.00 frames. ], tot_loss[loss=3.218, NarTop10Accuracy=0.6821, over 6012.96 frames. ], batch size: 9, lr: 5.30e-03
2024-08-06 18:01:26,884 INFO [trainer.py:765] (4/8) Epoch 16, batch 2400, train_loss[loss=2.958, NarTop10Accuracy=0.7378, over 5133.00 frames. ], tot_loss[loss=3.192, NarTop10Accuracy=0.6873, over 5769.63 frames. ], batch size: 7, lr: 5.30e-03
2024-08-06 18:01:50,406 INFO [trainer.py:765] (4/8) Epoch 16, batch 2500, train_loss[loss=3.076, NarTop10Accuracy=0.7211, over 4974.00 frames. ], tot_loss[loss=3.17, NarTop10Accuracy=0.6919, over 5457.92 frames. ], batch size: 7, lr: 5.29e-03
2024-08-06 18:02:10,325 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 18:03:08,530 INFO [trainer.py:765] (4/8) Epoch 17, batch 100, train_loss[loss=3.182, NarTop10Accuracy=0.6836, over 7341.00 frames. ], tot_loss[loss=3.133, NarTop10Accuracy=0.699, over 2359.24 frames. ], batch size: 32, lr: 5.12e-03
2024-08-06 18:03:45,145 INFO [trainer.py:765] (4/8) Epoch 17, batch 200, train_loss[loss=3.378, NarTop10Accuracy=0.6556, over 6774.00 frames. ], tot_loss[loss=3.155, NarTop10Accuracy=0.6946, over 3840.47 frames. ], batch size: 17, lr: 5.12e-03
2024-08-06 18:04:19,590 INFO [trainer.py:765] (4/8) Epoch 17, batch 300, train_loss[loss=3.318, NarTop10Accuracy=0.6642, over 7053.00 frames. ], tot_loss[loss=3.174, NarTop10Accuracy=0.691, over 4644.19 frames. ], batch size: 22, lr: 5.11e-03
2024-08-06 18:04:48,401 INFO [trainer.py:765] (4/8) Epoch 17, batch 400, train_loss[loss=3.34, NarTop10Accuracy=0.6667, over 5130.00 frames. ], tot_loss[loss=3.169, NarTop10Accuracy=0.6919, over 5086.87 frames. ], batch size: 7, lr: 5.10e-03
2024-08-06 18:05:24,680 INFO [trainer.py:765] (4/8) Epoch 17, batch 500, train_loss[loss=2.869, NarTop10Accuracy=0.7528, over 6042.00 frames. ], tot_loss[loss=3.155, NarTop10Accuracy=0.6947, over 5375.87 frames. ], batch size: 11, lr: 5.10e-03
2024-08-06 18:05:58,739 INFO [trainer.py:765] (4/8) Epoch 17, batch 600, train_loss[loss=3.128, NarTop10Accuracy=0.7005, over 5661.00 frames. ], tot_loss[loss=3.167, NarTop10Accuracy=0.692, over 5649.46 frames. ], batch size: 9, lr: 5.09e-03
2024-08-06 18:06:32,474 INFO [trainer.py:765] (4/8) Epoch 17, batch 700, train_loss[loss=3.099, NarTop10Accuracy=0.7032, over 4932.00 frames. ], tot_loss[loss=3.171, NarTop10Accuracy=0.6914, over 5722.54 frames. ], batch size: 6, lr: 5.08e-03
2024-08-06 18:07:02,723 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 18:07:10,763 INFO [trainer.py:811] (4/8) Epoch 17, validation: loss=3.018, NarTop10Accuracy=0.7223, over 1905321.00 frames. 
2024-08-06 18:07:10,763 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 18:07:11,312 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.649e+02 2.005e+02 2.161e+02 2.341e+02 3.806e+02, threshold=4.323e+02, percent-clipped=0.0
2024-08-06 18:07:14,353 INFO [trainer.py:765] (4/8) Epoch 17, batch 800, train_loss[loss=3.056, NarTop10Accuracy=0.7064, over 4974.00 frames. ], tot_loss[loss=3.179, NarTop10Accuracy=0.6895, over 5764.47 frames. ], batch size: 6, lr: 5.08e-03
2024-08-06 18:07:49,721 INFO [trainer.py:765] (4/8) Epoch 17, batch 900, train_loss[loss=3.641, NarTop10Accuracy=0.5983, over 6834.00 frames. ], tot_loss[loss=3.162, NarTop10Accuracy=0.6937, over 5800.29 frames. ], batch size: 14, lr: 5.07e-03
2024-08-06 18:08:21,598 INFO [trainer.py:765] (4/8) Epoch 17, batch 1000, train_loss[loss=3.228, NarTop10Accuracy=0.6782, over 6636.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6915, over 5918.32 frames. ], batch size: 14, lr: 5.07e-03
2024-08-06 18:09:03,106 INFO [trainer.py:765] (4/8) Epoch 17, batch 1100, train_loss[loss=2.987, NarTop10Accuracy=0.7317, over 6861.00 frames. ], tot_loss[loss=3.174, NarTop10Accuracy=0.6913, over 5958.86 frames. ], batch size: 17, lr: 5.06e-03
2024-08-06 18:09:36,746 INFO [trainer.py:765] (4/8) Epoch 17, batch 1200, train_loss[loss=3.137, NarTop10Accuracy=0.7013, over 7086.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6913, over 5945.42 frames. ], batch size: 31, lr: 5.06e-03
2024-08-06 18:10:10,689 INFO [trainer.py:765] (4/8) Epoch 17, batch 1300, train_loss[loss=3.157, NarTop10Accuracy=0.6878, over 5130.00 frames. ], tot_loss[loss=3.173, NarTop10Accuracy=0.691, over 6016.86 frames. ], batch size: 6, lr: 5.05e-03
2024-08-06 18:10:48,027 INFO [trainer.py:765] (4/8) Epoch 17, batch 1400, train_loss[loss=3.293, NarTop10Accuracy=0.6625, over 6069.00 frames. ], tot_loss[loss=3.179, NarTop10Accuracy=0.6896, over 6040.31 frames. ], batch size: 11, lr: 5.04e-03
2024-08-06 18:11:19,105 INFO [trainer.py:765] (4/8) Epoch 17, batch 1500, train_loss[loss=3.382, NarTop10Accuracy=0.6485, over 6426.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6908, over 5994.62 frames. ], batch size: 50, lr: 5.04e-03
2024-08-06 18:11:46,855 INFO [trainer.py:765] (4/8) Epoch 17, batch 1600, train_loss[loss=3.053, NarTop10Accuracy=0.7188, over 6915.00 frames. ], tot_loss[loss=3.159, NarTop10Accuracy=0.6936, over 5949.67 frames. ], batch size: 22, lr: 5.03e-03
2024-08-06 18:12:13,509 INFO [trainer.py:765] (4/8) Epoch 17, batch 1700, train_loss[loss=3.461, NarTop10Accuracy=0.6349, over 6135.00 frames. ], tot_loss[loss=3.178, NarTop10Accuracy=0.6899, over 5928.61 frames. ], batch size: 13, lr: 5.03e-03
2024-08-06 18:12:40,002 INFO [trainer.py:765] (4/8) Epoch 17, batch 1800, train_loss[loss=3.003, NarTop10Accuracy=0.7269, over 7110.00 frames. ], tot_loss[loss=3.191, NarTop10Accuracy=0.6874, over 5984.00 frames. ], batch size: 22, lr: 5.02e-03
2024-08-06 18:13:06,380 INFO [trainer.py:765] (4/8) Epoch 17, batch 1900, train_loss[loss=3.06, NarTop10Accuracy=0.7112, over 6234.00 frames. ], tot_loss[loss=3.194, NarTop10Accuracy=0.6864, over 6023.04 frames. ], batch size: 50, lr: 5.01e-03
2024-08-06 18:13:31,923 INFO [trainer.py:765] (4/8) Epoch 17, batch 2000, train_loss[loss=3.642, NarTop10Accuracy=0.5937, over 6303.00 frames. ], tot_loss[loss=3.171, NarTop10Accuracy=0.6916, over 5999.45 frames. ], batch size: 50, lr: 5.01e-03
2024-08-06 18:13:57,228 INFO [trainer.py:765] (4/8) Epoch 17, batch 2100, train_loss[loss=3.091, NarTop10Accuracy=0.7081, over 3918.00 frames. ], tot_loss[loss=3.178, NarTop10Accuracy=0.6895, over 5970.73 frames. ], batch size: 4, lr: 5.00e-03
2024-08-06 18:14:22,435 INFO [trainer.py:765] (4/8) Epoch 17, batch 2200, train_loss[loss=3.031, NarTop10Accuracy=0.7221, over 7464.00 frames. ], tot_loss[loss=3.194, NarTop10Accuracy=0.6863, over 5992.56 frames. ], batch size: 32, lr: 5.00e-03
2024-08-06 18:14:47,592 INFO [trainer.py:765] (4/8) Epoch 17, batch 2300, train_loss[loss=2.974, NarTop10Accuracy=0.7284, over 5556.00 frames. ], tot_loss[loss=3.187, NarTop10Accuracy=0.6879, over 6012.82 frames. ], batch size: 9, lr: 4.99e-03
2024-08-06 18:15:12,061 INFO [trainer.py:765] (4/8) Epoch 17, batch 2400, train_loss[loss=2.8, NarTop10Accuracy=0.7584, over 4938.00 frames. ], tot_loss[loss=3.181, NarTop10Accuracy=0.6892, over 5756.92 frames. ], batch size: 7, lr: 4.99e-03
2024-08-06 18:15:35,515 INFO [trainer.py:765] (4/8) Epoch 17, batch 2500, train_loss[loss=2.918, NarTop10Accuracy=0.7484, over 4989.00 frames. ], tot_loss[loss=3.165, NarTop10Accuracy=0.6922, over 5464.80 frames. ], batch size: 7, lr: 4.98e-03
2024-08-06 18:15:54,872 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 18:16:49,907 INFO [trainer.py:765] (4/8) Epoch 18, batch 100, train_loss[loss=3.029, NarTop10Accuracy=0.7226, over 7365.00 frames. ], tot_loss[loss=3.177, NarTop10Accuracy=0.6901, over 2368.58 frames. ], batch size: 31, lr: 4.83e-03
2024-08-06 18:17:24,749 INFO [trainer.py:765] (4/8) Epoch 18, batch 200, train_loss[loss=2.965, NarTop10Accuracy=0.7407, over 6648.00 frames. ], tot_loss[loss=3.161, NarTop10Accuracy=0.6936, over 3843.54 frames. ], batch size: 17, lr: 4.83e-03
2024-08-06 18:17:27,716 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 18:17:35,926 INFO [trainer.py:811] (4/8) Epoch 18, validation: loss=3.062, NarTop10Accuracy=0.7137, over 1905321.00 frames. 
2024-08-06 18:17:35,927 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 18:17:36,528 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.649e+02 2.024e+02 2.164e+02 2.334e+02 7.024e+02, threshold=4.329e+02, percent-clipped=0.1
2024-08-06 18:18:06,912 INFO [trainer.py:765] (4/8) Epoch 18, batch 300, train_loss[loss=3.472, NarTop10Accuracy=0.6257, over 6885.00 frames. ], tot_loss[loss=3.169, NarTop10Accuracy=0.6917, over 4635.62 frames. ], batch size: 22, lr: 4.82e-03
2024-08-06 18:18:38,183 INFO [trainer.py:765] (4/8) Epoch 18, batch 400, train_loss[loss=3.439, NarTop10Accuracy=0.6397, over 5064.00 frames. ], tot_loss[loss=3.159, NarTop10Accuracy=0.6942, over 5094.89 frames. ], batch size: 7, lr: 4.81e-03
2024-08-06 18:19:13,599 INFO [trainer.py:765] (4/8) Epoch 18, batch 500, train_loss[loss=3.169, NarTop10Accuracy=0.6886, over 6048.00 frames. ], tot_loss[loss=3.158, NarTop10Accuracy=0.6941, over 5374.70 frames. ], batch size: 11, lr: 4.81e-03
2024-08-06 18:19:48,151 INFO [trainer.py:765] (4/8) Epoch 18, batch 600, train_loss[loss=3.294, NarTop10Accuracy=0.6617, over 5679.00 frames. ], tot_loss[loss=3.143, NarTop10Accuracy=0.6968, over 5643.95 frames. ], batch size: 9, lr: 4.80e-03
2024-08-06 18:20:23,869 INFO [trainer.py:765] (4/8) Epoch 18, batch 700, train_loss[loss=3.568, NarTop10Accuracy=0.6091, over 5070.00 frames. ], tot_loss[loss=3.153, NarTop10Accuracy=0.6948, over 5713.82 frames. ], batch size: 6, lr: 4.80e-03
2024-08-06 18:21:01,026 INFO [trainer.py:765] (4/8) Epoch 18, batch 800, train_loss[loss=2.775, NarTop10Accuracy=0.7717, over 4959.00 frames. ], tot_loss[loss=3.165, NarTop10Accuracy=0.692, over 5776.07 frames. ], batch size: 6, lr: 4.79e-03
2024-08-06 18:21:32,408 INFO [trainer.py:765] (4/8) Epoch 18, batch 900, train_loss[loss=2.808, NarTop10Accuracy=0.7621, over 6369.00 frames. ], tot_loss[loss=3.143, NarTop10Accuracy=0.6963, over 5793.71 frames. ], batch size: 13, lr: 4.79e-03
2024-08-06 18:22:11,192 INFO [trainer.py:765] (4/8) Epoch 18, batch 1000, train_loss[loss=2.96, NarTop10Accuracy=0.7429, over 6702.00 frames. ], tot_loss[loss=3.16, NarTop10Accuracy=0.6932, over 5903.53 frames. ], batch size: 14, lr: 4.78e-03
2024-08-06 18:22:46,969 INFO [trainer.py:765] (4/8) Epoch 18, batch 1100, train_loss[loss=3.371, NarTop10Accuracy=0.65, over 6870.00 frames. ], tot_loss[loss=3.161, NarTop10Accuracy=0.6932, over 5936.97 frames. ], batch size: 17, lr: 4.78e-03
2024-08-06 18:23:18,605 INFO [trainer.py:765] (4/8) Epoch 18, batch 1200, train_loss[loss=3.548, NarTop10Accuracy=0.6113, over 7047.00 frames. ], tot_loss[loss=3.177, NarTop10Accuracy=0.6901, over 5916.77 frames. ], batch size: 31, lr: 4.77e-03
2024-08-06 18:24:00,099 INFO [trainer.py:765] (4/8) Epoch 18, batch 1300, train_loss[loss=2.829, NarTop10Accuracy=0.7589, over 4377.00 frames. ], tot_loss[loss=3.159, NarTop10Accuracy=0.6936, over 5976.56 frames. ], batch size: 5, lr: 4.77e-03
2024-08-06 18:24:29,574 INFO [trainer.py:765] (4/8) Epoch 18, batch 1400, train_loss[loss=2.922, NarTop10Accuracy=0.739, over 5976.00 frames. ], tot_loss[loss=3.157, NarTop10Accuracy=0.6943, over 6012.04 frames. ], batch size: 11, lr: 4.76e-03
2024-08-06 18:25:00,307 INFO [trainer.py:765] (4/8) Epoch 18, batch 1500, train_loss[loss=3.166, NarTop10Accuracy=0.6993, over 6912.00 frames. ], tot_loss[loss=3.15, NarTop10Accuracy=0.6959, over 5965.22 frames. ], batch size: 50, lr: 4.76e-03
2024-08-06 18:25:28,085 INFO [trainer.py:765] (4/8) Epoch 18, batch 1600, train_loss[loss=3.003, NarTop10Accuracy=0.7275, over 6993.00 frames. ], tot_loss[loss=3.163, NarTop10Accuracy=0.693, over 5943.60 frames. ], batch size: 22, lr: 4.75e-03
2024-08-06 18:25:54,687 INFO [trainer.py:765] (4/8) Epoch 18, batch 1700, train_loss[loss=3.045, NarTop10Accuracy=0.7206, over 6225.00 frames. ], tot_loss[loss=3.163, NarTop10Accuracy=0.6932, over 5909.36 frames. ], batch size: 13, lr: 4.75e-03
2024-08-06 18:26:21,196 INFO [trainer.py:765] (4/8) Epoch 18, batch 1800, train_loss[loss=3.471, NarTop10Accuracy=0.6277, over 7221.00 frames. ], tot_loss[loss=3.156, NarTop10Accuracy=0.6946, over 5971.48 frames. ], batch size: 22, lr: 4.74e-03
2024-08-06 18:26:47,567 INFO [trainer.py:765] (4/8) Epoch 18, batch 1900, train_loss[loss=3.151, NarTop10Accuracy=0.7008, over 5955.00 frames. ], tot_loss[loss=3.161, NarTop10Accuracy=0.6936, over 6011.46 frames. ], batch size: 50, lr: 4.74e-03
2024-08-06 18:27:13,176 INFO [trainer.py:765] (4/8) Epoch 18, batch 2000, train_loss[loss=3.123, NarTop10Accuracy=0.71, over 6390.00 frames. ], tot_loss[loss=3.159, NarTop10Accuracy=0.6937, over 6005.72 frames. ], batch size: 51, lr: 4.73e-03
2024-08-06 18:27:38,528 INFO [trainer.py:765] (4/8) Epoch 18, batch 2100, train_loss[loss=3.405, NarTop10Accuracy=0.6472, over 3927.00 frames. ], tot_loss[loss=3.154, NarTop10Accuracy=0.6952, over 5983.71 frames. ], batch size: 4, lr: 4.73e-03
2024-08-06 18:28:03,812 INFO [trainer.py:765] (4/8) Epoch 18, batch 2200, train_loss[loss=3.236, NarTop10Accuracy=0.6777, over 7302.00 frames. ], tot_loss[loss=3.16, NarTop10Accuracy=0.6938, over 6024.12 frames. ], batch size: 32, lr: 4.72e-03
2024-08-06 18:28:06,571 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 18:28:14,649 INFO [trainer.py:811] (4/8) Epoch 18, validation: loss=3.028, NarTop10Accuracy=0.7201, over 1905321.00 frames. 
2024-08-06 18:28:14,650 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 18:28:15,147 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.654e+02 2.054e+02 2.220e+02 2.384e+02 3.992e+02, threshold=4.441e+02, percent-clipped=0.0
2024-08-06 18:28:37,097 INFO [trainer.py:765] (4/8) Epoch 18, batch 2300, train_loss[loss=2.902, NarTop10Accuracy=0.7464, over 5775.00 frames. ], tot_loss[loss=3.178, NarTop10Accuracy=0.69, over 6027.29 frames. ], batch size: 9, lr: 4.72e-03
2024-08-06 18:29:01,592 INFO [trainer.py:765] (4/8) Epoch 18, batch 2400, train_loss[loss=2.992, NarTop10Accuracy=0.7298, over 5250.00 frames. ], tot_loss[loss=3.146, NarTop10Accuracy=0.6961, over 5778.92 frames. ], batch size: 7, lr: 4.71e-03
2024-08-06 18:29:25,027 INFO [trainer.py:765] (4/8) Epoch 18, batch 2500, train_loss[loss=2.952, NarTop10Accuracy=0.7342, over 5712.00 frames. ], tot_loss[loss=3.123, NarTop10Accuracy=0.7007, over 5470.41 frames. ], batch size: 8, lr: 4.71e-03
2024-08-06 18:29:45,425 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 18:30:41,231 INFO [trainer.py:765] (4/8) Epoch 19, batch 100, train_loss[loss=2.9, NarTop10Accuracy=0.7513, over 7290.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6921, over 2348.62 frames. ], batch size: 31, lr: 4.57e-03
2024-08-06 18:31:15,602 INFO [trainer.py:765] (4/8) Epoch 19, batch 200, train_loss[loss=2.925, NarTop10Accuracy=0.743, over 6936.00 frames. ], tot_loss[loss=3.16, NarTop10Accuracy=0.6948, over 3864.35 frames. ], batch size: 17, lr: 4.57e-03
2024-08-06 18:31:47,468 INFO [trainer.py:765] (4/8) Epoch 19, batch 300, train_loss[loss=3.624, NarTop10Accuracy=0.6026, over 7107.00 frames. ], tot_loss[loss=3.136, NarTop10Accuracy=0.6995, over 4659.53 frames. ], batch size: 22, lr: 4.56e-03
2024-08-06 18:32:20,355 INFO [trainer.py:765] (4/8) Epoch 19, batch 400, train_loss[loss=3.135, NarTop10Accuracy=0.7054, over 5139.00 frames. ], tot_loss[loss=3.137, NarTop10Accuracy=0.6991, over 5115.54 frames. ], batch size: 7, lr: 4.56e-03
2024-08-06 18:32:50,335 INFO [trainer.py:765] (4/8) Epoch 19, batch 500, train_loss[loss=3.04, NarTop10Accuracy=0.7218, over 6102.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6997, over 5376.75 frames. ], batch size: 11, lr: 4.55e-03
2024-08-06 18:33:29,609 INFO [trainer.py:765] (4/8) Epoch 19, batch 600, train_loss[loss=3.139, NarTop10Accuracy=0.702, over 5616.00 frames. ], tot_loss[loss=3.139, NarTop10Accuracy=0.6984, over 5639.53 frames. ], batch size: 9, lr: 4.55e-03
2024-08-06 18:34:03,592 INFO [trainer.py:765] (4/8) Epoch 19, batch 700, train_loss[loss=2.843, NarTop10Accuracy=0.7609, over 4947.00 frames. ], tot_loss[loss=3.153, NarTop10Accuracy=0.6955, over 5707.22 frames. ], batch size: 6, lr: 4.54e-03
2024-08-06 18:34:35,178 INFO [trainer.py:765] (4/8) Epoch 19, batch 800, train_loss[loss=3.217, NarTop10Accuracy=0.6856, over 5052.00 frames. ], tot_loss[loss=3.159, NarTop10Accuracy=0.6941, over 5760.31 frames. ], batch size: 6, lr: 4.54e-03
2024-08-06 18:35:10,262 INFO [trainer.py:765] (4/8) Epoch 19, batch 900, train_loss[loss=2.833, NarTop10Accuracy=0.766, over 6816.00 frames. ], tot_loss[loss=3.146, NarTop10Accuracy=0.6965, over 5784.60 frames. ], batch size: 14, lr: 4.53e-03
2024-08-06 18:35:48,637 INFO [trainer.py:765] (4/8) Epoch 19, batch 1000, train_loss[loss=3.341, NarTop10Accuracy=0.6549, over 6075.00 frames. ], tot_loss[loss=3.146, NarTop10Accuracy=0.6963, over 5893.46 frames. ], batch size: 13, lr: 4.53e-03
2024-08-06 18:36:20,938 INFO [trainer.py:765] (4/8) Epoch 19, batch 1100, train_loss[loss=2.935, NarTop10Accuracy=0.7461, over 6873.00 frames. ], tot_loss[loss=3.158, NarTop10Accuracy=0.6941, over 5937.71 frames. ], batch size: 17, lr: 4.52e-03
2024-08-06 18:36:57,130 INFO [trainer.py:765] (4/8) Epoch 19, batch 1200, train_loss[loss=2.996, NarTop10Accuracy=0.7321, over 7149.00 frames. ], tot_loss[loss=3.165, NarTop10Accuracy=0.6925, over 5914.13 frames. ], batch size: 31, lr: 4.52e-03
2024-08-06 18:37:35,314 INFO [trainer.py:765] (4/8) Epoch 19, batch 1300, train_loss[loss=2.848, NarTop10Accuracy=0.764, over 5058.00 frames. ], tot_loss[loss=3.165, NarTop10Accuracy=0.6927, over 5978.76 frames. ], batch size: 6, lr: 4.51e-03
2024-08-06 18:38:04,679 INFO [trainer.py:765] (4/8) Epoch 19, batch 1400, train_loss[loss=2.948, NarTop10Accuracy=0.7415, over 6204.00 frames. ], tot_loss[loss=3.167, NarTop10Accuracy=0.6922, over 6015.85 frames. ], batch size: 11, lr: 4.51e-03
2024-08-06 18:38:34,550 INFO [trainer.py:765] (4/8) Epoch 19, batch 1500, train_loss[loss=3.468, NarTop10Accuracy=0.6341, over 6342.00 frames. ], tot_loss[loss=3.149, NarTop10Accuracy=0.6963, over 5952.45 frames. ], batch size: 50, lr: 4.50e-03
2024-08-06 18:39:02,311 INFO [trainer.py:765] (4/8) Epoch 19, batch 1600, train_loss[loss=3.508, NarTop10Accuracy=0.6211, over 7083.00 frames. ], tot_loss[loss=3.143, NarTop10Accuracy=0.697, over 5937.18 frames. ], batch size: 22, lr: 4.50e-03
2024-08-06 18:39:11,590 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 18:39:19,795 INFO [trainer.py:811] (4/8) Epoch 19, validation: loss=2.958, NarTop10Accuracy=0.7345, over 1905321.00 frames. 
2024-08-06 18:39:19,796 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 18:39:20,378 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.633e+02 2.040e+02 2.194e+02 2.364e+02 6.410e+02, threshold=4.387e+02, percent-clipped=0.2
2024-08-06 18:39:37,191 INFO [trainer.py:765] (4/8) Epoch 19, batch 1700, train_loss[loss=3.584, NarTop10Accuracy=0.5954, over 6216.00 frames. ], tot_loss[loss=3.148, NarTop10Accuracy=0.6959, over 5920.54 frames. ], batch size: 13, lr: 4.49e-03
2024-08-06 18:40:03,789 INFO [trainer.py:765] (4/8) Epoch 19, batch 1800, train_loss[loss=3.492, NarTop10Accuracy=0.6226, over 7170.00 frames. ], tot_loss[loss=3.147, NarTop10Accuracy=0.696, over 5977.92 frames. ], batch size: 22, lr: 4.49e-03
2024-08-06 18:40:30,217 INFO [trainer.py:765] (4/8) Epoch 19, batch 1900, train_loss[loss=3.09, NarTop10Accuracy=0.7142, over 6000.00 frames. ], tot_loss[loss=3.143, NarTop10Accuracy=0.6971, over 6012.01 frames. ], batch size: 50, lr: 4.49e-03
2024-08-06 18:40:55,793 INFO [trainer.py:765] (4/8) Epoch 19, batch 2000, train_loss[loss=3.318, NarTop10Accuracy=0.6694, over 6228.00 frames. ], tot_loss[loss=3.147, NarTop10Accuracy=0.6965, over 5973.11 frames. ], batch size: 51, lr: 4.48e-03
2024-08-06 18:41:21,183 INFO [trainer.py:765] (4/8) Epoch 19, batch 2100, train_loss[loss=2.74, NarTop10Accuracy=0.7822, over 3834.00 frames. ], tot_loss[loss=3.135, NarTop10Accuracy=0.6989, over 5956.48 frames. ], batch size: 4, lr: 4.48e-03
2024-08-06 18:41:46,455 INFO [trainer.py:765] (4/8) Epoch 19, batch 2200, train_loss[loss=3.201, NarTop10Accuracy=0.6864, over 7083.00 frames. ], tot_loss[loss=3.147, NarTop10Accuracy=0.6965, over 5993.65 frames. ], batch size: 31, lr: 4.47e-03
2024-08-06 18:42:11,559 INFO [trainer.py:765] (4/8) Epoch 19, batch 2300, train_loss[loss=3.155, NarTop10Accuracy=0.6922, over 5673.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6916, over 6010.88 frames. ], batch size: 9, lr: 4.47e-03
2024-08-06 18:42:35,989 INFO [trainer.py:765] (4/8) Epoch 19, batch 2400, train_loss[loss=2.983, NarTop10Accuracy=0.7362, over 5187.00 frames. ], tot_loss[loss=3.151, NarTop10Accuracy=0.6957, over 5767.20 frames. ], batch size: 7, lr: 4.46e-03
2024-08-06 18:42:59,690 INFO [trainer.py:765] (4/8) Epoch 19, batch 2500, train_loss[loss=2.936, NarTop10Accuracy=0.7433, over 5244.00 frames. ], tot_loss[loss=3.126, NarTop10Accuracy=0.7003, over 5474.71 frames. ], batch size: 7, lr: 4.46e-03
2024-08-06 18:43:19,728 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 18:44:22,973 INFO [trainer.py:765] (4/8) Epoch 20, batch 100, train_loss[loss=3.157, NarTop10Accuracy=0.6848, over 7143.00 frames. ], tot_loss[loss=3.155, NarTop10Accuracy=0.6944, over 2374.53 frames. ], batch size: 31, lr: 4.34e-03
2024-08-06 18:44:58,379 INFO [trainer.py:765] (4/8) Epoch 20, batch 200, train_loss[loss=3.388, NarTop10Accuracy=0.6508, over 6732.00 frames. ], tot_loss[loss=3.124, NarTop10Accuracy=0.7016, over 3862.70 frames. ], batch size: 17, lr: 4.33e-03
2024-08-06 18:45:32,278 INFO [trainer.py:765] (4/8) Epoch 20, batch 300, train_loss[loss=3.448, NarTop10Accuracy=0.6391, over 7038.00 frames. ], tot_loss[loss=3.11, NarTop10Accuracy=0.7043, over 4657.07 frames. ], batch size: 22, lr: 4.33e-03
2024-08-06 18:46:05,128 INFO [trainer.py:765] (4/8) Epoch 20, batch 400, train_loss[loss=2.815, NarTop10Accuracy=0.7658, over 5145.00 frames. ], tot_loss[loss=3.114, NarTop10Accuracy=0.7037, over 5117.48 frames. ], batch size: 7, lr: 4.32e-03
2024-08-06 18:46:35,769 INFO [trainer.py:765] (4/8) Epoch 20, batch 500, train_loss[loss=2.898, NarTop10Accuracy=0.7467, over 6099.00 frames. ], tot_loss[loss=3.128, NarTop10Accuracy=0.7007, over 5380.02 frames. ], batch size: 11, lr: 4.32e-03
2024-08-06 18:47:13,255 INFO [trainer.py:765] (4/8) Epoch 20, batch 600, train_loss[loss=3.098, NarTop10Accuracy=0.7082, over 5730.00 frames. ], tot_loss[loss=3.128, NarTop10Accuracy=0.7008, over 5655.13 frames. ], batch size: 9, lr: 4.31e-03
2024-08-06 18:47:44,481 INFO [trainer.py:765] (4/8) Epoch 20, batch 700, train_loss[loss=2.929, NarTop10Accuracy=0.7545, over 4257.00 frames. ], tot_loss[loss=3.121, NarTop10Accuracy=0.7023, over 5731.19 frames. ], batch size: 5, lr: 4.31e-03
2024-08-06 18:48:21,015 INFO [trainer.py:765] (4/8) Epoch 20, batch 800, train_loss[loss=2.876, NarTop10Accuracy=0.7537, over 4263.00 frames. ], tot_loss[loss=3.135, NarTop10Accuracy=0.6992, over 5773.71 frames. ], batch size: 5, lr: 4.31e-03
2024-08-06 18:48:56,534 INFO [trainer.py:765] (4/8) Epoch 20, batch 900, train_loss[loss=2.941, NarTop10Accuracy=0.7384, over 6141.00 frames. ], tot_loss[loss=3.13, NarTop10Accuracy=0.7, over 5787.54 frames. ], batch size: 13, lr: 4.30e-03
2024-08-06 18:49:29,805 INFO [trainer.py:765] (4/8) Epoch 20, batch 1000, train_loss[loss=3.377, NarTop10Accuracy=0.6405, over 6696.00 frames. ], tot_loss[loss=3.158, NarTop10Accuracy=0.6941, over 5874.60 frames. ], batch size: 14, lr: 4.30e-03
2024-08-06 18:49:52,236 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 18:50:00,327 INFO [trainer.py:811] (4/8) Epoch 20, validation: loss=2.962, NarTop10Accuracy=0.7336, over 1905321.00 frames. 
2024-08-06 18:50:00,327 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 18:50:00,874 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.681e+02 2.061e+02 2.223e+02 2.401e+02 3.871e+02, threshold=4.447e+02, percent-clipped=0.0
2024-08-06 18:50:15,427 INFO [trainer.py:765] (4/8) Epoch 20, batch 1100, train_loss[loss=3.254, NarTop10Accuracy=0.6853, over 6855.00 frames. ], tot_loss[loss=3.149, NarTop10Accuracy=0.696, over 5928.47 frames. ], batch size: 17, lr: 4.29e-03
2024-08-06 18:50:53,775 INFO [trainer.py:765] (4/8) Epoch 20, batch 1200, train_loss[loss=2.978, NarTop10Accuracy=0.7352, over 7182.00 frames. ], tot_loss[loss=3.151, NarTop10Accuracy=0.6955, over 5939.76 frames. ], batch size: 31, lr: 4.29e-03
2024-08-06 18:51:25,129 INFO [trainer.py:765] (4/8) Epoch 20, batch 1300, train_loss[loss=3.438, NarTop10Accuracy=0.6375, over 4989.00 frames. ], tot_loss[loss=3.142, NarTop10Accuracy=0.6971, over 6013.11 frames. ], batch size: 6, lr: 4.29e-03
2024-08-06 18:51:59,314 INFO [trainer.py:765] (4/8) Epoch 20, batch 1400, train_loss[loss=2.934, NarTop10Accuracy=0.7284, over 6201.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6991, over 6028.56 frames. ], batch size: 11, lr: 4.28e-03
2024-08-06 18:52:32,805 INFO [trainer.py:765] (4/8) Epoch 20, batch 1500, train_loss[loss=3.314, NarTop10Accuracy=0.6715, over 6138.00 frames. ], tot_loss[loss=3.143, NarTop10Accuracy=0.6965, over 5950.47 frames. ], batch size: 50, lr: 4.28e-03
2024-08-06 18:53:00,635 INFO [trainer.py:765] (4/8) Epoch 20, batch 1600, train_loss[loss=2.902, NarTop10Accuracy=0.7565, over 7092.00 frames. ], tot_loss[loss=3.151, NarTop10Accuracy=0.6948, over 5933.29 frames. ], batch size: 22, lr: 4.27e-03
2024-08-06 18:53:27,328 INFO [trainer.py:765] (4/8) Epoch 20, batch 1700, train_loss[loss=3.378, NarTop10Accuracy=0.6451, over 6147.00 frames. ], tot_loss[loss=3.142, NarTop10Accuracy=0.697, over 5927.19 frames. ], batch size: 13, lr: 4.27e-03
2024-08-06 18:53:53,850 INFO [trainer.py:765] (4/8) Epoch 20, batch 1800, train_loss[loss=2.919, NarTop10Accuracy=0.7358, over 7104.00 frames. ], tot_loss[loss=3.129, NarTop10Accuracy=0.6996, over 5993.80 frames. ], batch size: 22, lr: 4.26e-03
2024-08-06 18:54:20,315 INFO [trainer.py:765] (4/8) Epoch 20, batch 1900, train_loss[loss=3.091, NarTop10Accuracy=0.7079, over 6321.00 frames. ], tot_loss[loss=3.158, NarTop10Accuracy=0.6938, over 6024.87 frames. ], batch size: 51, lr: 4.26e-03
2024-08-06 18:54:45,890 INFO [trainer.py:765] (4/8) Epoch 20, batch 2000, train_loss[loss=3.752, NarTop10Accuracy=0.5663, over 6015.00 frames. ], tot_loss[loss=3.161, NarTop10Accuracy=0.6934, over 6005.99 frames. ], batch size: 50, lr: 4.26e-03
2024-08-06 18:55:11,182 INFO [trainer.py:765] (4/8) Epoch 20, batch 2100, train_loss[loss=3.485, NarTop10Accuracy=0.615, over 3921.00 frames. ], tot_loss[loss=3.157, NarTop10Accuracy=0.6939, over 5978.23 frames. ], batch size: 4, lr: 4.25e-03
2024-08-06 18:55:36,414 INFO [trainer.py:765] (4/8) Epoch 20, batch 2200, train_loss[loss=2.856, NarTop10Accuracy=0.7518, over 7188.00 frames. ], tot_loss[loss=3.161, NarTop10Accuracy=0.6932, over 6012.55 frames. ], batch size: 31, lr: 4.25e-03
2024-08-06 18:56:01,635 INFO [trainer.py:765] (4/8) Epoch 20, batch 2300, train_loss[loss=3.383, NarTop10Accuracy=0.652, over 5721.00 frames. ], tot_loss[loss=3.172, NarTop10Accuracy=0.6911, over 6029.86 frames. ], batch size: 9, lr: 4.24e-03
2024-08-06 18:56:26,049 INFO [trainer.py:765] (4/8) Epoch 20, batch 2400, train_loss[loss=3.127, NarTop10Accuracy=0.6996, over 5070.00 frames. ], tot_loss[loss=3.155, NarTop10Accuracy=0.6942, over 5776.68 frames. ], batch size: 7, lr: 4.24e-03
2024-08-06 18:56:49,565 INFO [trainer.py:765] (4/8) Epoch 20, batch 2500, train_loss[loss=2.885, NarTop10Accuracy=0.7506, over 5163.00 frames. ], tot_loss[loss=3.118, NarTop10Accuracy=0.7017, over 5470.79 frames. ], batch size: 7, lr: 4.24e-03
2024-08-06 18:57:09,584 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 18:58:09,585 INFO [trainer.py:765] (4/8) Epoch 21, batch 100, train_loss[loss=3.205, NarTop10Accuracy=0.6843, over 7176.00 frames. ], tot_loss[loss=3.116, NarTop10Accuracy=0.703, over 2361.36 frames. ], batch size: 31, lr: 4.13e-03
2024-08-06 18:58:40,418 INFO [trainer.py:765] (4/8) Epoch 21, batch 200, train_loss[loss=2.817, NarTop10Accuracy=0.763, over 6792.00 frames. ], tot_loss[loss=3.127, NarTop10Accuracy=0.7008, over 3857.19 frames. ], batch size: 17, lr: 4.12e-03
2024-08-06 18:59:13,334 INFO [trainer.py:765] (4/8) Epoch 21, batch 300, train_loss[loss=2.929, NarTop10Accuracy=0.7434, over 7020.00 frames. ], tot_loss[loss=3.129, NarTop10Accuracy=0.7006, over 4649.72 frames. ], batch size: 22, lr: 4.12e-03
2024-08-06 18:59:48,151 INFO [trainer.py:765] (4/8) Epoch 21, batch 400, train_loss[loss=2.841, NarTop10Accuracy=0.7571, over 5136.00 frames. ], tot_loss[loss=3.11, NarTop10Accuracy=0.7042, over 5101.63 frames. ], batch size: 7, lr: 4.11e-03
2024-08-06 19:00:16,840 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:00:25,075 INFO [trainer.py:811] (4/8) Epoch 21, validation: loss=2.992, NarTop10Accuracy=0.7268, over 1905321.00 frames. 
2024-08-06 19:00:25,076 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:00:25,622 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.727e+02 2.071e+02 2.224e+02 2.387e+02 3.839e+02, threshold=4.447e+02, percent-clipped=0.0
2024-08-06 19:00:29,891 INFO [trainer.py:765] (4/8) Epoch 21, batch 500, train_loss[loss=2.881, NarTop10Accuracy=0.7474, over 6117.00 frames. ], tot_loss[loss=3.118, NarTop10Accuracy=0.7027, over 5387.79 frames. ], batch size: 11, lr: 4.11e-03
2024-08-06 19:01:03,328 INFO [trainer.py:765] (4/8) Epoch 21, batch 600, train_loss[loss=3.596, NarTop10Accuracy=0.6064, over 5739.00 frames. ], tot_loss[loss=3.107, NarTop10Accuracy=0.7047, over 5638.46 frames. ], batch size: 9, lr: 4.11e-03
2024-08-06 19:01:39,388 INFO [trainer.py:765] (4/8) Epoch 21, batch 700, train_loss[loss=2.804, NarTop10Accuracy=0.7592, over 4227.00 frames. ], tot_loss[loss=3.115, NarTop10Accuracy=0.7029, over 5708.79 frames. ], batch size: 5, lr: 4.10e-03
2024-08-06 19:02:18,047 INFO [trainer.py:765] (4/8) Epoch 21, batch 800, train_loss[loss=2.961, NarTop10Accuracy=0.7324, over 5154.00 frames. ], tot_loss[loss=3.135, NarTop10Accuracy=0.699, over 5767.45 frames. ], batch size: 6, lr: 4.10e-03
2024-08-06 19:02:48,663 INFO [trainer.py:765] (4/8) Epoch 21, batch 900, train_loss[loss=2.934, NarTop10Accuracy=0.7363, over 6138.00 frames. ], tot_loss[loss=3.134, NarTop10Accuracy=0.6992, over 5794.50 frames. ], batch size: 13, lr: 4.09e-03
2024-08-06 19:03:25,801 INFO [trainer.py:765] (4/8) Epoch 21, batch 1000, train_loss[loss=2.879, NarTop10Accuracy=0.7515, over 6780.00 frames. ], tot_loss[loss=3.133, NarTop10Accuracy=0.6994, over 5895.64 frames. ], batch size: 14, lr: 4.09e-03
2024-08-06 19:04:07,207 INFO [trainer.py:765] (4/8) Epoch 21, batch 1100, train_loss[loss=3.484, NarTop10Accuracy=0.6211, over 6786.00 frames. ], tot_loss[loss=3.15, NarTop10Accuracy=0.6958, over 5932.54 frames. ], batch size: 17, lr: 4.09e-03
2024-08-06 19:04:38,462 INFO [trainer.py:765] (4/8) Epoch 21, batch 1200, train_loss[loss=3.271, NarTop10Accuracy=0.6668, over 7185.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6994, over 5931.22 frames. ], batch size: 31, lr: 4.08e-03
2024-08-06 19:05:15,316 INFO [trainer.py:765] (4/8) Epoch 21, batch 1300, train_loss[loss=2.889, NarTop10Accuracy=0.7548, over 5115.00 frames. ], tot_loss[loss=3.115, NarTop10Accuracy=0.703, over 5998.88 frames. ], batch size: 6, lr: 4.08e-03
2024-08-06 19:05:55,559 INFO [trainer.py:765] (4/8) Epoch 21, batch 1400, train_loss[loss=3.391, NarTop10Accuracy=0.6456, over 6075.00 frames. ], tot_loss[loss=3.117, NarTop10Accuracy=0.7024, over 6018.86 frames. ], batch size: 11, lr: 4.07e-03
2024-08-06 19:06:23,600 INFO [trainer.py:765] (4/8) Epoch 21, batch 1500, train_loss[loss=3.315, NarTop10Accuracy=0.6599, over 6108.00 frames. ], tot_loss[loss=3.131, NarTop10Accuracy=0.6994, over 5957.13 frames. ], batch size: 50, lr: 4.07e-03
2024-08-06 19:06:51,462 INFO [trainer.py:765] (4/8) Epoch 21, batch 1600, train_loss[loss=2.988, NarTop10Accuracy=0.7368, over 7119.00 frames. ], tot_loss[loss=3.133, NarTop10Accuracy=0.6989, over 5924.21 frames. ], batch size: 22, lr: 4.07e-03
2024-08-06 19:07:18,212 INFO [trainer.py:765] (4/8) Epoch 21, batch 1700, train_loss[loss=3.273, NarTop10Accuracy=0.6708, over 6132.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6992, over 5898.75 frames. ], batch size: 13, lr: 4.06e-03
2024-08-06 19:07:44,809 INFO [trainer.py:765] (4/8) Epoch 21, batch 1800, train_loss[loss=2.732, NarTop10Accuracy=0.773, over 7023.00 frames. ], tot_loss[loss=3.126, NarTop10Accuracy=0.7003, over 5971.53 frames. ], batch size: 22, lr: 4.06e-03
2024-08-06 19:08:11,369 INFO [trainer.py:765] (4/8) Epoch 21, batch 1900, train_loss[loss=3.755, NarTop10Accuracy=0.5763, over 6399.00 frames. ], tot_loss[loss=3.138, NarTop10Accuracy=0.6979, over 6018.89 frames. ], batch size: 52, lr: 4.06e-03
2024-08-06 19:08:37,105 INFO [trainer.py:765] (4/8) Epoch 21, batch 2000, train_loss[loss=3.451, NarTop10Accuracy=0.6276, over 6315.00 frames. ], tot_loss[loss=3.131, NarTop10Accuracy=0.6989, over 6001.00 frames. ], batch size: 50, lr: 4.05e-03
2024-08-06 19:09:02,507 INFO [trainer.py:765] (4/8) Epoch 21, batch 2100, train_loss[loss=2.952, NarTop10Accuracy=0.7333, over 4815.00 frames. ], tot_loss[loss=3.137, NarTop10Accuracy=0.6973, over 5984.16 frames. ], batch size: 5, lr: 4.05e-03
2024-08-06 19:09:27,891 INFO [trainer.py:765] (4/8) Epoch 21, batch 2200, train_loss[loss=2.959, NarTop10Accuracy=0.7458, over 7419.00 frames. ], tot_loss[loss=3.14, NarTop10Accuracy=0.6969, over 6010.87 frames. ], batch size: 33, lr: 4.04e-03
2024-08-06 19:09:53,223 INFO [trainer.py:765] (4/8) Epoch 21, batch 2300, train_loss[loss=2.991, NarTop10Accuracy=0.7255, over 5661.00 frames. ], tot_loss[loss=3.155, NarTop10Accuracy=0.6942, over 6024.63 frames. ], batch size: 9, lr: 4.04e-03
2024-08-06 19:10:17,597 INFO [trainer.py:765] (4/8) Epoch 21, batch 2400, train_loss[loss=3.515, NarTop10Accuracy=0.6157, over 5055.00 frames. ], tot_loss[loss=3.14, NarTop10Accuracy=0.6971, over 5769.72 frames. ], batch size: 7, lr: 4.04e-03
2024-08-06 19:10:37,229 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:10:45,275 INFO [trainer.py:811] (4/8) Epoch 21, validation: loss=2.971, NarTop10Accuracy=0.7316, over 1905321.00 frames. 
2024-08-06 19:10:45,276 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:10:45,741 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.703e+02 2.100e+02 2.242e+02 2.407e+02 6.546e+02, threshold=4.484e+02, percent-clipped=0.1
2024-08-06 19:10:49,273 INFO [trainer.py:765] (4/8) Epoch 21, batch 2500, train_loss[loss=3.447, NarTop10Accuracy=0.6289, over 5151.00 frames. ], tot_loss[loss=3.101, NarTop10Accuracy=0.7046, over 5460.18 frames. ], batch size: 7, lr: 4.03e-03
2024-08-06 19:11:08,907 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 19:12:09,053 INFO [trainer.py:765] (4/8) Epoch 22, batch 100, train_loss[loss=2.879, NarTop10Accuracy=0.753, over 7245.00 frames. ], tot_loss[loss=3.093, NarTop10Accuracy=0.7074, over 2372.70 frames. ], batch size: 31, lr: 3.93e-03
2024-08-06 19:12:44,462 INFO [trainer.py:765] (4/8) Epoch 22, batch 200, train_loss[loss=3.181, NarTop10Accuracy=0.688, over 6723.00 frames. ], tot_loss[loss=3.094, NarTop10Accuracy=0.7069, over 3860.46 frames. ], batch size: 17, lr: 3.93e-03
2024-08-06 19:13:14,533 INFO [trainer.py:765] (4/8) Epoch 22, batch 300, train_loss[loss=3.021, NarTop10Accuracy=0.7251, over 7065.00 frames. ], tot_loss[loss=3.09, NarTop10Accuracy=0.7078, over 4658.66 frames. ], batch size: 22, lr: 3.93e-03
2024-08-06 19:13:49,228 INFO [trainer.py:765] (4/8) Epoch 22, batch 400, train_loss[loss=3.005, NarTop10Accuracy=0.7221, over 5157.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7079, over 5103.19 frames. ], batch size: 7, lr: 3.92e-03
2024-08-06 19:14:24,850 INFO [trainer.py:765] (4/8) Epoch 22, batch 500, train_loss[loss=3.065, NarTop10Accuracy=0.7073, over 6096.00 frames. ], tot_loss[loss=3.092, NarTop10Accuracy=0.7072, over 5383.89 frames. ], batch size: 11, lr: 3.92e-03
2024-08-06 19:14:55,701 INFO [trainer.py:765] (4/8) Epoch 22, batch 600, train_loss[loss=3.133, NarTop10Accuracy=0.7052, over 5820.00 frames. ], tot_loss[loss=3.128, NarTop10Accuracy=0.7004, over 5631.22 frames. ], batch size: 9, lr: 3.92e-03
2024-08-06 19:15:30,867 INFO [trainer.py:765] (4/8) Epoch 22, batch 700, train_loss[loss=3.287, NarTop10Accuracy=0.6621, over 5043.00 frames. ], tot_loss[loss=3.124, NarTop10Accuracy=0.7011, over 5717.54 frames. ], batch size: 6, lr: 3.91e-03
2024-08-06 19:16:10,664 INFO [trainer.py:765] (4/8) Epoch 22, batch 800, train_loss[loss=3.062, NarTop10Accuracy=0.7217, over 4416.00 frames. ], tot_loss[loss=3.111, NarTop10Accuracy=0.7034, over 5796.07 frames. ], batch size: 5, lr: 3.91e-03
2024-08-06 19:16:40,952 INFO [trainer.py:765] (4/8) Epoch 22, batch 900, train_loss[loss=2.986, NarTop10Accuracy=0.7223, over 6717.00 frames. ], tot_loss[loss=3.116, NarTop10Accuracy=0.7023, over 5807.46 frames. ], batch size: 14, lr: 3.90e-03
2024-08-06 19:17:16,433 INFO [trainer.py:765] (4/8) Epoch 22, batch 1000, train_loss[loss=3.067, NarTop10Accuracy=0.7107, over 6681.00 frames. ], tot_loss[loss=3.11, NarTop10Accuracy=0.7035, over 5905.43 frames. ], batch size: 14, lr: 3.90e-03
2024-08-06 19:17:52,085 INFO [trainer.py:765] (4/8) Epoch 22, batch 1100, train_loss[loss=2.958, NarTop10Accuracy=0.7383, over 6885.00 frames. ], tot_loss[loss=3.114, NarTop10Accuracy=0.7024, over 5940.31 frames. ], batch size: 17, lr: 3.90e-03
2024-08-06 19:18:25,926 INFO [trainer.py:765] (4/8) Epoch 22, batch 1200, train_loss[loss=2.923, NarTop10Accuracy=0.7381, over 7122.00 frames. ], tot_loss[loss=3.096, NarTop10Accuracy=0.7064, over 5945.16 frames. ], batch size: 31, lr: 3.89e-03
2024-08-06 19:19:01,252 INFO [trainer.py:765] (4/8) Epoch 22, batch 1300, train_loss[loss=3.004, NarTop10Accuracy=0.7245, over 5013.00 frames. ], tot_loss[loss=3.09, NarTop10Accuracy=0.7075, over 6001.98 frames. ], batch size: 6, lr: 3.89e-03
2024-08-06 19:19:33,317 INFO [trainer.py:765] (4/8) Epoch 22, batch 1400, train_loss[loss=2.851, NarTop10Accuracy=0.7603, over 6102.00 frames. ], tot_loss[loss=3.111, NarTop10Accuracy=0.7033, over 6000.56 frames. ], batch size: 11, lr: 3.89e-03
2024-08-06 19:20:03,830 INFO [trainer.py:765] (4/8) Epoch 22, batch 1500, train_loss[loss=3.557, NarTop10Accuracy=0.6203, over 6453.00 frames. ], tot_loss[loss=3.107, NarTop10Accuracy=0.704, over 5942.42 frames. ], batch size: 52, lr: 3.88e-03
2024-08-06 19:20:31,646 INFO [trainer.py:765] (4/8) Epoch 22, batch 1600, train_loss[loss=3.089, NarTop10Accuracy=0.7136, over 7086.00 frames. ], tot_loss[loss=3.126, NarTop10Accuracy=0.7, over 5937.22 frames. ], batch size: 22, lr: 3.88e-03
2024-08-06 19:20:58,417 INFO [trainer.py:765] (4/8) Epoch 22, batch 1700, train_loss[loss=3.179, NarTop10Accuracy=0.6955, over 6303.00 frames. ], tot_loss[loss=3.12, NarTop10Accuracy=0.7018, over 5935.38 frames. ], batch size: 13, lr: 3.88e-03
2024-08-06 19:21:25,010 INFO [trainer.py:765] (4/8) Epoch 22, batch 1800, train_loss[loss=2.825, NarTop10Accuracy=0.7637, over 6930.00 frames. ], tot_loss[loss=3.114, NarTop10Accuracy=0.703, over 5990.46 frames. ], batch size: 22, lr: 3.87e-03
2024-08-06 19:21:51,371 INFO [trainer.py:765] (4/8) Epoch 22, batch 1900, train_loss[loss=3.095, NarTop10Accuracy=0.707, over 5583.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6989, over 6037.43 frames. ], batch size: 50, lr: 3.87e-03
2024-08-06 19:21:53,109 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:22:01,088 INFO [trainer.py:811] (4/8) Epoch 22, validation: loss=3.009, NarTop10Accuracy=0.7241, over 1905321.00 frames. 
2024-08-06 19:22:01,089 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:22:01,575 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.670e+02 2.114e+02 2.276e+02 2.445e+02 4.438e+02, threshold=4.551e+02, percent-clipped=0.0
2024-08-06 19:22:24,818 INFO [trainer.py:765] (4/8) Epoch 22, batch 2000, train_loss[loss=3.448, NarTop10Accuracy=0.6338, over 5925.00 frames. ], tot_loss[loss=3.113, NarTop10Accuracy=0.7029, over 5999.15 frames. ], batch size: 50, lr: 3.87e-03
2024-08-06 19:22:50,041 INFO [trainer.py:765] (4/8) Epoch 22, batch 2100, train_loss[loss=3.163, NarTop10Accuracy=0.6882, over 3873.00 frames. ], tot_loss[loss=3.107, NarTop10Accuracy=0.7044, over 5978.72 frames. ], batch size: 4, lr: 3.86e-03
2024-08-06 19:23:15,230 INFO [trainer.py:765] (4/8) Epoch 22, batch 2200, train_loss[loss=3.002, NarTop10Accuracy=0.7307, over 7191.00 frames. ], tot_loss[loss=3.111, NarTop10Accuracy=0.7034, over 6018.00 frames. ], batch size: 31, lr: 3.86e-03
2024-08-06 19:23:40,314 INFO [trainer.py:765] (4/8) Epoch 22, batch 2300, train_loss[loss=3.162, NarTop10Accuracy=0.6926, over 5694.00 frames. ], tot_loss[loss=3.132, NarTop10Accuracy=0.6995, over 6024.27 frames. ], batch size: 9, lr: 3.86e-03
2024-08-06 19:24:04,602 INFO [trainer.py:765] (4/8) Epoch 22, batch 2400, train_loss[loss=3.111, NarTop10Accuracy=0.6908, over 5178.00 frames. ], tot_loss[loss=3.118, NarTop10Accuracy=0.7016, over 5772.89 frames. ], batch size: 7, lr: 3.85e-03
2024-08-06 19:24:28,024 INFO [trainer.py:765] (4/8) Epoch 22, batch 2500, train_loss[loss=3.134, NarTop10Accuracy=0.6905, over 5106.00 frames. ], tot_loss[loss=3.103, NarTop10Accuracy=0.7042, over 5464.66 frames. ], batch size: 7, lr: 3.85e-03
2024-08-06 19:24:47,502 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 19:25:45,385 INFO [trainer.py:765] (4/8) Epoch 23, batch 100, train_loss[loss=3.086, NarTop10Accuracy=0.7069, over 7146.00 frames. ], tot_loss[loss=3.125, NarTop10Accuracy=0.6995, over 2369.78 frames. ], batch size: 31, lr: 3.76e-03
2024-08-06 19:26:21,308 INFO [trainer.py:765] (4/8) Epoch 23, batch 200, train_loss[loss=3.432, NarTop10Accuracy=0.6437, over 6807.00 frames. ], tot_loss[loss=3.127, NarTop10Accuracy=0.7004, over 3846.47 frames. ], batch size: 17, lr: 3.76e-03
2024-08-06 19:26:57,602 INFO [trainer.py:765] (4/8) Epoch 23, batch 300, train_loss[loss=3.039, NarTop10Accuracy=0.724, over 7047.00 frames. ], tot_loss[loss=3.101, NarTop10Accuracy=0.7057, over 4653.10 frames. ], batch size: 22, lr: 3.75e-03
2024-08-06 19:27:26,540 INFO [trainer.py:765] (4/8) Epoch 23, batch 400, train_loss[loss=3.333, NarTop10Accuracy=0.6584, over 5310.00 frames. ], tot_loss[loss=3.115, NarTop10Accuracy=0.7031, over 5120.45 frames. ], batch size: 7, lr: 3.75e-03
2024-08-06 19:27:59,712 INFO [trainer.py:765] (4/8) Epoch 23, batch 500, train_loss[loss=3.452, NarTop10Accuracy=0.6294, over 6096.00 frames. ], tot_loss[loss=3.12, NarTop10Accuracy=0.7017, over 5390.83 frames. ], batch size: 11, lr: 3.75e-03
2024-08-06 19:28:35,883 INFO [trainer.py:765] (4/8) Epoch 23, batch 600, train_loss[loss=3.096, NarTop10Accuracy=0.6956, over 5733.00 frames. ], tot_loss[loss=3.11, NarTop10Accuracy=0.7037, over 5658.62 frames. ], batch size: 9, lr: 3.74e-03
2024-08-06 19:29:11,367 INFO [trainer.py:765] (4/8) Epoch 23, batch 700, train_loss[loss=2.988, NarTop10Accuracy=0.7081, over 5022.00 frames. ], tot_loss[loss=3.091, NarTop10Accuracy=0.7076, over 5730.31 frames. ], batch size: 6, lr: 3.74e-03
2024-08-06 19:29:43,612 INFO [trainer.py:765] (4/8) Epoch 23, batch 800, train_loss[loss=2.571, NarTop10Accuracy=0.8085, over 5208.00 frames. ], tot_loss[loss=3.101, NarTop10Accuracy=0.7058, over 5781.01 frames. ], batch size: 6, lr: 3.74e-03
2024-08-06 19:30:19,390 INFO [trainer.py:765] (4/8) Epoch 23, batch 900, train_loss[loss=3.338, NarTop10Accuracy=0.6499, over 6201.00 frames. ], tot_loss[loss=3.097, NarTop10Accuracy=0.706, over 5808.74 frames. ], batch size: 13, lr: 3.73e-03
2024-08-06 19:30:58,195 INFO [trainer.py:765] (4/8) Epoch 23, batch 1000, train_loss[loss=2.973, NarTop10Accuracy=0.7267, over 6081.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7078, over 5912.17 frames. ], batch size: 13, lr: 3.73e-03
2024-08-06 19:31:31,520 INFO [trainer.py:765] (4/8) Epoch 23, batch 1100, train_loss[loss=2.97, NarTop10Accuracy=0.7267, over 6936.00 frames. ], tot_loss[loss=3.095, NarTop10Accuracy=0.7068, over 5945.26 frames. ], batch size: 17, lr: 3.73e-03
2024-08-06 19:32:08,517 INFO [trainer.py:765] (4/8) Epoch 23, batch 1200, train_loss[loss=3.041, NarTop10Accuracy=0.7182, over 7545.00 frames. ], tot_loss[loss=3.109, NarTop10Accuracy=0.7036, over 5942.40 frames. ], batch size: 31, lr: 3.72e-03
2024-08-06 19:32:46,937 INFO [trainer.py:765] (4/8) Epoch 23, batch 1300, train_loss[loss=3.12, NarTop10Accuracy=0.7042, over 5184.00 frames. ], tot_loss[loss=3.114, NarTop10Accuracy=0.7026, over 6013.04 frames. ], batch size: 6, lr: 3.72e-03
2024-08-06 19:32:56,403 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:33:04,722 INFO [trainer.py:811] (4/8) Epoch 23, validation: loss=2.893, NarTop10Accuracy=0.7468, over 1905321.00 frames. 
2024-08-06 19:33:04,723 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:33:05,262 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.759e+02 2.108e+02 2.273e+02 2.457e+02 3.966e+02, threshold=4.546e+02, percent-clipped=0.0
2024-08-06 19:33:27,407 INFO [trainer.py:765] (4/8) Epoch 23, batch 1400, train_loss[loss=2.755, NarTop10Accuracy=0.7746, over 6051.00 frames. ], tot_loss[loss=3.113, NarTop10Accuracy=0.7024, over 6015.56 frames. ], batch size: 11, lr: 3.72e-03
2024-08-06 19:33:58,215 INFO [trainer.py:765] (4/8) Epoch 23, batch 1500, train_loss[loss=3.243, NarTop10Accuracy=0.68, over 6009.00 frames. ], tot_loss[loss=3.1, NarTop10Accuracy=0.7055, over 5965.01 frames. ], batch size: 51, lr: 3.71e-03
2024-08-06 19:34:26,015 INFO [trainer.py:765] (4/8) Epoch 23, batch 1600, train_loss[loss=2.921, NarTop10Accuracy=0.7523, over 7125.00 frames. ], tot_loss[loss=3.112, NarTop10Accuracy=0.7036, over 5936.58 frames. ], batch size: 22, lr: 3.71e-03
2024-08-06 19:34:52,782 INFO [trainer.py:765] (4/8) Epoch 23, batch 1700, train_loss[loss=3.242, NarTop10Accuracy=0.6812, over 6138.00 frames. ], tot_loss[loss=3.128, NarTop10Accuracy=0.6999, over 5906.50 frames. ], batch size: 13, lr: 3.71e-03
2024-08-06 19:35:19,261 INFO [trainer.py:765] (4/8) Epoch 23, batch 1800, train_loss[loss=2.925, NarTop10Accuracy=0.7395, over 7185.00 frames. ], tot_loss[loss=3.122, NarTop10Accuracy=0.7014, over 5972.12 frames. ], batch size: 22, lr: 3.70e-03
2024-08-06 19:35:45,596 INFO [trainer.py:765] (4/8) Epoch 23, batch 1900, train_loss[loss=3.453, NarTop10Accuracy=0.6272, over 5844.00 frames. ], tot_loss[loss=3.13, NarTop10Accuracy=0.6994, over 6012.01 frames. ], batch size: 50, lr: 3.70e-03
2024-08-06 19:36:11,170 INFO [trainer.py:765] (4/8) Epoch 23, batch 2000, train_loss[loss=3.636, NarTop10Accuracy=0.5919, over 5745.00 frames. ], tot_loss[loss=3.112, NarTop10Accuracy=0.7032, over 6002.16 frames. ], batch size: 50, lr: 3.70e-03
2024-08-06 19:36:36,517 INFO [trainer.py:765] (4/8) Epoch 23, batch 2100, train_loss[loss=3.155, NarTop10Accuracy=0.6875, over 4869.00 frames. ], tot_loss[loss=3.116, NarTop10Accuracy=0.7026, over 5981.68 frames. ], batch size: 5, lr: 3.69e-03
2024-08-06 19:37:01,908 INFO [trainer.py:765] (4/8) Epoch 23, batch 2200, train_loss[loss=3.105, NarTop10Accuracy=0.6991, over 7353.00 frames. ], tot_loss[loss=3.128, NarTop10Accuracy=0.7003, over 6018.75 frames. ], batch size: 31, lr: 3.69e-03
2024-08-06 19:37:27,060 INFO [trainer.py:765] (4/8) Epoch 23, batch 2300, train_loss[loss=3.082, NarTop10Accuracy=0.7067, over 5706.00 frames. ], tot_loss[loss=3.122, NarTop10Accuracy=0.7015, over 6030.69 frames. ], batch size: 9, lr: 3.69e-03
2024-08-06 19:37:51,424 INFO [trainer.py:765] (4/8) Epoch 23, batch 2400, train_loss[loss=3.006, NarTop10Accuracy=0.7267, over 5187.00 frames. ], tot_loss[loss=3.12, NarTop10Accuracy=0.7011, over 5774.99 frames. ], batch size: 7, lr: 3.69e-03
2024-08-06 19:38:15,052 INFO [trainer.py:765] (4/8) Epoch 23, batch 2500, train_loss[loss=3.241, NarTop10Accuracy=0.6763, over 5367.00 frames. ], tot_loss[loss=3.102, NarTop10Accuracy=0.7048, over 5472.77 frames. ], batch size: 7, lr: 3.68e-03
2024-08-06 19:38:35,167 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 19:39:37,632 INFO [trainer.py:765] (4/8) Epoch 24, batch 100, train_loss[loss=3.519, NarTop10Accuracy=0.6201, over 7362.00 frames. ], tot_loss[loss=3.122, NarTop10Accuracy=0.7021, over 2367.63 frames. ], batch size: 31, lr: 3.60e-03
2024-08-06 19:40:10,190 INFO [trainer.py:765] (4/8) Epoch 24, batch 200, train_loss[loss=2.892, NarTop10Accuracy=0.7499, over 6792.00 frames. ], tot_loss[loss=3.107, NarTop10Accuracy=0.7051, over 3854.13 frames. ], batch size: 17, lr: 3.60e-03
2024-08-06 19:40:40,556 INFO [trainer.py:765] (4/8) Epoch 24, batch 300, train_loss[loss=2.902, NarTop10Accuracy=0.7481, over 6957.00 frames. ], tot_loss[loss=3.094, NarTop10Accuracy=0.7071, over 4654.29 frames. ], batch size: 22, lr: 3.59e-03
2024-08-06 19:41:18,234 INFO [trainer.py:765] (4/8) Epoch 24, batch 400, train_loss[loss=2.896, NarTop10Accuracy=0.7516, over 5121.00 frames. ], tot_loss[loss=3.097, NarTop10Accuracy=0.7067, over 5102.44 frames. ], batch size: 7, lr: 3.59e-03
2024-08-06 19:41:50,323 INFO [trainer.py:765] (4/8) Epoch 24, batch 500, train_loss[loss=2.859, NarTop10Accuracy=0.7547, over 6012.00 frames. ], tot_loss[loss=3.087, NarTop10Accuracy=0.7089, over 5366.43 frames. ], batch size: 11, lr: 3.59e-03
2024-08-06 19:42:21,453 INFO [trainer.py:765] (4/8) Epoch 24, batch 600, train_loss[loss=2.767, NarTop10Accuracy=0.7793, over 5820.00 frames. ], tot_loss[loss=3.081, NarTop10Accuracy=0.7099, over 5638.91 frames. ], batch size: 9, lr: 3.58e-03
2024-08-06 19:42:52,843 INFO [trainer.py:765] (4/8) Epoch 24, batch 700, train_loss[loss=2.938, NarTop10Accuracy=0.7404, over 5082.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7096, over 5733.59 frames. ], batch size: 6, lr: 3.58e-03
2024-08-06 19:43:17,381 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:43:25,410 INFO [trainer.py:811] (4/8) Epoch 24, validation: loss=3.021, NarTop10Accuracy=0.7204, over 1905321.00 frames. 
2024-08-06 19:43:25,411 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:43:28,563 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.744e+02 2.113e+02 2.282e+02 2.472e+02 2.357e+03, threshold=4.564e+02, percent-clipped=0.2
2024-08-06 19:43:40,815 INFO [trainer.py:765] (4/8) Epoch 24, batch 800, train_loss[loss=2.692, NarTop10Accuracy=0.7914, over 4347.00 frames. ], tot_loss[loss=3.083, NarTop10Accuracy=0.7089, over 5790.68 frames. ], batch size: 5, lr: 3.58e-03
2024-08-06 19:44:11,410 INFO [trainer.py:765] (4/8) Epoch 24, batch 900, train_loss[loss=2.895, NarTop10Accuracy=0.7525, over 6243.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7092, over 5809.37 frames. ], batch size: 13, lr: 3.57e-03
2024-08-06 19:44:47,490 INFO [trainer.py:765] (4/8) Epoch 24, batch 1000, train_loss[loss=3.213, NarTop10Accuracy=0.6809, over 6690.00 frames. ], tot_loss[loss=3.095, NarTop10Accuracy=0.7064, over 5916.91 frames. ], batch size: 14, lr: 3.57e-03
2024-08-06 19:45:27,108 INFO [trainer.py:765] (4/8) Epoch 24, batch 1100, train_loss[loss=3.54, NarTop10Accuracy=0.6127, over 6744.00 frames. ], tot_loss[loss=3.108, NarTop10Accuracy=0.7039, over 5931.46 frames. ], batch size: 17, lr: 3.57e-03
2024-08-06 19:45:58,438 INFO [trainer.py:765] (4/8) Epoch 24, batch 1200, train_loss[loss=2.998, NarTop10Accuracy=0.7293, over 7824.00 frames. ], tot_loss[loss=3.102, NarTop10Accuracy=0.7053, over 5923.96 frames. ], batch size: 33, lr: 3.57e-03
2024-08-06 19:46:30,295 INFO [trainer.py:765] (4/8) Epoch 24, batch 1300, train_loss[loss=3.398, NarTop10Accuracy=0.6415, over 5130.00 frames. ], tot_loss[loss=3.096, NarTop10Accuracy=0.7065, over 5999.08 frames. ], batch size: 6, lr: 3.56e-03
2024-08-06 19:47:07,860 INFO [trainer.py:765] (4/8) Epoch 24, batch 1400, train_loss[loss=3.297, NarTop10Accuracy=0.6608, over 6051.00 frames. ], tot_loss[loss=3.104, NarTop10Accuracy=0.7041, over 6022.05 frames. ], batch size: 11, lr: 3.56e-03
2024-08-06 19:47:40,957 INFO [trainer.py:765] (4/8) Epoch 24, batch 1500, train_loss[loss=3.294, NarTop10Accuracy=0.6601, over 5769.00 frames. ], tot_loss[loss=3.118, NarTop10Accuracy=0.7015, over 5955.75 frames. ], batch size: 50, lr: 3.56e-03
2024-08-06 19:48:08,676 INFO [trainer.py:765] (4/8) Epoch 24, batch 1600, train_loss[loss=3.378, NarTop10Accuracy=0.648, over 7017.00 frames. ], tot_loss[loss=3.122, NarTop10Accuracy=0.7009, over 5931.66 frames. ], batch size: 22, lr: 3.55e-03
2024-08-06 19:48:35,269 INFO [trainer.py:765] (4/8) Epoch 24, batch 1700, train_loss[loss=2.978, NarTop10Accuracy=0.7344, over 6357.00 frames. ], tot_loss[loss=3.123, NarTop10Accuracy=0.7011, over 5920.12 frames. ], batch size: 13, lr: 3.55e-03
2024-08-06 19:49:01,638 INFO [trainer.py:765] (4/8) Epoch 24, batch 1800, train_loss[loss=2.837, NarTop10Accuracy=0.759, over 7002.00 frames. ], tot_loss[loss=3.13, NarTop10Accuracy=0.6999, over 5983.59 frames. ], batch size: 22, lr: 3.55e-03
2024-08-06 19:49:28,043 INFO [trainer.py:765] (4/8) Epoch 24, batch 1900, train_loss[loss=3.542, NarTop10Accuracy=0.6165, over 6057.00 frames. ], tot_loss[loss=3.138, NarTop10Accuracy=0.6979, over 6037.89 frames. ], batch size: 51, lr: 3.55e-03
2024-08-06 19:49:53,533 INFO [trainer.py:765] (4/8) Epoch 24, batch 2000, train_loss[loss=3.427, NarTop10Accuracy=0.6304, over 6171.00 frames. ], tot_loss[loss=3.112, NarTop10Accuracy=0.7031, over 6004.46 frames. ], batch size: 50, lr: 3.54e-03
2024-08-06 19:50:18,820 INFO [trainer.py:765] (4/8) Epoch 24, batch 2100, train_loss[loss=2.915, NarTop10Accuracy=0.7425, over 3894.00 frames. ], tot_loss[loss=3.108, NarTop10Accuracy=0.7036, over 5986.14 frames. ], batch size: 4, lr: 3.54e-03
2024-08-06 19:50:43,942 INFO [trainer.py:765] (4/8) Epoch 24, batch 2200, train_loss[loss=3.525, NarTop10Accuracy=0.612, over 7332.00 frames. ], tot_loss[loss=3.111, NarTop10Accuracy=0.7031, over 6011.51 frames. ], batch size: 31, lr: 3.54e-03
2024-08-06 19:51:09,024 INFO [trainer.py:765] (4/8) Epoch 24, batch 2300, train_loss[loss=2.893, NarTop10Accuracy=0.7436, over 5874.00 frames. ], tot_loss[loss=3.107, NarTop10Accuracy=0.7041, over 6017.27 frames. ], batch size: 9, lr: 3.53e-03
2024-08-06 19:51:33,349 INFO [trainer.py:765] (4/8) Epoch 24, batch 2400, train_loss[loss=3.091, NarTop10Accuracy=0.6994, over 5121.00 frames. ], tot_loss[loss=3.098, NarTop10Accuracy=0.7061, over 5784.23 frames. ], batch size: 7, lr: 3.53e-03
2024-08-06 19:51:56,783 INFO [trainer.py:765] (4/8) Epoch 24, batch 2500, train_loss[loss=2.914, NarTop10Accuracy=0.7364, over 5202.00 frames. ], tot_loss[loss=3.074, NarTop10Accuracy=0.7108, over 5496.14 frames. ], batch size: 7, lr: 3.53e-03
2024-08-06 19:52:16,696 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 19:53:22,198 INFO [trainer.py:765] (4/8) Epoch 25, batch 100, train_loss[loss=3.347, NarTop10Accuracy=0.6591, over 7377.00 frames. ], tot_loss[loss=3.088, NarTop10Accuracy=0.7077, over 2373.68 frames. ], batch size: 31, lr: 3.45e-03
2024-08-06 19:53:47,262 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 19:53:55,329 INFO [trainer.py:811] (4/8) Epoch 25, validation: loss=2.96, NarTop10Accuracy=0.7332, over 1905321.00 frames. 
2024-08-06 19:53:55,329 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 19:53:55,916 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.693e+02 2.155e+02 2.306e+02 2.475e+02 6.485e+02, threshold=4.611e+02, percent-clipped=0.1
2024-08-06 19:54:01,177 INFO [trainer.py:765] (4/8) Epoch 25, batch 200, train_loss[loss=2.975, NarTop10Accuracy=0.7315, over 6816.00 frames. ], tot_loss[loss=3.091, NarTop10Accuracy=0.7071, over 3865.82 frames. ], batch size: 17, lr: 3.45e-03
2024-08-06 19:54:35,647 INFO [trainer.py:765] (4/8) Epoch 25, batch 300, train_loss[loss=3.185, NarTop10Accuracy=0.6857, over 7179.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7092, over 4666.15 frames. ], batch size: 22, lr: 3.45e-03
2024-08-06 19:55:12,958 INFO [trainer.py:765] (4/8) Epoch 25, batch 400, train_loss[loss=3.079, NarTop10Accuracy=0.706, over 5007.00 frames. ], tot_loss[loss=3.087, NarTop10Accuracy=0.7083, over 5120.05 frames. ], batch size: 7, lr: 3.44e-03
2024-08-06 19:55:43,738 INFO [trainer.py:765] (4/8) Epoch 25, batch 500, train_loss[loss=2.956, NarTop10Accuracy=0.7355, over 6123.00 frames. ], tot_loss[loss=3.081, NarTop10Accuracy=0.7093, over 5397.54 frames. ], batch size: 11, lr: 3.44e-03
2024-08-06 19:56:14,815 INFO [trainer.py:765] (4/8) Epoch 25, batch 600, train_loss[loss=2.685, NarTop10Accuracy=0.7863, over 5712.00 frames. ], tot_loss[loss=3.081, NarTop10Accuracy=0.7095, over 5656.11 frames. ], batch size: 9, lr: 3.44e-03
2024-08-06 19:56:55,496 INFO [trainer.py:765] (4/8) Epoch 25, batch 700, train_loss[loss=2.768, NarTop10Accuracy=0.7731, over 5088.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7106, over 5723.73 frames. ], batch size: 6, lr: 3.43e-03
2024-08-06 19:57:30,136 INFO [trainer.py:765] (4/8) Epoch 25, batch 800, train_loss[loss=2.886, NarTop10Accuracy=0.7409, over 5151.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7096, over 5782.17 frames. ], batch size: 6, lr: 3.43e-03
2024-08-06 19:58:00,679 INFO [trainer.py:765] (4/8) Epoch 25, batch 900, train_loss[loss=3.147, NarTop10Accuracy=0.6905, over 6627.00 frames. ], tot_loss[loss=3.072, NarTop10Accuracy=0.7112, over 5805.03 frames. ], batch size: 14, lr: 3.43e-03
2024-08-06 19:58:37,639 INFO [trainer.py:765] (4/8) Epoch 25, batch 1000, train_loss[loss=2.779, NarTop10Accuracy=0.7712, over 6306.00 frames. ], tot_loss[loss=3.084, NarTop10Accuracy=0.7084, over 5909.80 frames. ], batch size: 13, lr: 3.43e-03
2024-08-06 19:59:14,855 INFO [trainer.py:765] (4/8) Epoch 25, batch 1100, train_loss[loss=3.354, NarTop10Accuracy=0.6544, over 6879.00 frames. ], tot_loss[loss=3.093, NarTop10Accuracy=0.7069, over 5925.86 frames. ], batch size: 17, lr: 3.42e-03
2024-08-06 19:59:49,039 INFO [trainer.py:765] (4/8) Epoch 25, batch 1200, train_loss[loss=3.25, NarTop10Accuracy=0.6747, over 7533.00 frames. ], tot_loss[loss=3.093, NarTop10Accuracy=0.707, over 5919.76 frames. ], batch size: 31, lr: 3.42e-03
2024-08-06 20:00:25,598 INFO [trainer.py:765] (4/8) Epoch 25, batch 1300, train_loss[loss=2.775, NarTop10Accuracy=0.773, over 5223.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.7083, over 5990.98 frames. ], batch size: 6, lr: 3.42e-03
2024-08-06 20:01:02,016 INFO [trainer.py:765] (4/8) Epoch 25, batch 1400, train_loss[loss=2.676, NarTop10Accuracy=0.7903, over 6057.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7093, over 6011.13 frames. ], batch size: 11, lr: 3.42e-03
2024-08-06 20:01:32,823 INFO [trainer.py:765] (4/8) Epoch 25, batch 1500, train_loss[loss=3.161, NarTop10Accuracy=0.6859, over 5940.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7082, over 5944.63 frames. ], batch size: 50, lr: 3.41e-03
2024-08-06 20:02:00,625 INFO [trainer.py:765] (4/8) Epoch 25, batch 1600, train_loss[loss=2.874, NarTop10Accuracy=0.7532, over 7167.00 frames. ], tot_loss[loss=3.084, NarTop10Accuracy=0.7087, over 5930.19 frames. ], batch size: 22, lr: 3.41e-03
2024-08-06 20:02:27,360 INFO [trainer.py:765] (4/8) Epoch 25, batch 1700, train_loss[loss=3.105, NarTop10Accuracy=0.7152, over 6549.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.7081, over 5917.43 frames. ], batch size: 14, lr: 3.41e-03
2024-08-06 20:02:53,853 INFO [trainer.py:765] (4/8) Epoch 25, batch 1800, train_loss[loss=3.393, NarTop10Accuracy=0.6539, over 7029.00 frames. ], tot_loss[loss=3.094, NarTop10Accuracy=0.7062, over 5993.40 frames. ], batch size: 22, lr: 3.40e-03
2024-08-06 20:03:20,340 INFO [trainer.py:765] (4/8) Epoch 25, batch 1900, train_loss[loss=3.151, NarTop10Accuracy=0.7035, over 6246.00 frames. ], tot_loss[loss=3.108, NarTop10Accuracy=0.7037, over 6018.07 frames. ], batch size: 50, lr: 3.40e-03
2024-08-06 20:03:45,933 INFO [trainer.py:765] (4/8) Epoch 25, batch 2000, train_loss[loss=3.524, NarTop10Accuracy=0.622, over 5985.00 frames. ], tot_loss[loss=3.122, NarTop10Accuracy=0.7011, over 5982.15 frames. ], batch size: 51, lr: 3.40e-03
2024-08-06 20:04:11,246 INFO [trainer.py:765] (4/8) Epoch 25, batch 2100, train_loss[loss=2.774, NarTop10Accuracy=0.7727, over 4050.00 frames. ], tot_loss[loss=3.111, NarTop10Accuracy=0.7031, over 5976.10 frames. ], batch size: 4, lr: 3.40e-03
2024-08-06 20:04:31,409 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:04:39,343 INFO [trainer.py:811] (4/8) Epoch 25, validation: loss=2.999, NarTop10Accuracy=0.7251, over 1905321.00 frames. 
2024-08-06 20:04:39,344 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 28075MB
2024-08-06 20:04:39,840 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.755e+02 2.185e+02 2.339e+02 2.507e+02 3.640e+02, threshold=4.678e+02, percent-clipped=0.0
2024-08-06 20:04:44,512 INFO [trainer.py:765] (4/8) Epoch 25, batch 2200, train_loss[loss=3.22, NarTop10Accuracy=0.6776, over 7413.00 frames. ], tot_loss[loss=3.113, NarTop10Accuracy=0.7027, over 6022.23 frames. ], batch size: 31, lr: 3.39e-03
2024-08-06 20:05:09,644 INFO [trainer.py:765] (4/8) Epoch 25, batch 2300, train_loss[loss=3.073, NarTop10Accuracy=0.7046, over 5772.00 frames. ], tot_loss[loss=3.115, NarTop10Accuracy=0.7026, over 6021.93 frames. ], batch size: 9, lr: 3.39e-03
2024-08-06 20:05:34,140 INFO [trainer.py:765] (4/8) Epoch 25, batch 2400, train_loss[loss=2.839, NarTop10Accuracy=0.7637, over 5358.00 frames. ], tot_loss[loss=3.094, NarTop10Accuracy=0.7067, over 5787.24 frames. ], batch size: 7, lr: 3.39e-03
2024-08-06 20:05:57,845 INFO [trainer.py:765] (4/8) Epoch 25, batch 2500, train_loss[loss=2.69, NarTop10Accuracy=0.7922, over 5256.00 frames. ], tot_loss[loss=3.07, NarTop10Accuracy=0.7114, over 5489.24 frames. ], batch size: 7, lr: 3.39e-03
2024-08-06 20:06:17,423 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 20:07:19,304 INFO [trainer.py:765] (4/8) Epoch 26, batch 100, train_loss[loss=3.09, NarTop10Accuracy=0.7109, over 7302.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7088, over 2359.91 frames. ], batch size: 31, lr: 3.32e-03
2024-08-06 20:07:52,382 INFO [trainer.py:765] (4/8) Epoch 26, batch 200, train_loss[loss=2.735, NarTop10Accuracy=0.7737, over 6768.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7094, over 3859.16 frames. ], batch size: 17, lr: 3.31e-03
2024-08-06 20:08:24,733 INFO [trainer.py:765] (4/8) Epoch 26, batch 300, train_loss[loss=3.035, NarTop10Accuracy=0.7168, over 7152.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.709, over 4674.21 frames. ], batch size: 22, lr: 3.31e-03
2024-08-06 20:08:58,184 INFO [trainer.py:765] (4/8) Epoch 26, batch 400, train_loss[loss=2.993, NarTop10Accuracy=0.735, over 5307.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7092, over 5100.65 frames. ], batch size: 7, lr: 3.31e-03
2024-08-06 20:09:33,147 INFO [trainer.py:765] (4/8) Epoch 26, batch 500, train_loss[loss=2.798, NarTop10Accuracy=0.7736, over 6075.00 frames. ], tot_loss[loss=3.09, NarTop10Accuracy=0.7072, over 5382.86 frames. ], batch size: 11, lr: 3.30e-03
2024-08-06 20:10:03,890 INFO [trainer.py:765] (4/8) Epoch 26, batch 600, train_loss[loss=2.816, NarTop10Accuracy=0.7683, over 5694.00 frames. ], tot_loss[loss=3.067, NarTop10Accuracy=0.7121, over 5650.30 frames. ], batch size: 9, lr: 3.30e-03
2024-08-06 20:10:39,872 INFO [trainer.py:765] (4/8) Epoch 26, batch 700, train_loss[loss=3.35, NarTop10Accuracy=0.656, over 5016.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.708, over 5737.27 frames. ], batch size: 6, lr: 3.30e-03
2024-08-06 20:11:19,061 INFO [trainer.py:765] (4/8) Epoch 26, batch 800, train_loss[loss=2.856, NarTop10Accuracy=0.7507, over 5121.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.7084, over 5785.32 frames. ], batch size: 6, lr: 3.30e-03
2024-08-06 20:11:49,315 INFO [trainer.py:765] (4/8) Epoch 26, batch 900, train_loss[loss=2.847, NarTop10Accuracy=0.7529, over 6621.00 frames. ], tot_loss[loss=3.08, NarTop10Accuracy=0.709, over 5792.43 frames. ], batch size: 14, lr: 3.29e-03
2024-08-06 20:12:25,973 INFO [trainer.py:765] (4/8) Epoch 26, batch 1000, train_loss[loss=2.912, NarTop10Accuracy=0.7513, over 6114.00 frames. ], tot_loss[loss=3.091, NarTop10Accuracy=0.7068, over 5898.38 frames. ], batch size: 13, lr: 3.29e-03
2024-08-06 20:13:06,377 INFO [trainer.py:765] (4/8) Epoch 26, batch 1100, train_loss[loss=3.354, NarTop10Accuracy=0.6608, over 6750.00 frames. ], tot_loss[loss=3.099, NarTop10Accuracy=0.7053, over 5941.19 frames. ], batch size: 17, lr: 3.29e-03
2024-08-06 20:13:37,536 INFO [trainer.py:765] (4/8) Epoch 26, batch 1200, train_loss[loss=3.23, NarTop10Accuracy=0.6689, over 7224.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7086, over 5944.88 frames. ], batch size: 31, lr: 3.29e-03
2024-08-06 20:14:13,696 INFO [trainer.py:765] (4/8) Epoch 26, batch 1300, train_loss[loss=2.864, NarTop10Accuracy=0.7516, over 5178.00 frames. ], tot_loss[loss=3.083, NarTop10Accuracy=0.7089, over 6019.74 frames. ], batch size: 6, lr: 3.28e-03
2024-08-06 20:14:50,538 INFO [trainer.py:765] (4/8) Epoch 26, batch 1400, train_loss[loss=2.788, NarTop10Accuracy=0.7678, over 6054.00 frames. ], tot_loss[loss=3.084, NarTop10Accuracy=0.7085, over 6029.11 frames. ], batch size: 11, lr: 3.28e-03
2024-08-06 20:15:21,155 INFO [trainer.py:765] (4/8) Epoch 26, batch 1500, train_loss[loss=3.083, NarTop10Accuracy=0.7145, over 6162.00 frames. ], tot_loss[loss=3.083, NarTop10Accuracy=0.709, over 5971.23 frames. ], batch size: 50, lr: 3.28e-03
2024-08-06 20:15:48,979 INFO [trainer.py:765] (4/8) Epoch 26, batch 1600, train_loss[loss=3.027, NarTop10Accuracy=0.722, over 6963.00 frames. ], tot_loss[loss=3.076, NarTop10Accuracy=0.7103, over 5956.04 frames. ], batch size: 22, lr: 3.28e-03
2024-08-06 20:15:50,002 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:15:58,239 INFO [trainer.py:811] (4/8) Epoch 26, validation: loss=2.899, NarTop10Accuracy=0.7457, over 1905321.00 frames. 
2024-08-06 20:15:58,239 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 20:15:58,778 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.752e+02 2.166e+02 2.322e+02 2.511e+02 3.952e+02, threshold=4.644e+02, percent-clipped=0.0
2024-08-06 20:16:23,951 INFO [trainer.py:765] (4/8) Epoch 26, batch 1700, train_loss[loss=3.215, NarTop10Accuracy=0.6768, over 6105.00 frames. ], tot_loss[loss=3.066, NarTop10Accuracy=0.7124, over 5950.76 frames. ], batch size: 13, lr: 3.28e-03
2024-08-06 20:16:50,426 INFO [trainer.py:765] (4/8) Epoch 26, batch 1800, train_loss[loss=2.797, NarTop10Accuracy=0.769, over 7122.00 frames. ], tot_loss[loss=3.079, NarTop10Accuracy=0.7099, over 6006.22 frames. ], batch size: 22, lr: 3.27e-03
2024-08-06 20:17:16,840 INFO [trainer.py:765] (4/8) Epoch 26, batch 1900, train_loss[loss=2.969, NarTop10Accuracy=0.7331, over 6306.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7081, over 6030.76 frames. ], batch size: 50, lr: 3.27e-03
2024-08-06 20:17:42,379 INFO [trainer.py:765] (4/8) Epoch 26, batch 2000, train_loss[loss=3.584, NarTop10Accuracy=0.6109, over 5538.00 frames. ], tot_loss[loss=3.098, NarTop10Accuracy=0.7065, over 5992.64 frames. ], batch size: 51, lr: 3.27e-03
2024-08-06 20:18:07,562 INFO [trainer.py:765] (4/8) Epoch 26, batch 2100, train_loss[loss=3.006, NarTop10Accuracy=0.7165, over 3972.00 frames. ], tot_loss[loss=3.098, NarTop10Accuracy=0.7062, over 5966.98 frames. ], batch size: 4, lr: 3.27e-03
2024-08-06 20:18:32,776 INFO [trainer.py:765] (4/8) Epoch 26, batch 2200, train_loss[loss=2.863, NarTop10Accuracy=0.7521, over 7107.00 frames. ], tot_loss[loss=3.094, NarTop10Accuracy=0.707, over 6004.37 frames. ], batch size: 31, lr: 3.26e-03
2024-08-06 20:18:57,897 INFO [trainer.py:765] (4/8) Epoch 26, batch 2300, train_loss[loss=3.147, NarTop10Accuracy=0.6805, over 5625.00 frames. ], tot_loss[loss=3.098, NarTop10Accuracy=0.706, over 6024.40 frames. ], batch size: 9, lr: 3.26e-03
2024-08-06 20:19:22,205 INFO [trainer.py:765] (4/8) Epoch 26, batch 2400, train_loss[loss=2.786, NarTop10Accuracy=0.7699, over 5187.00 frames. ], tot_loss[loss=3.073, NarTop10Accuracy=0.7108, over 5774.19 frames. ], batch size: 7, lr: 3.26e-03
2024-08-06 20:19:45,651 INFO [trainer.py:765] (4/8) Epoch 26, batch 2500, train_loss[loss=2.736, NarTop10Accuracy=0.7783, over 5256.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7153, over 5478.16 frames. ], batch size: 7, lr: 3.26e-03
2024-08-06 20:20:05,796 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 20:21:04,874 INFO [trainer.py:765] (4/8) Epoch 27, batch 100, train_loss[loss=3.222, NarTop10Accuracy=0.6768, over 7059.00 frames. ], tot_loss[loss=3.07, NarTop10Accuracy=0.7115, over 2361.67 frames. ], batch size: 31, lr: 3.19e-03
2024-08-06 20:21:39,783 INFO [trainer.py:765] (4/8) Epoch 27, batch 200, train_loss[loss=2.689, NarTop10Accuracy=0.7887, over 6657.00 frames. ], tot_loss[loss=3.077, NarTop10Accuracy=0.7101, over 3855.36 frames. ], batch size: 17, lr: 3.19e-03
2024-08-06 20:22:13,049 INFO [trainer.py:765] (4/8) Epoch 27, batch 300, train_loss[loss=2.907, NarTop10Accuracy=0.7549, over 7101.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7102, over 4665.99 frames. ], batch size: 22, lr: 3.18e-03
2024-08-06 20:22:43,557 INFO [trainer.py:765] (4/8) Epoch 27, batch 400, train_loss[loss=2.713, NarTop10Accuracy=0.7812, over 5238.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7138, over 5114.69 frames. ], batch size: 7, lr: 3.18e-03
2024-08-06 20:23:18,084 INFO [trainer.py:765] (4/8) Epoch 27, batch 500, train_loss[loss=2.769, NarTop10Accuracy=0.7797, over 6015.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7163, over 5385.67 frames. ], batch size: 11, lr: 3.18e-03
2024-08-06 20:23:51,435 INFO [trainer.py:765] (4/8) Epoch 27, batch 600, train_loss[loss=3.099, NarTop10Accuracy=0.6934, over 5871.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7162, over 5651.35 frames. ], batch size: 9, lr: 3.18e-03
2024-08-06 20:24:24,975 INFO [trainer.py:765] (4/8) Epoch 27, batch 700, train_loss[loss=2.745, NarTop10Accuracy=0.7885, over 5085.00 frames. ], tot_loss[loss=3.045, NarTop10Accuracy=0.7166, over 5712.54 frames. ], batch size: 6, lr: 3.18e-03
2024-08-06 20:25:03,407 INFO [trainer.py:765] (4/8) Epoch 27, batch 800, train_loss[loss=3.283, NarTop10Accuracy=0.6593, over 5160.00 frames. ], tot_loss[loss=3.072, NarTop10Accuracy=0.7112, over 5771.59 frames. ], batch size: 6, lr: 3.17e-03
2024-08-06 20:25:34,176 INFO [trainer.py:765] (4/8) Epoch 27, batch 900, train_loss[loss=3.197, NarTop10Accuracy=0.6875, over 6654.00 frames. ], tot_loss[loss=3.072, NarTop10Accuracy=0.7109, over 5808.87 frames. ], batch size: 14, lr: 3.17e-03
2024-08-06 20:26:10,097 INFO [trainer.py:765] (4/8) Epoch 27, batch 1000, train_loss[loss=2.853, NarTop10Accuracy=0.7598, over 6078.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.7105, over 5912.63 frames. ], batch size: 13, lr: 3.17e-03
2024-08-06 20:26:18,314 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:26:26,346 INFO [trainer.py:811] (4/8) Epoch 27, validation: loss=2.95, NarTop10Accuracy=0.735, over 1905321.00 frames. 
2024-08-06 20:26:26,347 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 20:26:26,877 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.789e+02 2.166e+02 2.331e+02 2.512e+02 4.284e+02, threshold=4.663e+02, percent-clipped=0.0
2024-08-06 20:26:50,899 INFO [trainer.py:765] (4/8) Epoch 27, batch 1100, train_loss[loss=2.958, NarTop10Accuracy=0.7336, over 6738.00 frames. ], tot_loss[loss=3.074, NarTop10Accuracy=0.7105, over 5943.89 frames. ], batch size: 17, lr: 3.17e-03
2024-08-06 20:27:24,545 INFO [trainer.py:765] (4/8) Epoch 27, batch 1200, train_loss[loss=2.927, NarTop10Accuracy=0.7482, over 6921.00 frames. ], tot_loss[loss=3.068, NarTop10Accuracy=0.7119, over 5933.88 frames. ], batch size: 31, lr: 3.16e-03
2024-08-06 20:27:58,568 INFO [trainer.py:765] (4/8) Epoch 27, batch 1300, train_loss[loss=2.709, NarTop10Accuracy=0.7849, over 4986.00 frames. ], tot_loss[loss=3.063, NarTop10Accuracy=0.7132, over 6001.94 frames. ], batch size: 6, lr: 3.16e-03
2024-08-06 20:28:36,745 INFO [trainer.py:765] (4/8) Epoch 27, batch 1400, train_loss[loss=3.367, NarTop10Accuracy=0.653, over 6060.00 frames. ], tot_loss[loss=3.084, NarTop10Accuracy=0.7088, over 6030.68 frames. ], batch size: 11, lr: 3.16e-03
2024-08-06 20:29:04,632 INFO [trainer.py:765] (4/8) Epoch 27, batch 1500, train_loss[loss=2.96, NarTop10Accuracy=0.7308, over 5748.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.7104, over 5970.03 frames. ], batch size: 50, lr: 3.16e-03
2024-08-06 20:29:32,362 INFO [trainer.py:765] (4/8) Epoch 27, batch 1600, train_loss[loss=2.976, NarTop10Accuracy=0.7348, over 7134.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7088, over 5940.84 frames. ], batch size: 22, lr: 3.15e-03
2024-08-06 20:29:58,977 INFO [trainer.py:765] (4/8) Epoch 27, batch 1700, train_loss[loss=3.079, NarTop10Accuracy=0.7087, over 6204.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7099, over 5911.56 frames. ], batch size: 13, lr: 3.15e-03
2024-08-06 20:30:25,463 INFO [trainer.py:765] (4/8) Epoch 27, batch 1800, train_loss[loss=3.504, NarTop10Accuracy=0.6256, over 6912.00 frames. ], tot_loss[loss=3.083, NarTop10Accuracy=0.7089, over 5973.77 frames. ], batch size: 22, lr: 3.15e-03
2024-08-06 20:30:51,845 INFO [trainer.py:765] (4/8) Epoch 27, batch 1900, train_loss[loss=3.085, NarTop10Accuracy=0.7188, over 6060.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7083, over 6014.56 frames. ], batch size: 51, lr: 3.15e-03
2024-08-06 20:31:17,390 INFO [trainer.py:765] (4/8) Epoch 27, batch 2000, train_loss[loss=3.081, NarTop10Accuracy=0.7073, over 6183.00 frames. ], tot_loss[loss=3.081, NarTop10Accuracy=0.7098, over 5994.65 frames. ], batch size: 50, lr: 3.15e-03
2024-08-06 20:31:42,659 INFO [trainer.py:765] (4/8) Epoch 27, batch 2100, train_loss[loss=2.721, NarTop10Accuracy=0.7772, over 3939.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7097, over 5963.82 frames. ], batch size: 4, lr: 3.14e-03
2024-08-06 20:32:07,804 INFO [trainer.py:765] (4/8) Epoch 27, batch 2200, train_loss[loss=3.447, NarTop10Accuracy=0.6339, over 7152.00 frames. ], tot_loss[loss=3.09, NarTop10Accuracy=0.708, over 6013.00 frames. ], batch size: 31, lr: 3.14e-03
2024-08-06 20:32:32,941 INFO [trainer.py:765] (4/8) Epoch 27, batch 2300, train_loss[loss=2.782, NarTop10Accuracy=0.7626, over 5748.00 frames. ], tot_loss[loss=3.087, NarTop10Accuracy=0.7086, over 6028.75 frames. ], batch size: 9, lr: 3.14e-03
2024-08-06 20:32:57,246 INFO [trainer.py:765] (4/8) Epoch 27, batch 2400, train_loss[loss=2.725, NarTop10Accuracy=0.7866, over 5109.00 frames. ], tot_loss[loss=3.088, NarTop10Accuracy=0.7081, over 5760.84 frames. ], batch size: 7, lr: 3.14e-03
2024-08-06 20:33:20,615 INFO [trainer.py:765] (4/8) Epoch 27, batch 2500, train_loss[loss=3.552, NarTop10Accuracy=0.6141, over 5055.00 frames. ], tot_loss[loss=3.055, NarTop10Accuracy=0.7146, over 5460.59 frames. ], batch size: 7, lr: 3.13e-03
2024-08-06 20:33:40,897 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 20:34:35,828 INFO [trainer.py:765] (4/8) Epoch 28, batch 100, train_loss[loss=3.021, NarTop10Accuracy=0.7284, over 7467.00 frames. ], tot_loss[loss=3.087, NarTop10Accuracy=0.7089, over 2360.22 frames. ], batch size: 31, lr: 3.07e-03
2024-08-06 20:35:07,393 INFO [trainer.py:765] (4/8) Epoch 28, batch 200, train_loss[loss=2.752, NarTop10Accuracy=0.7733, over 6726.00 frames. ], tot_loss[loss=3.091, NarTop10Accuracy=0.7077, over 3870.05 frames. ], batch size: 17, lr: 3.07e-03
2024-08-06 20:35:45,421 INFO [trainer.py:765] (4/8) Epoch 28, batch 300, train_loss[loss=2.992, NarTop10Accuracy=0.7186, over 7185.00 frames. ], tot_loss[loss=3.084, NarTop10Accuracy=0.7092, over 4676.21 frames. ], batch size: 22, lr: 3.07e-03
2024-08-06 20:36:15,864 INFO [trainer.py:765] (4/8) Epoch 28, batch 400, train_loss[loss=3.32, NarTop10Accuracy=0.6632, over 5304.00 frames. ], tot_loss[loss=3.092, NarTop10Accuracy=0.7075, over 5131.27 frames. ], batch size: 7, lr: 3.07e-03
2024-08-06 20:36:32,406 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:36:40,530 INFO [trainer.py:811] (4/8) Epoch 28, validation: loss=2.963, NarTop10Accuracy=0.7327, over 1905321.00 frames. 
2024-08-06 20:36:40,531 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 20:36:41,102 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.761e+02 2.179e+02 2.348e+02 2.536e+02 3.573e+02, threshold=4.696e+02, percent-clipped=0.0
2024-08-06 20:36:56,664 INFO [trainer.py:765] (4/8) Epoch 28, batch 500, train_loss[loss=3.235, NarTop10Accuracy=0.6836, over 6018.00 frames. ], tot_loss[loss=3.076, NarTop10Accuracy=0.7106, over 5400.81 frames. ], batch size: 11, lr: 3.06e-03
2024-08-06 20:37:29,463 INFO [trainer.py:765] (4/8) Epoch 28, batch 600, train_loss[loss=3.2, NarTop10Accuracy=0.6826, over 5604.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7103, over 5665.52 frames. ], batch size: 9, lr: 3.06e-03
2024-08-06 20:38:08,891 INFO [trainer.py:765] (4/8) Epoch 28, batch 700, train_loss[loss=3.189, NarTop10Accuracy=0.6837, over 4359.00 frames. ], tot_loss[loss=3.087, NarTop10Accuracy=0.7081, over 5731.66 frames. ], batch size: 5, lr: 3.06e-03
2024-08-06 20:38:42,489 INFO [trainer.py:765] (4/8) Epoch 28, batch 800, train_loss[loss=2.88, NarTop10Accuracy=0.7597, over 5121.00 frames. ], tot_loss[loss=3.062, NarTop10Accuracy=0.7131, over 5796.08 frames. ], batch size: 6, lr: 3.06e-03
2024-08-06 20:39:15,507 INFO [trainer.py:765] (4/8) Epoch 28, batch 900, train_loss[loss=3.314, NarTop10Accuracy=0.6669, over 6177.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7144, over 5812.40 frames. ], batch size: 13, lr: 3.06e-03
2024-08-06 20:39:53,241 INFO [trainer.py:765] (4/8) Epoch 28, batch 1000, train_loss[loss=3.357, NarTop10Accuracy=0.6525, over 6138.00 frames. ], tot_loss[loss=3.056, NarTop10Accuracy=0.714, over 5915.67 frames. ], batch size: 13, lr: 3.05e-03
2024-08-06 20:40:25,868 INFO [trainer.py:765] (4/8) Epoch 28, batch 1100, train_loss[loss=2.728, NarTop10Accuracy=0.7824, over 6891.00 frames. ], tot_loss[loss=3.068, NarTop10Accuracy=0.7113, over 5948.49 frames. ], batch size: 17, lr: 3.05e-03
2024-08-06 20:40:59,419 INFO [trainer.py:765] (4/8) Epoch 28, batch 1200, train_loss[loss=3.253, NarTop10Accuracy=0.6707, over 7251.00 frames. ], tot_loss[loss=3.077, NarTop10Accuracy=0.7096, over 5915.87 frames. ], batch size: 31, lr: 3.05e-03
2024-08-06 20:41:38,681 INFO [trainer.py:765] (4/8) Epoch 28, batch 1300, train_loss[loss=3.048, NarTop10Accuracy=0.7133, over 5106.00 frames. ], tot_loss[loss=3.08, NarTop10Accuracy=0.7093, over 5977.92 frames. ], batch size: 6, lr: 3.05e-03
2024-08-06 20:42:13,048 INFO [trainer.py:765] (4/8) Epoch 28, batch 1400, train_loss[loss=2.944, NarTop10Accuracy=0.7333, over 6069.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7078, over 6001.98 frames. ], batch size: 11, lr: 3.04e-03
2024-08-06 20:42:43,171 INFO [trainer.py:765] (4/8) Epoch 28, batch 1500, train_loss[loss=3.512, NarTop10Accuracy=0.6283, over 6204.00 frames. ], tot_loss[loss=3.07, NarTop10Accuracy=0.7112, over 5943.45 frames. ], batch size: 52, lr: 3.04e-03
2024-08-06 20:43:11,081 INFO [trainer.py:765] (4/8) Epoch 28, batch 1600, train_loss[loss=2.91, NarTop10Accuracy=0.7533, over 7191.00 frames. ], tot_loss[loss=3.081, NarTop10Accuracy=0.7091, over 5907.85 frames. ], batch size: 22, lr: 3.04e-03
2024-08-06 20:43:37,786 INFO [trainer.py:765] (4/8) Epoch 28, batch 1700, train_loss[loss=3.05, NarTop10Accuracy=0.7236, over 6315.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7084, over 5899.35 frames. ], batch size: 13, lr: 3.04e-03
2024-08-06 20:44:04,326 INFO [trainer.py:765] (4/8) Epoch 28, batch 1800, train_loss[loss=3.049, NarTop10Accuracy=0.7114, over 7155.00 frames. ], tot_loss[loss=3.079, NarTop10Accuracy=0.7094, over 5964.19 frames. ], batch size: 22, lr: 3.04e-03
2024-08-06 20:44:30,757 INFO [trainer.py:765] (4/8) Epoch 28, batch 1900, train_loss[loss=3.135, NarTop10Accuracy=0.6959, over 5802.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7096, over 6032.14 frames. ], batch size: 50, lr: 3.03e-03
2024-08-06 20:44:56,329 INFO [trainer.py:765] (4/8) Epoch 28, batch 2000, train_loss[loss=3.035, NarTop10Accuracy=0.7242, over 6129.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7144, over 6003.81 frames. ], batch size: 50, lr: 3.03e-03
2024-08-06 20:45:21,651 INFO [trainer.py:765] (4/8) Epoch 28, batch 2100, train_loss[loss=2.859, NarTop10Accuracy=0.7544, over 4896.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7145, over 5974.99 frames. ], batch size: 5, lr: 3.03e-03
2024-08-06 20:45:47,077 INFO [trainer.py:765] (4/8) Epoch 28, batch 2200, train_loss[loss=2.925, NarTop10Accuracy=0.7513, over 7329.00 frames. ], tot_loss[loss=3.068, NarTop10Accuracy=0.7117, over 6010.66 frames. ], batch size: 31, lr: 3.03e-03
2024-08-06 20:46:12,308 INFO [trainer.py:765] (4/8) Epoch 28, batch 2300, train_loss[loss=3.128, NarTop10Accuracy=0.6931, over 5805.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7085, over 6017.37 frames. ], batch size: 9, lr: 3.03e-03
2024-08-06 20:46:36,808 INFO [trainer.py:765] (4/8) Epoch 28, batch 2400, train_loss[loss=2.916, NarTop10Accuracy=0.7418, over 5127.00 frames. ], tot_loss[loss=3.085, NarTop10Accuracy=0.7085, over 5762.47 frames. ], batch size: 7, lr: 3.02e-03
2024-08-06 20:46:48,595 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:46:56,604 INFO [trainer.py:811] (4/8) Epoch 28, validation: loss=2.931, NarTop10Accuracy=0.7396, over 1905321.00 frames. 
2024-08-06 20:46:56,605 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 20:46:57,082 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.745e+02 2.201e+02 2.381e+02 2.551e+02 4.872e+02, threshold=4.762e+02, percent-clipped=0.1
2024-08-06 20:47:08,293 INFO [trainer.py:765] (4/8) Epoch 28, batch 2500, train_loss[loss=2.93, NarTop10Accuracy=0.7366, over 5019.00 frames. ], tot_loss[loss=3.069, NarTop10Accuracy=0.7115, over 5459.01 frames. ], batch size: 7, lr: 3.02e-03
2024-08-06 20:47:28,264 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 20:48:21,053 INFO [trainer.py:765] (4/8) Epoch 29, batch 100, train_loss[loss=2.978, NarTop10Accuracy=0.7259, over 7143.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7103, over 2369.30 frames. ], batch size: 31, lr: 2.96e-03
2024-08-06 20:48:53,406 INFO [trainer.py:765] (4/8) Epoch 29, batch 200, train_loss[loss=3.337, NarTop10Accuracy=0.6571, over 6753.00 frames. ], tot_loss[loss=3.053, NarTop10Accuracy=0.7155, over 3863.72 frames. ], batch size: 17, lr: 2.96e-03
2024-08-06 20:49:27,477 INFO [trainer.py:765] (4/8) Epoch 29, batch 300, train_loss[loss=3.042, NarTop10Accuracy=0.7184, over 7020.00 frames. ], tot_loss[loss=3.045, NarTop10Accuracy=0.7173, over 4686.46 frames. ], batch size: 22, lr: 2.96e-03
2024-08-06 20:49:56,053 INFO [trainer.py:765] (4/8) Epoch 29, batch 400, train_loss[loss=3.303, NarTop10Accuracy=0.6631, over 5259.00 frames. ], tot_loss[loss=3.066, NarTop10Accuracy=0.7128, over 5133.87 frames. ], batch size: 7, lr: 2.96e-03
2024-08-06 20:50:29,436 INFO [trainer.py:765] (4/8) Epoch 29, batch 500, train_loss[loss=3.265, NarTop10Accuracy=0.6727, over 5988.00 frames. ], tot_loss[loss=3.051, NarTop10Accuracy=0.7156, over 5376.77 frames. ], batch size: 11, lr: 2.96e-03
2024-08-06 20:51:00,025 INFO [trainer.py:765] (4/8) Epoch 29, batch 600, train_loss[loss=2.728, NarTop10Accuracy=0.7852, over 5706.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7154, over 5636.78 frames. ], batch size: 9, lr: 2.95e-03
2024-08-06 20:51:35,678 INFO [trainer.py:765] (4/8) Epoch 29, batch 700, train_loss[loss=2.843, NarTop10Accuracy=0.77, over 4968.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.7104, over 5726.08 frames. ], batch size: 6, lr: 2.95e-03
2024-08-06 20:52:10,725 INFO [trainer.py:765] (4/8) Epoch 29, batch 800, train_loss[loss=2.744, NarTop10Accuracy=0.789, over 5142.00 frames. ], tot_loss[loss=3.064, NarTop10Accuracy=0.7127, over 5804.02 frames. ], batch size: 6, lr: 2.95e-03
2024-08-06 20:52:40,743 INFO [trainer.py:765] (4/8) Epoch 29, batch 900, train_loss[loss=2.754, NarTop10Accuracy=0.7662, over 6231.00 frames. ], tot_loss[loss=3.07, NarTop10Accuracy=0.7113, over 5809.21 frames. ], batch size: 13, lr: 2.95e-03
2024-08-06 20:53:16,862 INFO [trainer.py:765] (4/8) Epoch 29, batch 1000, train_loss[loss=3.333, NarTop10Accuracy=0.6516, over 6189.00 frames. ], tot_loss[loss=3.076, NarTop10Accuracy=0.7102, over 5906.77 frames. ], batch size: 13, lr: 2.95e-03
2024-08-06 20:53:52,903 INFO [trainer.py:765] (4/8) Epoch 29, batch 1100, train_loss[loss=3.202, NarTop10Accuracy=0.6857, over 6780.00 frames. ], tot_loss[loss=3.082, NarTop10Accuracy=0.7088, over 5923.47 frames. ], batch size: 17, lr: 2.94e-03
2024-08-06 20:54:23,691 INFO [trainer.py:765] (4/8) Epoch 29, batch 1200, train_loss[loss=3.103, NarTop10Accuracy=0.7101, over 7014.00 frames. ], tot_loss[loss=3.076, NarTop10Accuracy=0.7104, over 5905.59 frames. ], batch size: 31, lr: 2.94e-03
2024-08-06 20:55:01,429 INFO [trainer.py:765] (4/8) Epoch 29, batch 1300, train_loss[loss=2.748, NarTop10Accuracy=0.7694, over 4884.00 frames. ], tot_loss[loss=3.072, NarTop10Accuracy=0.7112, over 5981.01 frames. ], batch size: 6, lr: 2.94e-03
2024-08-06 20:55:32,557 INFO [trainer.py:765] (4/8) Epoch 29, batch 1400, train_loss[loss=3.263, NarTop10Accuracy=0.6711, over 6057.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.7105, over 6007.72 frames. ], batch size: 11, lr: 2.94e-03
2024-08-06 20:56:04,359 INFO [trainer.py:765] (4/8) Epoch 29, batch 1500, train_loss[loss=3.322, NarTop10Accuracy=0.6586, over 6621.00 frames. ], tot_loss[loss=3.077, NarTop10Accuracy=0.7101, over 5970.65 frames. ], batch size: 50, lr: 2.94e-03
2024-08-06 20:56:32,042 INFO [trainer.py:765] (4/8) Epoch 29, batch 1600, train_loss[loss=3.377, NarTop10Accuracy=0.6445, over 7233.00 frames. ], tot_loss[loss=3.08, NarTop10Accuracy=0.7092, over 5957.40 frames. ], batch size: 22, lr: 2.93e-03
2024-08-06 20:56:58,640 INFO [trainer.py:765] (4/8) Epoch 29, batch 1700, train_loss[loss=2.787, NarTop10Accuracy=0.7714, over 6315.00 frames. ], tot_loss[loss=3.08, NarTop10Accuracy=0.7098, over 5944.53 frames. ], batch size: 13, lr: 2.93e-03
2024-08-06 20:57:25,001 INFO [trainer.py:765] (4/8) Epoch 29, batch 1800, train_loss[loss=3.111, NarTop10Accuracy=0.7049, over 7134.00 frames. ], tot_loss[loss=3.071, NarTop10Accuracy=0.7115, over 5989.54 frames. ], batch size: 22, lr: 2.93e-03
2024-08-06 20:57:44,622 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 20:57:52,863 INFO [trainer.py:811] (4/8) Epoch 29, validation: loss=2.897, NarTop10Accuracy=0.7458, over 1905321.00 frames. 
2024-08-06 20:57:52,864 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 20:57:53,424 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.772e+02 2.206e+02 2.380e+02 2.554e+02 4.464e+02, threshold=4.759e+02, percent-clipped=0.0
2024-08-06 20:57:59,756 INFO [trainer.py:765] (4/8) Epoch 29, batch 1900, train_loss[loss=3.002, NarTop10Accuracy=0.7232, over 5916.00 frames. ], tot_loss[loss=3.088, NarTop10Accuracy=0.708, over 6027.08 frames. ], batch size: 50, lr: 2.93e-03
2024-08-06 20:58:25,308 INFO [trainer.py:765] (4/8) Epoch 29, batch 2000, train_loss[loss=3.553, NarTop10Accuracy=0.6196, over 6162.00 frames. ], tot_loss[loss=3.086, NarTop10Accuracy=0.7083, over 6007.44 frames. ], batch size: 50, lr: 2.93e-03
2024-08-06 20:58:50,629 INFO [trainer.py:765] (4/8) Epoch 29, batch 2100, train_loss[loss=2.972, NarTop10Accuracy=0.7445, over 3918.00 frames. ], tot_loss[loss=3.091, NarTop10Accuracy=0.7072, over 5978.34 frames. ], batch size: 4, lr: 2.92e-03
2024-08-06 20:59:15,805 INFO [trainer.py:765] (4/8) Epoch 29, batch 2200, train_loss[loss=2.912, NarTop10Accuracy=0.7416, over 7332.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.71, over 6024.18 frames. ], batch size: 31, lr: 2.92e-03
2024-08-06 20:59:40,910 INFO [trainer.py:765] (4/8) Epoch 29, batch 2300, train_loss[loss=2.892, NarTop10Accuracy=0.7432, over 5685.00 frames. ], tot_loss[loss=3.089, NarTop10Accuracy=0.7076, over 6031.24 frames. ], batch size: 9, lr: 2.92e-03
2024-08-06 21:00:05,155 INFO [trainer.py:765] (4/8) Epoch 29, batch 2400, train_loss[loss=2.771, NarTop10Accuracy=0.7738, over 5142.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.7103, over 5780.43 frames. ], batch size: 7, lr: 2.92e-03
2024-08-06 21:00:28,742 INFO [trainer.py:765] (4/8) Epoch 29, batch 2500, train_loss[loss=3.455, NarTop10Accuracy=0.6368, over 4998.00 frames. ], tot_loss[loss=3.053, NarTop10Accuracy=0.7145, over 5465.58 frames. ], batch size: 7, lr: 2.92e-03
2024-08-06 21:00:48,843 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 21:01:41,716 INFO [trainer.py:765] (4/8) Epoch 30, batch 100, train_loss[loss=2.94, NarTop10Accuracy=0.7407, over 7221.00 frames. ], tot_loss[loss=3.034, NarTop10Accuracy=0.7192, over 2374.84 frames. ], batch size: 31, lr: 2.86e-03
2024-08-06 21:02:17,013 INFO [trainer.py:765] (4/8) Epoch 30, batch 200, train_loss[loss=2.974, NarTop10Accuracy=0.729, over 7116.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7206, over 3856.93 frames. ], batch size: 18, lr: 2.86e-03
2024-08-06 21:02:51,342 INFO [trainer.py:765] (4/8) Epoch 30, batch 300, train_loss[loss=2.897, NarTop10Accuracy=0.7431, over 7185.00 frames. ], tot_loss[loss=3.01, NarTop10Accuracy=0.7236, over 4649.91 frames. ], batch size: 23, lr: 2.86e-03
2024-08-06 21:03:21,642 INFO [trainer.py:765] (4/8) Epoch 30, batch 400, train_loss[loss=2.724, NarTop10Accuracy=0.7788, over 4983.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.7199, over 5098.70 frames. ], batch size: 7, lr: 2.86e-03
2024-08-06 21:03:58,545 INFO [trainer.py:765] (4/8) Epoch 30, batch 500, train_loss[loss=3.384, NarTop10Accuracy=0.6466, over 6057.00 frames. ], tot_loss[loss=3.035, NarTop10Accuracy=0.7183, over 5386.76 frames. ], batch size: 11, lr: 2.86e-03
2024-08-06 21:04:31,655 INFO [trainer.py:765] (4/8) Epoch 30, batch 600, train_loss[loss=2.926, NarTop10Accuracy=0.7388, over 5784.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7176, over 5655.39 frames. ], batch size: 9, lr: 2.85e-03
2024-08-06 21:05:03,525 INFO [trainer.py:765] (4/8) Epoch 30, batch 700, train_loss[loss=3.064, NarTop10Accuracy=0.7164, over 4971.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7206, over 5712.17 frames. ], batch size: 6, lr: 2.85e-03
2024-08-06 21:05:44,131 INFO [trainer.py:765] (4/8) Epoch 30, batch 800, train_loss[loss=2.996, NarTop10Accuracy=0.73, over 4377.00 frames. ], tot_loss[loss=3.026, NarTop10Accuracy=0.7205, over 5777.87 frames. ], batch size: 5, lr: 2.85e-03
2024-08-06 21:06:14,843 INFO [trainer.py:765] (4/8) Epoch 30, batch 900, train_loss[loss=2.794, NarTop10Accuracy=0.7623, over 6147.00 frames. ], tot_loss[loss=3.026, NarTop10Accuracy=0.7202, over 5793.63 frames. ], batch size: 13, lr: 2.85e-03
2024-08-06 21:06:48,951 INFO [trainer.py:765] (4/8) Epoch 30, batch 1000, train_loss[loss=2.926, NarTop10Accuracy=0.7458, over 6726.00 frames. ], tot_loss[loss=3.065, NarTop10Accuracy=0.7122, over 5898.88 frames. ], batch size: 14, lr: 2.85e-03
2024-08-06 21:07:25,936 INFO [trainer.py:765] (4/8) Epoch 30, batch 1100, train_loss[loss=3.312, NarTop10Accuracy=0.6665, over 7044.00 frames. ], tot_loss[loss=3.079, NarTop10Accuracy=0.7098, over 5929.50 frames. ], batch size: 17, lr: 2.84e-03
2024-08-06 21:08:02,380 INFO [trainer.py:765] (4/8) Epoch 30, batch 1200, train_loss[loss=2.955, NarTop10Accuracy=0.7347, over 7407.00 frames. ], tot_loss[loss=3.073, NarTop10Accuracy=0.7109, over 5924.34 frames. ], batch size: 31, lr: 2.84e-03
2024-08-06 21:08:35,370 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 21:08:43,457 INFO [trainer.py:811] (4/8) Epoch 30, validation: loss=2.93, NarTop10Accuracy=0.7391, over 1905321.00 frames. 
2024-08-06 21:08:43,458 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 21:08:44,197 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.770e+02 2.209e+02 2.377e+02 2.553e+02 3.956e+02, threshold=4.754e+02, percent-clipped=0.0
2024-08-06 21:08:44,202 INFO [trainer.py:765] (4/8) Epoch 30, batch 1300, train_loss[loss=2.951, NarTop10Accuracy=0.725, over 4974.00 frames. ], tot_loss[loss=3.066, NarTop10Accuracy=0.7125, over 5999.01 frames. ], batch size: 6, lr: 2.84e-03
2024-08-06 21:09:22,396 INFO [trainer.py:765] (4/8) Epoch 30, batch 1400, train_loss[loss=2.774, NarTop10Accuracy=0.7655, over 6126.00 frames. ], tot_loss[loss=3.064, NarTop10Accuracy=0.7123, over 6038.06 frames. ], batch size: 11, lr: 2.84e-03
2024-08-06 21:09:52,373 INFO [trainer.py:765] (4/8) Epoch 30, batch 1500, train_loss[loss=2.989, NarTop10Accuracy=0.7303, over 6330.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7131, over 5966.95 frames. ], batch size: 50, lr: 2.84e-03
2024-08-06 21:10:20,083 INFO [trainer.py:765] (4/8) Epoch 30, batch 1600, train_loss[loss=3.067, NarTop10Accuracy=0.7135, over 7056.00 frames. ], tot_loss[loss=3.066, NarTop10Accuracy=0.7118, over 5930.81 frames. ], batch size: 22, lr: 2.84e-03
2024-08-06 21:10:46,678 INFO [trainer.py:765] (4/8) Epoch 30, batch 1700, train_loss[loss=3.107, NarTop10Accuracy=0.6914, over 6276.00 frames. ], tot_loss[loss=3.075, NarTop10Accuracy=0.71, over 5916.21 frames. ], batch size: 13, lr: 2.83e-03
2024-08-06 21:11:13,057 INFO [trainer.py:765] (4/8) Epoch 30, batch 1800, train_loss[loss=3.424, NarTop10Accuracy=0.636, over 7260.00 frames. ], tot_loss[loss=3.074, NarTop10Accuracy=0.7102, over 5973.96 frames. ], batch size: 22, lr: 2.83e-03
2024-08-06 21:11:39,417 INFO [trainer.py:765] (4/8) Epoch 30, batch 1900, train_loss[loss=3.016, NarTop10Accuracy=0.7295, over 6252.00 frames. ], tot_loss[loss=3.083, NarTop10Accuracy=0.7093, over 6012.22 frames. ], batch size: 50, lr: 2.83e-03
2024-08-06 21:12:04,825 INFO [trainer.py:765] (4/8) Epoch 30, batch 2000, train_loss[loss=3.348, NarTop10Accuracy=0.6578, over 5970.00 frames. ], tot_loss[loss=3.067, NarTop10Accuracy=0.7123, over 5975.72 frames. ], batch size: 50, lr: 2.83e-03
2024-08-06 21:12:30,087 INFO [trainer.py:765] (4/8) Epoch 30, batch 2100, train_loss[loss=2.78, NarTop10Accuracy=0.7686, over 3852.00 frames. ], tot_loss[loss=3.069, NarTop10Accuracy=0.7119, over 5954.97 frames. ], batch size: 4, lr: 2.83e-03
2024-08-06 21:12:55,224 INFO [trainer.py:765] (4/8) Epoch 30, batch 2200, train_loss[loss=3.067, NarTop10Accuracy=0.7186, over 7128.00 frames. ], tot_loss[loss=3.073, NarTop10Accuracy=0.7111, over 5995.09 frames. ], batch size: 31, lr: 2.82e-03
2024-08-06 21:13:20,296 INFO [trainer.py:765] (4/8) Epoch 30, batch 2300, train_loss[loss=2.869, NarTop10Accuracy=0.7583, over 5562.00 frames. ], tot_loss[loss=3.092, NarTop10Accuracy=0.7078, over 5986.58 frames. ], batch size: 9, lr: 2.82e-03
2024-08-06 21:13:44,490 INFO [trainer.py:765] (4/8) Epoch 30, batch 2400, train_loss[loss=2.769, NarTop10Accuracy=0.7661, over 5865.00 frames. ], tot_loss[loss=3.053, NarTop10Accuracy=0.7151, over 5763.64 frames. ], batch size: 8, lr: 2.82e-03
2024-08-06 21:14:07,986 INFO [trainer.py:765] (4/8) Epoch 30, batch 2500, train_loss[loss=2.984, NarTop10Accuracy=0.7213, over 5145.00 frames. ], tot_loss[loss=3.047, NarTop10Accuracy=0.7155, over 5475.79 frames. ], batch size: 7, lr: 2.82e-03
2024-08-06 21:14:28,015 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 21:15:23,633 INFO [trainer.py:765] (4/8) Epoch 31, batch 100, train_loss[loss=3.491, NarTop10Accuracy=0.6241, over 7152.00 frames. ], tot_loss[loss=3.066, NarTop10Accuracy=0.7121, over 2365.48 frames. ], batch size: 31, lr: 2.77e-03
2024-08-06 21:15:55,127 INFO [trainer.py:765] (4/8) Epoch 31, batch 200, train_loss[loss=2.973, NarTop10Accuracy=0.7293, over 6642.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7176, over 3856.68 frames. ], batch size: 17, lr: 2.77e-03
2024-08-06 21:16:31,216 INFO [trainer.py:765] (4/8) Epoch 31, batch 300, train_loss[loss=2.912, NarTop10Accuracy=0.7535, over 7083.00 frames. ], tot_loss[loss=3.031, NarTop10Accuracy=0.7197, over 4667.27 frames. ], batch size: 22, lr: 2.77e-03
2024-08-06 21:17:01,625 INFO [trainer.py:765] (4/8) Epoch 31, batch 400, train_loss[loss=3.13, NarTop10Accuracy=0.697, over 5196.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7158, over 5106.83 frames. ], batch size: 7, lr: 2.76e-03
2024-08-06 21:17:35,724 INFO [trainer.py:765] (4/8) Epoch 31, batch 500, train_loss[loss=2.773, NarTop10Accuracy=0.7735, over 6108.00 frames. ], tot_loss[loss=3.044, NarTop10Accuracy=0.7166, over 5390.36 frames. ], batch size: 11, lr: 2.76e-03
2024-08-06 21:18:07,084 INFO [trainer.py:765] (4/8) Epoch 31, batch 600, train_loss[loss=2.831, NarTop10Accuracy=0.7624, over 5772.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7135, over 5665.82 frames. ], batch size: 9, lr: 2.76e-03
2024-08-06 21:18:44,610 INFO [trainer.py:765] (4/8) Epoch 31, batch 700, train_loss[loss=3.355, NarTop10Accuracy=0.6475, over 5106.00 frames. ], tot_loss[loss=3.063, NarTop10Accuracy=0.7131, over 5725.36 frames. ], batch size: 6, lr: 2.76e-03
2024-08-06 21:18:51,096 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 21:18:59,276 INFO [trainer.py:811] (4/8) Epoch 31, validation: loss=2.984, NarTop10Accuracy=0.7279, over 1905321.00 frames. 
2024-08-06 21:18:59,277 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 21:18:59,986 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.824e+02 2.222e+02 2.378e+02 2.557e+02 4.306e+02, threshold=4.755e+02, percent-clipped=0.0
2024-08-06 21:19:24,246 INFO [trainer.py:765] (4/8) Epoch 31, batch 800, train_loss[loss=2.767, NarTop10Accuracy=0.7755, over 5766.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.715, over 5782.59 frames. ], batch size: 7, lr: 2.76e-03
2024-08-06 21:19:56,951 INFO [trainer.py:765] (4/8) Epoch 31, batch 900, train_loss[loss=3.376, NarTop10Accuracy=0.6357, over 6222.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7152, over 5793.97 frames. ], batch size: 13, lr: 2.76e-03
2024-08-06 21:20:33,311 INFO [trainer.py:765] (4/8) Epoch 31, batch 1000, train_loss[loss=3.416, NarTop10Accuracy=0.6416, over 6252.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7158, over 5895.82 frames. ], batch size: 13, lr: 2.75e-03
2024-08-06 21:21:10,216 INFO [trainer.py:765] (4/8) Epoch 31, batch 1100, train_loss[loss=3.252, NarTop10Accuracy=0.6785, over 6831.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7157, over 5927.55 frames. ], batch size: 17, lr: 2.75e-03
2024-08-06 21:21:41,120 INFO [trainer.py:765] (4/8) Epoch 31, batch 1200, train_loss[loss=2.84, NarTop10Accuracy=0.7537, over 7236.00 frames. ], tot_loss[loss=3.036, NarTop10Accuracy=0.7179, over 5925.90 frames. ], batch size: 31, lr: 2.75e-03
2024-08-06 21:22:19,742 INFO [trainer.py:765] (4/8) Epoch 31, batch 1300, train_loss[loss=2.791, NarTop10Accuracy=0.767, over 4335.00 frames. ], tot_loss[loss=3.062, NarTop10Accuracy=0.7126, over 5988.64 frames. ], batch size: 5, lr: 2.75e-03
2024-08-06 21:22:53,534 INFO [trainer.py:765] (4/8) Epoch 31, batch 1400, train_loss[loss=2.917, NarTop10Accuracy=0.7442, over 6108.00 frames. ], tot_loss[loss=3.069, NarTop10Accuracy=0.7108, over 6035.39 frames. ], batch size: 11, lr: 2.75e-03
2024-08-06 21:23:21,269 INFO [trainer.py:765] (4/8) Epoch 31, batch 1500, train_loss[loss=3.395, NarTop10Accuracy=0.6437, over 6054.00 frames. ], tot_loss[loss=3.057, NarTop10Accuracy=0.7133, over 5948.87 frames. ], batch size: 53, lr: 2.74e-03
2024-08-06 21:23:49,005 INFO [trainer.py:765] (4/8) Epoch 31, batch 1600, train_loss[loss=3.325, NarTop10Accuracy=0.6554, over 7347.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7146, over 5931.36 frames. ], batch size: 22, lr: 2.74e-03
2024-08-06 21:24:15,512 INFO [trainer.py:765] (4/8) Epoch 31, batch 1700, train_loss[loss=3.44, NarTop10Accuracy=0.6424, over 6231.00 frames. ], tot_loss[loss=3.058, NarTop10Accuracy=0.7138, over 5931.41 frames. ], batch size: 13, lr: 2.74e-03
2024-08-06 21:24:41,996 INFO [trainer.py:765] (4/8) Epoch 31, batch 1800, train_loss[loss=2.8, NarTop10Accuracy=0.7746, over 6915.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.716, over 6002.60 frames. ], batch size: 22, lr: 2.74e-03
2024-08-06 21:25:08,357 INFO [trainer.py:765] (4/8) Epoch 31, batch 1900, train_loss[loss=3.233, NarTop10Accuracy=0.6855, over 5991.00 frames. ], tot_loss[loss=3.062, NarTop10Accuracy=0.7131, over 6032.47 frames. ], batch size: 51, lr: 2.74e-03
2024-08-06 21:25:33,773 INFO [trainer.py:765] (4/8) Epoch 31, batch 2000, train_loss[loss=3.029, NarTop10Accuracy=0.7255, over 5760.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7149, over 5995.51 frames. ], batch size: 51, lr: 2.74e-03
2024-08-06 21:25:59,106 INFO [trainer.py:765] (4/8) Epoch 31, batch 2100, train_loss[loss=2.756, NarTop10Accuracy=0.7752, over 4764.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7152, over 5983.00 frames. ], batch size: 5, lr: 2.73e-03
2024-08-06 21:26:24,238 INFO [trainer.py:765] (4/8) Epoch 31, batch 2200, train_loss[loss=2.998, NarTop10Accuracy=0.7278, over 7290.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7173, over 6025.61 frames. ], batch size: 31, lr: 2.73e-03
2024-08-06 21:26:49,322 INFO [trainer.py:765] (4/8) Epoch 31, batch 2300, train_loss[loss=2.816, NarTop10Accuracy=0.7636, over 5712.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7141, over 6040.57 frames. ], batch size: 9, lr: 2.73e-03
2024-08-06 21:27:13,607 INFO [trainer.py:765] (4/8) Epoch 31, batch 2400, train_loss[loss=2.916, NarTop10Accuracy=0.7451, over 5055.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7147, over 5776.56 frames. ], batch size: 7, lr: 2.73e-03
2024-08-06 21:27:37,027 INFO [trainer.py:765] (4/8) Epoch 31, batch 2500, train_loss[loss=2.937, NarTop10Accuracy=0.7417, over 5625.00 frames. ], tot_loss[loss=3.036, NarTop10Accuracy=0.7175, over 5492.51 frames. ], batch size: 8, lr: 2.73e-03
2024-08-06 21:27:57,405 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 21:28:49,392 INFO [trainer.py:765] (4/8) Epoch 32, batch 100, train_loss[loss=2.878, NarTop10Accuracy=0.7504, over 7425.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7156, over 2356.78 frames. ], batch size: 32, lr: 2.68e-03
2024-08-06 21:29:08,160 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 21:29:16,392 INFO [trainer.py:811] (4/8) Epoch 32, validation: loss=2.919, NarTop10Accuracy=0.7409, over 1905321.00 frames. 
2024-08-06 21:29:16,393 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 21:29:16,939 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.842e+02 2.253e+02 2.413e+02 2.600e+02 5.680e+02, threshold=4.826e+02, percent-clipped=0.1
2024-08-06 21:29:32,272 INFO [trainer.py:765] (4/8) Epoch 32, batch 200, train_loss[loss=3.229, NarTop10Accuracy=0.6773, over 6852.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7148, over 3860.33 frames. ], batch size: 17, lr: 2.68e-03
2024-08-06 21:30:05,278 INFO [trainer.py:765] (4/8) Epoch 32, batch 300, train_loss[loss=3.124, NarTop10Accuracy=0.7065, over 7299.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7163, over 4653.59 frames. ], batch size: 22, lr: 2.68e-03
2024-08-06 21:30:34,102 INFO [trainer.py:765] (4/8) Epoch 32, batch 400, train_loss[loss=2.901, NarTop10Accuracy=0.7522, over 5292.00 frames. ], tot_loss[loss=3.057, NarTop10Accuracy=0.7137, over 5106.33 frames. ], batch size: 7, lr: 2.68e-03
2024-08-06 21:31:13,529 INFO [trainer.py:765] (4/8) Epoch 32, batch 500, train_loss[loss=3.172, NarTop10Accuracy=0.6922, over 6096.00 frames. ], tot_loss[loss=3.047, NarTop10Accuracy=0.7157, over 5394.92 frames. ], batch size: 11, lr: 2.67e-03
2024-08-06 21:31:42,485 INFO [trainer.py:765] (4/8) Epoch 32, batch 600, train_loss[loss=3, NarTop10Accuracy=0.7107, over 5799.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7154, over 5661.76 frames. ], batch size: 9, lr: 2.67e-03
2024-08-06 21:32:17,028 INFO [trainer.py:765] (4/8) Epoch 32, batch 700, train_loss[loss=2.656, NarTop10Accuracy=0.7894, over 5040.00 frames. ], tot_loss[loss=3.041, NarTop10Accuracy=0.7171, over 5726.62 frames. ], batch size: 6, lr: 2.67e-03
2024-08-06 21:33:00,646 INFO [trainer.py:765] (4/8) Epoch 32, batch 800, train_loss[loss=3.241, NarTop10Accuracy=0.67, over 4941.00 frames. ], tot_loss[loss=3.047, NarTop10Accuracy=0.7161, over 5773.04 frames. ], batch size: 6, lr: 2.67e-03
2024-08-06 21:33:28,990 INFO [trainer.py:765] (4/8) Epoch 32, batch 900, train_loss[loss=2.779, NarTop10Accuracy=0.7718, over 6174.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.7189, over 5787.40 frames. ], batch size: 13, lr: 2.67e-03
2024-08-06 21:34:04,049 INFO [trainer.py:765] (4/8) Epoch 32, batch 1000, train_loss[loss=3.189, NarTop10Accuracy=0.6836, over 6312.00 frames. ], tot_loss[loss=3.042, NarTop10Accuracy=0.717, over 5889.43 frames. ], batch size: 13, lr: 2.67e-03
2024-08-06 21:34:46,674 INFO [trainer.py:765] (4/8) Epoch 32, batch 1100, train_loss[loss=3.116, NarTop10Accuracy=0.7002, over 6738.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7158, over 5924.21 frames. ], batch size: 17, lr: 2.66e-03
2024-08-06 21:35:18,171 INFO [trainer.py:765] (4/8) Epoch 32, batch 1200, train_loss[loss=3.178, NarTop10Accuracy=0.6906, over 7278.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7134, over 5910.36 frames. ], batch size: 31, lr: 2.66e-03
2024-08-06 21:35:52,800 INFO [trainer.py:765] (4/8) Epoch 32, batch 1300, train_loss[loss=3.139, NarTop10Accuracy=0.6973, over 5034.00 frames. ], tot_loss[loss=3.069, NarTop10Accuracy=0.7119, over 5991.54 frames. ], batch size: 6, lr: 2.66e-03
2024-08-06 21:36:29,478 INFO [trainer.py:765] (4/8) Epoch 32, batch 1400, train_loss[loss=3.39, NarTop10Accuracy=0.6439, over 5988.00 frames. ], tot_loss[loss=3.065, NarTop10Accuracy=0.7124, over 6022.31 frames. ], batch size: 11, lr: 2.66e-03
2024-08-06 21:37:04,733 INFO [trainer.py:765] (4/8) Epoch 32, batch 1500, train_loss[loss=3.475, NarTop10Accuracy=0.6307, over 6285.00 frames. ], tot_loss[loss=3.062, NarTop10Accuracy=0.7127, over 5950.74 frames. ], batch size: 51, lr: 2.66e-03
2024-08-06 21:37:32,521 INFO [trainer.py:765] (4/8) Epoch 32, batch 1600, train_loss[loss=3.04, NarTop10Accuracy=0.7166, over 7125.00 frames. ], tot_loss[loss=3.057, NarTop10Accuracy=0.714, over 5936.02 frames. ], batch size: 22, lr: 2.66e-03
2024-08-06 21:37:59,159 INFO [trainer.py:765] (4/8) Epoch 32, batch 1700, train_loss[loss=3.147, NarTop10Accuracy=0.6915, over 6336.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7148, over 5907.70 frames. ], batch size: 13, lr: 2.65e-03
2024-08-06 21:38:25,702 INFO [trainer.py:765] (4/8) Epoch 32, batch 1800, train_loss[loss=3.017, NarTop10Accuracy=0.7172, over 6990.00 frames. ], tot_loss[loss=3.059, NarTop10Accuracy=0.7134, over 5967.11 frames. ], batch size: 22, lr: 2.65e-03
2024-08-06 21:38:52,169 INFO [trainer.py:765] (4/8) Epoch 32, batch 1900, train_loss[loss=3.037, NarTop10Accuracy=0.7142, over 6618.00 frames. ], tot_loss[loss=3.078, NarTop10Accuracy=0.7094, over 6019.03 frames. ], batch size: 51, lr: 2.65e-03
2024-08-06 21:39:17,768 INFO [trainer.py:765] (4/8) Epoch 32, batch 2000, train_loss[loss=3.465, NarTop10Accuracy=0.6361, over 6279.00 frames. ], tot_loss[loss=3.062, NarTop10Accuracy=0.713, over 5995.00 frames. ], batch size: 51, lr: 2.65e-03
2024-08-06 21:39:43,178 INFO [trainer.py:765] (4/8) Epoch 32, batch 2100, train_loss[loss=2.64, NarTop10Accuracy=0.7981, over 4860.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7134, over 5982.03 frames. ], batch size: 5, lr: 2.65e-03
2024-08-06 21:39:54,781 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 21:40:02,942 INFO [trainer.py:811] (4/8) Epoch 32, validation: loss=2.886, NarTop10Accuracy=0.7482, over 1905321.00 frames. 
2024-08-06 21:40:02,942 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 21:40:03,423 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.874e+02 2.278e+02 2.449e+02 2.609e+02 8.207e+02, threshold=4.898e+02, percent-clipped=0.3
2024-08-06 21:40:16,629 INFO [trainer.py:765] (4/8) Epoch 32, batch 2200, train_loss[loss=3.148, NarTop10Accuracy=0.6995, over 7164.00 frames. ], tot_loss[loss=3.056, NarTop10Accuracy=0.7141, over 6016.36 frames. ], batch size: 31, lr: 2.65e-03
2024-08-06 21:40:41,718 INFO [trainer.py:765] (4/8) Epoch 32, batch 2300, train_loss[loss=3.229, NarTop10Accuracy=0.6759, over 5733.00 frames. ], tot_loss[loss=3.074, NarTop10Accuracy=0.71, over 6029.65 frames. ], batch size: 9, lr: 2.65e-03
2024-08-06 21:41:06,072 INFO [trainer.py:765] (4/8) Epoch 32, batch 2400, train_loss[loss=3.368, NarTop10Accuracy=0.6535, over 5154.00 frames. ], tot_loss[loss=3.056, NarTop10Accuracy=0.7138, over 5792.66 frames. ], batch size: 7, lr: 2.64e-03
2024-08-06 21:41:29,538 INFO [trainer.py:765] (4/8) Epoch 32, batch 2500, train_loss[loss=2.805, NarTop10Accuracy=0.7761, over 5169.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.7191, over 5502.42 frames. ], batch size: 7, lr: 2.64e-03
2024-08-06 21:41:49,894 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 21:42:47,615 INFO [trainer.py:765] (4/8) Epoch 33, batch 100, train_loss[loss=3.027, NarTop10Accuracy=0.7252, over 7278.00 frames. ], tot_loss[loss=3.009, NarTop10Accuracy=0.724, over 2361.08 frames. ], batch size: 31, lr: 2.60e-03
2024-08-06 21:43:22,368 INFO [trainer.py:765] (4/8) Epoch 33, batch 200, train_loss[loss=2.607, NarTop10Accuracy=0.8004, over 6774.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7208, over 3856.26 frames. ], batch size: 17, lr: 2.60e-03
2024-08-06 21:43:56,513 INFO [trainer.py:765] (4/8) Epoch 33, batch 300, train_loss[loss=3.424, NarTop10Accuracy=0.6305, over 6993.00 frames. ], tot_loss[loss=3.041, NarTop10Accuracy=0.7174, over 4653.87 frames. ], batch size: 22, lr: 2.60e-03
2024-08-06 21:44:30,316 INFO [trainer.py:765] (4/8) Epoch 33, batch 400, train_loss[loss=2.729, NarTop10Accuracy=0.7903, over 5181.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7179, over 5112.87 frames. ], batch size: 7, lr: 2.59e-03
2024-08-06 21:45:02,870 INFO [trainer.py:765] (4/8) Epoch 33, batch 500, train_loss[loss=2.738, NarTop10Accuracy=0.7864, over 6024.00 frames. ], tot_loss[loss=3.018, NarTop10Accuracy=0.7221, over 5391.26 frames. ], batch size: 11, lr: 2.59e-03
2024-08-06 21:45:36,226 INFO [trainer.py:765] (4/8) Epoch 33, batch 600, train_loss[loss=3.328, NarTop10Accuracy=0.6513, over 5796.00 frames. ], tot_loss[loss=3.042, NarTop10Accuracy=0.7171, over 5639.76 frames. ], batch size: 9, lr: 2.59e-03
2024-08-06 21:46:11,316 INFO [trainer.py:765] (4/8) Epoch 33, batch 700, train_loss[loss=2.85, NarTop10Accuracy=0.7637, over 4914.00 frames. ], tot_loss[loss=3.045, NarTop10Accuracy=0.7165, over 5712.53 frames. ], batch size: 6, lr: 2.59e-03
2024-08-06 21:46:46,169 INFO [trainer.py:765] (4/8) Epoch 33, batch 800, train_loss[loss=2.788, NarTop10Accuracy=0.7792, over 4950.00 frames. ], tot_loss[loss=3.045, NarTop10Accuracy=0.7166, over 5781.05 frames. ], batch size: 6, lr: 2.59e-03
2024-08-06 21:47:18,908 INFO [trainer.py:765] (4/8) Epoch 33, batch 900, train_loss[loss=3.326, NarTop10Accuracy=0.6602, over 6711.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7158, over 5782.58 frames. ], batch size: 14, lr: 2.59e-03
2024-08-06 21:47:57,316 INFO [trainer.py:765] (4/8) Epoch 33, batch 1000, train_loss[loss=2.979, NarTop10Accuracy=0.7255, over 6258.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7157, over 5898.63 frames. ], batch size: 13, lr: 2.58e-03
2024-08-06 21:48:30,908 INFO [trainer.py:765] (4/8) Epoch 33, batch 1100, train_loss[loss=2.948, NarTop10Accuracy=0.7318, over 6783.00 frames. ], tot_loss[loss=3.067, NarTop10Accuracy=0.7117, over 5937.13 frames. ], batch size: 17, lr: 2.58e-03
2024-08-06 21:49:06,659 INFO [trainer.py:765] (4/8) Epoch 33, batch 1200, train_loss[loss=2.861, NarTop10Accuracy=0.7565, over 7383.00 frames. ], tot_loss[loss=3.057, NarTop10Accuracy=0.7139, over 5934.46 frames. ], batch size: 31, lr: 2.58e-03
2024-08-06 21:49:42,815 INFO [trainer.py:765] (4/8) Epoch 33, batch 1300, train_loss[loss=3.039, NarTop10Accuracy=0.716, over 4209.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7146, over 5985.76 frames. ], batch size: 5, lr: 2.58e-03
2024-08-06 21:50:17,310 INFO [trainer.py:765] (4/8) Epoch 33, batch 1400, train_loss[loss=3.329, NarTop10Accuracy=0.6507, over 6165.00 frames. ], tot_loss[loss=3.057, NarTop10Accuracy=0.7137, over 6023.87 frames. ], batch size: 11, lr: 2.58e-03
2024-08-06 21:50:45,370 INFO [trainer.py:765] (4/8) Epoch 33, batch 1500, train_loss[loss=3.071, NarTop10Accuracy=0.7084, over 6180.00 frames. ], tot_loss[loss=3.055, NarTop10Accuracy=0.7141, over 5960.54 frames. ], batch size: 50, lr: 2.58e-03
2024-08-06 21:51:04,606 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 21:51:12,661 INFO [trainer.py:811] (4/8) Epoch 33, validation: loss=2.938, NarTop10Accuracy=0.7372, over 1905321.00 frames. 
2024-08-06 21:51:12,662 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 21:51:13,180 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.834e+02 2.250e+02 2.409e+02 2.586e+02 3.975e+02, threshold=4.818e+02, percent-clipped=0.0
2024-08-06 21:51:21,261 INFO [trainer.py:765] (4/8) Epoch 33, batch 1600, train_loss[loss=3.235, NarTop10Accuracy=0.673, over 7164.00 frames. ], tot_loss[loss=3.041, NarTop10Accuracy=0.7168, over 5918.04 frames. ], batch size: 22, lr: 2.57e-03
2024-08-06 21:51:47,923 INFO [trainer.py:765] (4/8) Epoch 33, batch 1700, train_loss[loss=2.706, NarTop10Accuracy=0.7863, over 6228.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.715, over 5903.25 frames. ], batch size: 13, lr: 2.57e-03
2024-08-06 21:52:14,392 INFO [trainer.py:765] (4/8) Epoch 33, batch 1800, train_loss[loss=2.801, NarTop10Accuracy=0.7639, over 6978.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7157, over 5971.86 frames. ], batch size: 22, lr: 2.57e-03
2024-08-06 21:52:40,856 INFO [trainer.py:765] (4/8) Epoch 33, batch 1900, train_loss[loss=3.4, NarTop10Accuracy=0.642, over 6387.00 frames. ], tot_loss[loss=3.069, NarTop10Accuracy=0.7116, over 6013.39 frames. ], batch size: 50, lr: 2.57e-03
2024-08-06 21:53:06,352 INFO [trainer.py:765] (4/8) Epoch 33, batch 2000, train_loss[loss=3.503, NarTop10Accuracy=0.6257, over 6066.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7179, over 6001.04 frames. ], batch size: 51, lr: 2.57e-03
2024-08-06 21:53:31,659 INFO [trainer.py:765] (4/8) Epoch 33, batch 2100, train_loss[loss=3.304, NarTop10Accuracy=0.6628, over 3978.00 frames. ], tot_loss[loss=3.044, NarTop10Accuracy=0.7165, over 5984.61 frames. ], batch size: 4, lr: 2.57e-03
2024-08-06 21:53:56,891 INFO [trainer.py:765] (4/8) Epoch 33, batch 2200, train_loss[loss=3.489, NarTop10Accuracy=0.6236, over 7164.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7142, over 6027.20 frames. ], batch size: 31, lr: 2.57e-03
2024-08-06 21:54:21,990 INFO [trainer.py:765] (4/8) Epoch 33, batch 2300, train_loss[loss=2.718, NarTop10Accuracy=0.7714, over 5673.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.7145, over 6011.90 frames. ], batch size: 9, lr: 2.56e-03
2024-08-06 21:54:46,430 INFO [trainer.py:765] (4/8) Epoch 33, batch 2400, train_loss[loss=2.682, NarTop10Accuracy=0.7835, over 5256.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7173, over 5768.78 frames. ], batch size: 7, lr: 2.56e-03
2024-08-06 21:55:09,862 INFO [trainer.py:765] (4/8) Epoch 33, batch 2500, train_loss[loss=2.694, NarTop10Accuracy=0.7814, over 5097.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7203, over 5471.32 frames. ], batch size: 7, lr: 2.56e-03
2024-08-06 21:55:29,791 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 21:56:24,721 INFO [trainer.py:765] (4/8) Epoch 34, batch 100, train_loss[loss=3.34, NarTop10Accuracy=0.6624, over 7140.00 frames. ], tot_loss[loss=3.044, NarTop10Accuracy=0.7174, over 2383.12 frames. ], batch size: 31, lr: 2.52e-03
2024-08-06 21:56:55,613 INFO [trainer.py:765] (4/8) Epoch 34, batch 200, train_loss[loss=3.074, NarTop10Accuracy=0.698, over 6816.00 frames. ], tot_loss[loss=3.02, NarTop10Accuracy=0.7216, over 3869.75 frames. ], batch size: 17, lr: 2.52e-03
2024-08-06 21:57:31,776 INFO [trainer.py:765] (4/8) Epoch 34, batch 300, train_loss[loss=2.889, NarTop10Accuracy=0.7526, over 7194.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.7198, over 4683.29 frames. ], batch size: 22, lr: 2.52e-03
2024-08-06 21:58:02,724 INFO [trainer.py:765] (4/8) Epoch 34, batch 400, train_loss[loss=3.275, NarTop10Accuracy=0.6601, over 5067.00 frames. ], tot_loss[loss=3.011, NarTop10Accuracy=0.7231, over 5115.84 frames. ], batch size: 7, lr: 2.52e-03
2024-08-06 21:58:34,689 INFO [trainer.py:765] (4/8) Epoch 34, batch 500, train_loss[loss=3.024, NarTop10Accuracy=0.7174, over 6063.00 frames. ], tot_loss[loss=3.024, NarTop10Accuracy=0.7205, over 5387.82 frames. ], batch size: 11, lr: 2.51e-03
2024-08-06 21:59:09,616 INFO [trainer.py:765] (4/8) Epoch 34, batch 600, train_loss[loss=2.903, NarTop10Accuracy=0.7439, over 5682.00 frames. ], tot_loss[loss=3.032, NarTop10Accuracy=0.7188, over 5662.63 frames. ], batch size: 9, lr: 2.51e-03
2024-08-06 21:59:46,056 INFO [trainer.py:765] (4/8) Epoch 34, batch 700, train_loss[loss=3.149, NarTop10Accuracy=0.6952, over 5097.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7178, over 5730.18 frames. ], batch size: 6, lr: 2.51e-03
2024-08-06 22:00:17,575 INFO [trainer.py:765] (4/8) Epoch 34, batch 800, train_loss[loss=2.951, NarTop10Accuracy=0.7371, over 5097.00 frames. ], tot_loss[loss=3.032, NarTop10Accuracy=0.7192, over 5772.11 frames. ], batch size: 6, lr: 2.51e-03
2024-08-06 22:00:49,874 INFO [trainer.py:765] (4/8) Epoch 34, batch 900, train_loss[loss=2.938, NarTop10Accuracy=0.7375, over 6558.00 frames. ], tot_loss[loss=3.033, NarTop10Accuracy=0.7191, over 5800.47 frames. ], batch size: 14, lr: 2.51e-03
2024-08-06 22:01:25,338 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:01:33,386 INFO [trainer.py:811] (4/8) Epoch 34, validation: loss=2.9, NarTop10Accuracy=0.7444, over 1905321.00 frames. 
2024-08-06 22:01:33,387 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:01:34,091 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.819e+02 2.259e+02 2.434e+02 2.615e+02 5.125e+02, threshold=4.868e+02, percent-clipped=0.1
2024-08-06 22:01:35,625 INFO [trainer.py:765] (4/8) Epoch 34, batch 1000, train_loss[loss=3.259, NarTop10Accuracy=0.6758, over 6630.00 frames. ], tot_loss[loss=3.042, NarTop10Accuracy=0.7173, over 5892.55 frames. ], batch size: 14, lr: 2.51e-03
2024-08-06 22:02:10,829 INFO [trainer.py:765] (4/8) Epoch 34, batch 1100, train_loss[loss=3.344, NarTop10Accuracy=0.6637, over 6816.00 frames. ], tot_loss[loss=3.047, NarTop10Accuracy=0.7158, over 5920.57 frames. ], batch size: 17, lr: 2.51e-03
2024-08-06 22:02:46,786 INFO [trainer.py:765] (4/8) Epoch 34, batch 1200, train_loss[loss=3.032, NarTop10Accuracy=0.7228, over 7095.00 frames. ], tot_loss[loss=3.043, NarTop10Accuracy=0.7162, over 5923.10 frames. ], batch size: 31, lr: 2.50e-03
2024-08-06 22:03:20,814 INFO [trainer.py:765] (4/8) Epoch 34, batch 1300, train_loss[loss=2.677, NarTop10Accuracy=0.7961, over 5121.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7174, over 5974.34 frames. ], batch size: 6, lr: 2.50e-03
2024-08-06 22:03:52,950 INFO [trainer.py:765] (4/8) Epoch 34, batch 1400, train_loss[loss=3.267, NarTop10Accuracy=0.6612, over 6009.00 frames. ], tot_loss[loss=3.042, NarTop10Accuracy=0.7167, over 5995.36 frames. ], batch size: 11, lr: 2.50e-03
2024-08-06 22:04:20,822 INFO [trainer.py:765] (4/8) Epoch 34, batch 1500, train_loss[loss=3.064, NarTop10Accuracy=0.7137, over 5715.00 frames. ], tot_loss[loss=3.043, NarTop10Accuracy=0.7167, over 5931.13 frames. ], batch size: 50, lr: 2.50e-03
2024-08-06 22:04:48,600 INFO [trainer.py:765] (4/8) Epoch 34, batch 1600, train_loss[loss=2.904, NarTop10Accuracy=0.7373, over 7278.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7154, over 5918.15 frames. ], batch size: 23, lr: 2.50e-03
2024-08-06 22:05:15,241 INFO [trainer.py:765] (4/8) Epoch 34, batch 1700, train_loss[loss=3.18, NarTop10Accuracy=0.6857, over 6630.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7177, over 5927.14 frames. ], batch size: 14, lr: 2.50e-03
2024-08-06 22:05:41,721 INFO [trainer.py:765] (4/8) Epoch 34, batch 1800, train_loss[loss=3.403, NarTop10Accuracy=0.6471, over 6927.00 frames. ], tot_loss[loss=3.045, NarTop10Accuracy=0.7164, over 5988.91 frames. ], batch size: 22, lr: 2.50e-03
2024-08-06 22:06:08,207 INFO [trainer.py:765] (4/8) Epoch 34, batch 1900, train_loss[loss=3.046, NarTop10Accuracy=0.7231, over 6171.00 frames. ], tot_loss[loss=3.067, NarTop10Accuracy=0.7115, over 6019.29 frames. ], batch size: 51, lr: 2.49e-03
2024-08-06 22:06:33,770 INFO [trainer.py:765] (4/8) Epoch 34, batch 2000, train_loss[loss=3.046, NarTop10Accuracy=0.7147, over 6222.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7148, over 5991.67 frames. ], batch size: 50, lr: 2.49e-03
2024-08-06 22:06:59,126 INFO [trainer.py:765] (4/8) Epoch 34, batch 2100, train_loss[loss=3.314, NarTop10Accuracy=0.6633, over 3864.00 frames. ], tot_loss[loss=3.065, NarTop10Accuracy=0.7117, over 5968.86 frames. ], batch size: 4, lr: 2.49e-03
2024-08-06 22:07:24,398 INFO [trainer.py:765] (4/8) Epoch 34, batch 2200, train_loss[loss=2.831, NarTop10Accuracy=0.7599, over 7446.00 frames. ], tot_loss[loss=3.064, NarTop10Accuracy=0.712, over 6021.31 frames. ], batch size: 31, lr: 2.49e-03
2024-08-06 22:07:49,535 INFO [trainer.py:765] (4/8) Epoch 34, batch 2300, train_loss[loss=2.656, NarTop10Accuracy=0.7986, over 5709.00 frames. ], tot_loss[loss=3.064, NarTop10Accuracy=0.7126, over 6036.20 frames. ], batch size: 9, lr: 2.49e-03
2024-08-06 22:08:14,059 INFO [trainer.py:765] (4/8) Epoch 34, batch 2400, train_loss[loss=3.481, NarTop10Accuracy=0.6323, over 5826.00 frames. ], tot_loss[loss=3.059, NarTop10Accuracy=0.7135, over 5791.17 frames. ], batch size: 8, lr: 2.49e-03
2024-08-06 22:08:37,648 INFO [trainer.py:765] (4/8) Epoch 34, batch 2500, train_loss[loss=2.764, NarTop10Accuracy=0.7703, over 5115.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7205, over 5495.06 frames. ], batch size: 7, lr: 2.49e-03
2024-08-06 22:08:57,618 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 22:09:52,640 INFO [trainer.py:765] (4/8) Epoch 35, batch 100, train_loss[loss=2.873, NarTop10Accuracy=0.7477, over 7347.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7175, over 2364.89 frames. ], batch size: 32, lr: 2.45e-03
2024-08-06 22:10:29,698 INFO [trainer.py:765] (4/8) Epoch 35, batch 200, train_loss[loss=3.205, NarTop10Accuracy=0.6853, over 6822.00 frames. ], tot_loss[loss=3.06, NarTop10Accuracy=0.7136, over 3835.11 frames. ], batch size: 17, lr: 2.45e-03
2024-08-06 22:11:04,942 INFO [trainer.py:765] (4/8) Epoch 35, batch 300, train_loss[loss=2.762, NarTop10Accuracy=0.7831, over 7230.00 frames. ], tot_loss[loss=3.034, NarTop10Accuracy=0.7192, over 4646.05 frames. ], batch size: 22, lr: 2.44e-03
2024-08-06 22:11:35,333 INFO [trainer.py:765] (4/8) Epoch 35, batch 400, train_loss[loss=3.006, NarTop10Accuracy=0.7304, over 5229.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.7202, over 5107.88 frames. ], batch size: 7, lr: 2.44e-03
2024-08-06 22:11:40,048 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:11:48,129 INFO [trainer.py:811] (4/8) Epoch 35, validation: loss=2.84, NarTop10Accuracy=0.7576, over 1905321.00 frames. 
2024-08-06 22:11:48,129 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:11:48,702 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.898e+02 2.275e+02 2.426e+02 2.615e+02 4.095e+02, threshold=4.852e+02, percent-clipped=0.0
2024-08-06 22:12:17,723 INFO [trainer.py:765] (4/8) Epoch 35, batch 500, train_loss[loss=2.74, NarTop10Accuracy=0.7759, over 6042.00 frames. ], tot_loss[loss=3.012, NarTop10Accuracy=0.7231, over 5386.42 frames. ], batch size: 11, lr: 2.44e-03
2024-08-06 22:12:51,425 INFO [trainer.py:765] (4/8) Epoch 35, batch 600, train_loss[loss=3.408, NarTop10Accuracy=0.6441, over 5664.00 frames. ], tot_loss[loss=3.031, NarTop10Accuracy=0.7193, over 5657.22 frames. ], batch size: 9, lr: 2.44e-03
2024-08-06 22:13:24,940 INFO [trainer.py:765] (4/8) Epoch 35, batch 700, train_loss[loss=2.898, NarTop10Accuracy=0.7518, over 5118.00 frames. ], tot_loss[loss=3.035, NarTop10Accuracy=0.7186, over 5734.90 frames. ], batch size: 6, lr: 2.44e-03
2024-08-06 22:14:01,384 INFO [trainer.py:765] (4/8) Epoch 35, batch 800, train_loss[loss=2.891, NarTop10Accuracy=0.7477, over 4965.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7172, over 5792.71 frames. ], batch size: 6, lr: 2.44e-03
2024-08-06 22:14:34,373 INFO [trainer.py:765] (4/8) Epoch 35, batch 900, train_loss[loss=3.11, NarTop10Accuracy=0.6978, over 6111.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7198, over 5801.60 frames. ], batch size: 13, lr: 2.44e-03
2024-08-06 22:15:09,372 INFO [trainer.py:765] (4/8) Epoch 35, batch 1000, train_loss[loss=2.936, NarTop10Accuracy=0.7477, over 6306.00 frames. ], tot_loss[loss=3.036, NarTop10Accuracy=0.718, over 5887.15 frames. ], batch size: 13, lr: 2.43e-03
2024-08-06 22:15:48,495 INFO [trainer.py:765] (4/8) Epoch 35, batch 1100, train_loss[loss=2.988, NarTop10Accuracy=0.7308, over 6894.00 frames. ], tot_loss[loss=3.042, NarTop10Accuracy=0.7169, over 5923.25 frames. ], batch size: 17, lr: 2.43e-03
2024-08-06 22:16:22,484 INFO [trainer.py:765] (4/8) Epoch 35, batch 1200, train_loss[loss=2.962, NarTop10Accuracy=0.7367, over 7143.00 frames. ], tot_loss[loss=3.031, NarTop10Accuracy=0.7193, over 5917.17 frames. ], batch size: 31, lr: 2.43e-03
2024-08-06 22:16:57,061 INFO [trainer.py:765] (4/8) Epoch 35, batch 1300, train_loss[loss=2.952, NarTop10Accuracy=0.7299, over 4383.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7206, over 5987.19 frames. ], batch size: 5, lr: 2.43e-03
2024-08-06 22:17:31,061 INFO [trainer.py:765] (4/8) Epoch 35, batch 1400, train_loss[loss=3.094, NarTop10Accuracy=0.7067, over 6036.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.7192, over 6012.92 frames. ], batch size: 11, lr: 2.43e-03
2024-08-06 22:18:03,062 INFO [trainer.py:765] (4/8) Epoch 35, batch 1500, train_loss[loss=3.011, NarTop10Accuracy=0.7325, over 6375.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7175, over 5956.35 frames. ], batch size: 50, lr: 2.43e-03
2024-08-06 22:18:30,728 INFO [trainer.py:765] (4/8) Epoch 35, batch 1600, train_loss[loss=2.833, NarTop10Accuracy=0.7589, over 7059.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7147, over 5932.81 frames. ], batch size: 22, lr: 2.43e-03
2024-08-06 22:18:57,320 INFO [trainer.py:765] (4/8) Epoch 35, batch 1700, train_loss[loss=2.846, NarTop10Accuracy=0.7632, over 6177.00 frames. ], tot_loss[loss=3.058, NarTop10Accuracy=0.7137, over 5924.68 frames. ], batch size: 13, lr: 2.42e-03
2024-08-06 22:19:23,703 INFO [trainer.py:765] (4/8) Epoch 35, batch 1800, train_loss[loss=3.454, NarTop10Accuracy=0.6254, over 7137.00 frames. ], tot_loss[loss=3.047, NarTop10Accuracy=0.7158, over 5965.68 frames. ], batch size: 22, lr: 2.42e-03
2024-08-06 22:19:50,202 INFO [trainer.py:765] (4/8) Epoch 35, batch 1900, train_loss[loss=3.262, NarTop10Accuracy=0.6741, over 5592.00 frames. ], tot_loss[loss=3.053, NarTop10Accuracy=0.7145, over 5994.65 frames. ], batch size: 53, lr: 2.42e-03
2024-08-06 22:20:15,763 INFO [trainer.py:765] (4/8) Epoch 35, batch 2000, train_loss[loss=2.987, NarTop10Accuracy=0.7268, over 6261.00 frames. ], tot_loss[loss=3.044, NarTop10Accuracy=0.7167, over 5980.51 frames. ], batch size: 50, lr: 2.42e-03
2024-08-06 22:20:41,045 INFO [trainer.py:765] (4/8) Epoch 35, batch 2100, train_loss[loss=2.708, NarTop10Accuracy=0.7814, over 4002.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7156, over 5944.87 frames. ], batch size: 4, lr: 2.42e-03
2024-08-06 22:21:06,226 INFO [trainer.py:765] (4/8) Epoch 35, batch 2200, train_loss[loss=2.994, NarTop10Accuracy=0.7322, over 7293.00 frames. ], tot_loss[loss=3.054, NarTop10Accuracy=0.714, over 6002.49 frames. ], batch size: 31, lr: 2.42e-03
2024-08-06 22:21:31,286 INFO [trainer.py:765] (4/8) Epoch 35, batch 2300, train_loss[loss=2.989, NarTop10Accuracy=0.73, over 5736.00 frames. ], tot_loss[loss=3.055, NarTop10Accuracy=0.7143, over 6011.68 frames. ], batch size: 9, lr: 2.42e-03
2024-08-06 22:21:55,648 INFO [trainer.py:765] (4/8) Epoch 35, batch 2400, train_loss[loss=3.248, NarTop10Accuracy=0.676, over 5208.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7155, over 5761.82 frames. ], batch size: 7, lr: 2.42e-03
2024-08-06 22:21:59,681 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:22:07,656 INFO [trainer.py:811] (4/8) Epoch 35, validation: loss=2.905, NarTop10Accuracy=0.7437, over 1905321.00 frames. 
2024-08-06 22:22:07,657 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:22:08,116 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.895e+02 2.316e+02 2.462e+02 2.653e+02 5.566e+02, threshold=4.923e+02, percent-clipped=0.1
2024-08-06 22:22:27,127 INFO [trainer.py:765] (4/8) Epoch 35, batch 2500, train_loss[loss=3.238, NarTop10Accuracy=0.6808, over 5049.00 frames. ], tot_loss[loss=3.026, NarTop10Accuracy=0.7197, over 5468.57 frames. ], batch size: 7, lr: 2.41e-03
2024-08-06 22:22:47,167 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 22:23:47,172 INFO [trainer.py:765] (4/8) Epoch 36, batch 100, train_loss[loss=3.256, NarTop10Accuracy=0.6793, over 7224.00 frames. ], tot_loss[loss=2.99, NarTop10Accuracy=0.7275, over 2363.42 frames. ], batch size: 31, lr: 2.38e-03
2024-08-06 22:24:22,494 INFO [trainer.py:765] (4/8) Epoch 36, batch 200, train_loss[loss=2.847, NarTop10Accuracy=0.7521, over 6900.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.7196, over 3848.51 frames. ], batch size: 17, lr: 2.38e-03
2024-08-06 22:24:54,721 INFO [trainer.py:765] (4/8) Epoch 36, batch 300, train_loss[loss=3.287, NarTop10Accuracy=0.6656, over 7002.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7177, over 4658.37 frames. ], batch size: 22, lr: 2.37e-03
2024-08-06 22:25:29,276 INFO [trainer.py:765] (4/8) Epoch 36, batch 400, train_loss[loss=2.992, NarTop10Accuracy=0.7297, over 5112.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7211, over 5099.85 frames. ], batch size: 7, lr: 2.37e-03
2024-08-06 22:26:01,818 INFO [trainer.py:765] (4/8) Epoch 36, batch 500, train_loss[loss=3.293, NarTop10Accuracy=0.6637, over 6066.00 frames. ], tot_loss[loss=3.022, NarTop10Accuracy=0.7209, over 5382.08 frames. ], batch size: 11, lr: 2.37e-03
2024-08-06 22:26:35,025 INFO [trainer.py:765] (4/8) Epoch 36, batch 600, train_loss[loss=3.018, NarTop10Accuracy=0.7316, over 5658.00 frames. ], tot_loss[loss=3.02, NarTop10Accuracy=0.7212, over 5654.37 frames. ], batch size: 9, lr: 2.37e-03
2024-08-06 22:27:10,990 INFO [trainer.py:765] (4/8) Epoch 36, batch 700, train_loss[loss=3.196, NarTop10Accuracy=0.6964, over 5049.00 frames. ], tot_loss[loss=3.022, NarTop10Accuracy=0.721, over 5738.49 frames. ], batch size: 6, lr: 2.37e-03
2024-08-06 22:27:44,915 INFO [trainer.py:765] (4/8) Epoch 36, batch 800, train_loss[loss=3.217, NarTop10Accuracy=0.6862, over 5043.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7177, over 5777.18 frames. ], batch size: 6, lr: 2.37e-03
2024-08-06 22:28:17,812 INFO [trainer.py:765] (4/8) Epoch 36, batch 900, train_loss[loss=2.855, NarTop10Accuracy=0.7642, over 6204.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.7198, over 5786.21 frames. ], batch size: 13, lr: 2.37e-03
2024-08-06 22:28:56,983 INFO [trainer.py:765] (4/8) Epoch 36, batch 1000, train_loss[loss=3.418, NarTop10Accuracy=0.6336, over 6255.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.7201, over 5906.80 frames. ], batch size: 13, lr: 2.37e-03
2024-08-06 22:29:29,364 INFO [trainer.py:765] (4/8) Epoch 36, batch 1100, train_loss[loss=2.949, NarTop10Accuracy=0.7363, over 6585.00 frames. ], tot_loss[loss=3.032, NarTop10Accuracy=0.719, over 5936.96 frames. ], batch size: 17, lr: 2.36e-03
2024-08-06 22:30:05,681 INFO [trainer.py:765] (4/8) Epoch 36, batch 1200, train_loss[loss=3.016, NarTop10Accuracy=0.722, over 7455.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.7199, over 5931.55 frames. ], batch size: 31, lr: 2.36e-03
2024-08-06 22:30:42,576 INFO [trainer.py:765] (4/8) Epoch 36, batch 1300, train_loss[loss=2.925, NarTop10Accuracy=0.7372, over 5067.00 frames. ], tot_loss[loss=3.034, NarTop10Accuracy=0.7182, over 5999.44 frames. ], batch size: 6, lr: 2.36e-03
2024-08-06 22:31:15,938 INFO [trainer.py:765] (4/8) Epoch 36, batch 1400, train_loss[loss=3.06, NarTop10Accuracy=0.707, over 5997.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7217, over 6029.95 frames. ], batch size: 11, lr: 2.36e-03
2024-08-06 22:31:43,748 INFO [trainer.py:765] (4/8) Epoch 36, batch 1500, train_loss[loss=3.441, NarTop10Accuracy=0.6392, over 6324.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.7197, over 5959.34 frames. ], batch size: 51, lr: 2.36e-03
2024-08-06 22:32:11,460 INFO [trainer.py:765] (4/8) Epoch 36, batch 1600, train_loss[loss=3.302, NarTop10Accuracy=0.6601, over 7011.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.72, over 5919.85 frames. ], batch size: 22, lr: 2.36e-03
2024-08-06 22:32:38,109 INFO [trainer.py:765] (4/8) Epoch 36, batch 1700, train_loss[loss=3.277, NarTop10Accuracy=0.6729, over 6228.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7163, over 5897.19 frames. ], batch size: 13, lr: 2.36e-03
2024-08-06 22:33:04,554 INFO [trainer.py:765] (4/8) Epoch 36, batch 1800, train_loss[loss=3.263, NarTop10Accuracy=0.6752, over 6879.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7177, over 5969.22 frames. ], batch size: 22, lr: 2.36e-03
2024-08-06 22:33:15,170 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:33:23,567 INFO [trainer.py:811] (4/8) Epoch 36, validation: loss=2.897, NarTop10Accuracy=0.7457, over 1905321.00 frames. 
2024-08-06 22:33:23,568 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:33:24,096 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.876e+02 2.309e+02 2.476e+02 2.664e+02 4.811e+02, threshold=4.951e+02, percent-clipped=0.0
2024-08-06 22:33:39,456 INFO [trainer.py:765] (4/8) Epoch 36, batch 1900, train_loss[loss=3.03, NarTop10Accuracy=0.7155, over 6357.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7174, over 6032.58 frames. ], batch size: 52, lr: 2.35e-03
2024-08-06 22:34:05,077 INFO [trainer.py:765] (4/8) Epoch 36, batch 2000, train_loss[loss=3.121, NarTop10Accuracy=0.7044, over 6417.00 frames. ], tot_loss[loss=3.035, NarTop10Accuracy=0.718, over 6007.57 frames. ], batch size: 50, lr: 2.35e-03
2024-08-06 22:34:30,514 INFO [trainer.py:765] (4/8) Epoch 36, batch 2100, train_loss[loss=2.574, NarTop10Accuracy=0.8028, over 3978.00 frames. ], tot_loss[loss=3.024, NarTop10Accuracy=0.7205, over 5974.25 frames. ], batch size: 4, lr: 2.35e-03
2024-08-06 22:34:55,938 INFO [trainer.py:765] (4/8) Epoch 36, batch 2200, train_loss[loss=3.476, NarTop10Accuracy=0.6317, over 7347.00 frames. ], tot_loss[loss=3.043, NarTop10Accuracy=0.7164, over 6008.11 frames. ], batch size: 31, lr: 2.35e-03
2024-08-06 22:35:21,145 INFO [trainer.py:765] (4/8) Epoch 36, batch 2300, train_loss[loss=3.308, NarTop10Accuracy=0.6717, over 5826.00 frames. ], tot_loss[loss=3.053, NarTop10Accuracy=0.7147, over 6010.95 frames. ], batch size: 9, lr: 2.35e-03
2024-08-06 22:35:45,600 INFO [trainer.py:765] (4/8) Epoch 36, batch 2400, train_loss[loss=3.27, NarTop10Accuracy=0.6749, over 5208.00 frames. ], tot_loss[loss=3.041, NarTop10Accuracy=0.7172, over 5766.04 frames. ], batch size: 7, lr: 2.35e-03
2024-08-06 22:36:09,182 INFO [trainer.py:765] (4/8) Epoch 36, batch 2500, train_loss[loss=2.748, NarTop10Accuracy=0.7793, over 5154.00 frames. ], tot_loss[loss=3.02, NarTop10Accuracy=0.7209, over 5463.98 frames. ], batch size: 7, lr: 2.35e-03
2024-08-06 22:36:28,978 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 22:37:29,725 INFO [trainer.py:765] (4/8) Epoch 37, batch 100, train_loss[loss=2.774, NarTop10Accuracy=0.7764, over 7125.00 frames. ], tot_loss[loss=3.046, NarTop10Accuracy=0.7163, over 2350.12 frames. ], batch size: 31, lr: 2.31e-03
2024-08-06 22:38:01,272 INFO [trainer.py:765] (4/8) Epoch 37, batch 200, train_loss[loss=2.85, NarTop10Accuracy=0.7613, over 6780.00 frames. ], tot_loss[loss=3.029, NarTop10Accuracy=0.7196, over 3853.10 frames. ], batch size: 17, lr: 2.31e-03
2024-08-06 22:38:35,956 INFO [trainer.py:765] (4/8) Epoch 37, batch 300, train_loss[loss=3.259, NarTop10Accuracy=0.6754, over 7260.00 frames. ], tot_loss[loss=3.022, NarTop10Accuracy=0.7207, over 4659.06 frames. ], batch size: 22, lr: 2.31e-03
2024-08-06 22:39:09,307 INFO [trainer.py:765] (4/8) Epoch 37, batch 400, train_loss[loss=2.609, NarTop10Accuracy=0.8035, over 5199.00 frames. ], tot_loss[loss=3.007, NarTop10Accuracy=0.724, over 5100.49 frames. ], batch size: 7, lr: 2.31e-03
2024-08-06 22:39:43,861 INFO [trainer.py:765] (4/8) Epoch 37, batch 500, train_loss[loss=3.402, NarTop10Accuracy=0.6435, over 6150.00 frames. ], tot_loss[loss=3.01, NarTop10Accuracy=0.7234, over 5372.75 frames. ], batch size: 11, lr: 2.31e-03
2024-08-06 22:40:17,333 INFO [trainer.py:765] (4/8) Epoch 37, batch 600, train_loss[loss=2.672, NarTop10Accuracy=0.7928, over 5586.00 frames. ], tot_loss[loss=3.015, NarTop10Accuracy=0.7224, over 5637.45 frames. ], batch size: 9, lr: 2.31e-03
2024-08-06 22:40:51,615 INFO [trainer.py:765] (4/8) Epoch 37, batch 700, train_loss[loss=3.139, NarTop10Accuracy=0.6935, over 4335.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7176, over 5708.04 frames. ], batch size: 5, lr: 2.30e-03
2024-08-06 22:41:30,564 INFO [trainer.py:765] (4/8) Epoch 37, batch 800, train_loss[loss=2.737, NarTop10Accuracy=0.7784, over 5271.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7174, over 5786.69 frames. ], batch size: 6, lr: 2.30e-03
2024-08-06 22:41:59,083 INFO [trainer.py:765] (4/8) Epoch 37, batch 900, train_loss[loss=2.79, NarTop10Accuracy=0.7692, over 6327.00 frames. ], tot_loss[loss=3.021, NarTop10Accuracy=0.7211, over 5807.61 frames. ], batch size: 13, lr: 2.30e-03
2024-08-06 22:42:38,267 INFO [trainer.py:765] (4/8) Epoch 37, batch 1000, train_loss[loss=3.071, NarTop10Accuracy=0.7094, over 6096.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7177, over 5911.43 frames. ], batch size: 13, lr: 2.30e-03
2024-08-06 22:43:15,907 INFO [trainer.py:765] (4/8) Epoch 37, batch 1100, train_loss[loss=2.937, NarTop10Accuracy=0.7423, over 6879.00 frames. ], tot_loss[loss=3.038, NarTop10Accuracy=0.7177, over 5946.54 frames. ], batch size: 17, lr: 2.30e-03
2024-08-06 22:43:47,740 INFO [trainer.py:765] (4/8) Epoch 37, batch 1200, train_loss[loss=2.812, NarTop10Accuracy=0.7608, over 7407.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.717, over 5918.52 frames. ], batch size: 32, lr: 2.30e-03
2024-08-06 22:44:11,754 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:44:20,075 INFO [trainer.py:811] (4/8) Epoch 37, validation: loss=2.92, NarTop10Accuracy=0.7407, over 1905321.00 frames. 
2024-08-06 22:44:20,076 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:44:20,606 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.887e+02 2.309e+02 2.481e+02 2.647e+02 8.766e+02, threshold=4.961e+02, percent-clipped=0.1
2024-08-06 22:44:32,784 INFO [trainer.py:765] (4/8) Epoch 37, batch 1300, train_loss[loss=2.777, NarTop10Accuracy=0.7739, over 5061.00 frames. ], tot_loss[loss=3.024, NarTop10Accuracy=0.7204, over 5985.15 frames. ], batch size: 6, lr: 2.30e-03
2024-08-06 22:45:10,388 INFO [trainer.py:765] (4/8) Epoch 37, batch 1400, train_loss[loss=2.722, NarTop10Accuracy=0.7788, over 6084.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7207, over 6018.55 frames. ], batch size: 11, lr: 2.30e-03
2024-08-06 22:45:40,513 INFO [trainer.py:765] (4/8) Epoch 37, batch 1500, train_loss[loss=3.005, NarTop10Accuracy=0.722, over 5826.00 frames. ], tot_loss[loss=3.033, NarTop10Accuracy=0.7187, over 5946.77 frames. ], batch size: 50, lr: 2.29e-03
2024-08-06 22:46:08,438 INFO [trainer.py:765] (4/8) Epoch 37, batch 1600, train_loss[loss=3.347, NarTop10Accuracy=0.6523, over 7125.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7173, over 5936.38 frames. ], batch size: 22, lr: 2.29e-03
2024-08-06 22:46:35,187 INFO [trainer.py:765] (4/8) Epoch 37, batch 1700, train_loss[loss=3.292, NarTop10Accuracy=0.6585, over 6486.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7181, over 5932.24 frames. ], batch size: 14, lr: 2.29e-03
2024-08-06 22:47:01,793 INFO [trainer.py:765] (4/8) Epoch 37, batch 1800, train_loss[loss=2.829, NarTop10Accuracy=0.7642, over 7197.00 frames. ], tot_loss[loss=3.032, NarTop10Accuracy=0.7189, over 5993.70 frames. ], batch size: 22, lr: 2.29e-03
2024-08-06 22:47:28,312 INFO [trainer.py:765] (4/8) Epoch 37, batch 1900, train_loss[loss=3.092, NarTop10Accuracy=0.7075, over 5967.00 frames. ], tot_loss[loss=3.035, NarTop10Accuracy=0.7183, over 6038.93 frames. ], batch size: 51, lr: 2.29e-03
2024-08-06 22:47:53,925 INFO [trainer.py:765] (4/8) Epoch 37, batch 2000, train_loss[loss=3.2, NarTop10Accuracy=0.6834, over 6075.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.72, over 6010.77 frames. ], batch size: 50, lr: 2.29e-03
2024-08-06 22:48:19,326 INFO [trainer.py:765] (4/8) Epoch 37, batch 2100, train_loss[loss=3.011, NarTop10Accuracy=0.7258, over 4746.00 frames. ], tot_loss[loss=3.036, NarTop10Accuracy=0.7187, over 5969.97 frames. ], batch size: 5, lr: 2.29e-03
2024-08-06 22:48:44,707 INFO [trainer.py:765] (4/8) Epoch 37, batch 2200, train_loss[loss=2.993, NarTop10Accuracy=0.7209, over 7194.00 frames. ], tot_loss[loss=3.05, NarTop10Accuracy=0.7156, over 6003.60 frames. ], batch size: 31, lr: 2.29e-03
2024-08-06 22:49:09,913 INFO [trainer.py:765] (4/8) Epoch 37, batch 2300, train_loss[loss=2.799, NarTop10Accuracy=0.7735, over 5745.00 frames. ], tot_loss[loss=3.051, NarTop10Accuracy=0.7157, over 6008.40 frames. ], batch size: 9, lr: 2.29e-03
2024-08-06 22:49:34,319 INFO [trainer.py:765] (4/8) Epoch 37, batch 2400, train_loss[loss=3.236, NarTop10Accuracy=0.6728, over 5298.00 frames. ], tot_loss[loss=3.028, NarTop10Accuracy=0.7198, over 5758.04 frames. ], batch size: 7, lr: 2.28e-03
2024-08-06 22:49:57,861 INFO [trainer.py:765] (4/8) Epoch 37, batch 2500, train_loss[loss=3.182, NarTop10Accuracy=0.6964, over 5145.00 frames. ], tot_loss[loss=2.997, NarTop10Accuracy=0.7258, over 5471.14 frames. ], batch size: 7, lr: 2.28e-03
2024-08-06 22:50:18,227 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 22:51:16,151 INFO [trainer.py:765] (4/8) Epoch 38, batch 100, train_loss[loss=3.046, NarTop10Accuracy=0.7191, over 7260.00 frames. ], tot_loss[loss=3.007, NarTop10Accuracy=0.7237, over 2366.02 frames. ], batch size: 31, lr: 2.25e-03
2024-08-06 22:51:53,013 INFO [trainer.py:765] (4/8) Epoch 38, batch 200, train_loss[loss=3.255, NarTop10Accuracy=0.6808, over 6918.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7217, over 3844.48 frames. ], batch size: 17, lr: 2.25e-03
2024-08-06 22:52:25,202 INFO [trainer.py:765] (4/8) Epoch 38, batch 300, train_loss[loss=2.92, NarTop10Accuracy=0.7453, over 7272.00 frames. ], tot_loss[loss=3.036, NarTop10Accuracy=0.7182, over 4655.82 frames. ], batch size: 23, lr: 2.25e-03
2024-08-06 22:52:55,626 INFO [trainer.py:765] (4/8) Epoch 38, batch 400, train_loss[loss=3.037, NarTop10Accuracy=0.7152, over 5118.00 frames. ], tot_loss[loss=3.014, NarTop10Accuracy=0.7224, over 5116.61 frames. ], batch size: 7, lr: 2.25e-03
2024-08-06 22:53:32,228 INFO [trainer.py:765] (4/8) Epoch 38, batch 500, train_loss[loss=2.857, NarTop10Accuracy=0.7574, over 6066.00 frames. ], tot_loss[loss=2.987, NarTop10Accuracy=0.7278, over 5387.24 frames. ], batch size: 11, lr: 2.25e-03
2024-08-06 22:54:05,497 INFO [trainer.py:765] (4/8) Epoch 38, batch 600, train_loss[loss=3.034, NarTop10Accuracy=0.7125, over 5754.00 frames. ], tot_loss[loss=3, NarTop10Accuracy=0.7253, over 5656.67 frames. ], batch size: 9, lr: 2.24e-03
2024-08-06 22:54:36,002 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 22:54:43,918 INFO [trainer.py:811] (4/8) Epoch 38, validation: loss=2.939, NarTop10Accuracy=0.7369, over 1905321.00 frames. 
2024-08-06 22:54:43,919 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 22:54:44,427 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.880e+02 2.313e+02 2.478e+02 2.663e+02 7.254e+02, threshold=4.957e+02, percent-clipped=0.3
2024-08-06 22:54:46,659 INFO [trainer.py:765] (4/8) Epoch 38, batch 700, train_loss[loss=2.668, NarTop10Accuracy=0.7802, over 5130.00 frames. ], tot_loss[loss=3.007, NarTop10Accuracy=0.7239, over 5728.30 frames. ], batch size: 6, lr: 2.24e-03
2024-08-06 22:55:24,938 INFO [trainer.py:765] (4/8) Epoch 38, batch 800, train_loss[loss=2.785, NarTop10Accuracy=0.7618, over 5049.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.72, over 5797.30 frames. ], batch size: 6, lr: 2.24e-03
2024-08-06 22:55:59,704 INFO [trainer.py:765] (4/8) Epoch 38, batch 900, train_loss[loss=2.872, NarTop10Accuracy=0.7542, over 6129.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7212, over 5818.39 frames. ], batch size: 13, lr: 2.24e-03
2024-08-06 22:56:32,090 INFO [trainer.py:765] (4/8) Epoch 38, batch 1000, train_loss[loss=3.355, NarTop10Accuracy=0.6465, over 6771.00 frames. ], tot_loss[loss=3.02, NarTop10Accuracy=0.7211, over 5916.58 frames. ], batch size: 14, lr: 2.24e-03
2024-08-06 22:57:08,991 INFO [trainer.py:765] (4/8) Epoch 38, batch 1100, train_loss[loss=3.092, NarTop10Accuracy=0.7051, over 6693.00 frames. ], tot_loss[loss=3.029, NarTop10Accuracy=0.7189, over 5958.97 frames. ], batch size: 17, lr: 2.24e-03
2024-08-06 22:57:42,662 INFO [trainer.py:765] (4/8) Epoch 38, batch 1200, train_loss[loss=2.827, NarTop10Accuracy=0.7617, over 7359.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.7196, over 5960.56 frames. ], batch size: 31, lr: 2.24e-03
2024-08-06 22:58:16,546 INFO [trainer.py:765] (4/8) Epoch 38, batch 1300, train_loss[loss=3.165, NarTop10Accuracy=0.6859, over 4941.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.7191, over 6005.81 frames. ], batch size: 6, lr: 2.24e-03
2024-08-06 22:58:49,810 INFO [trainer.py:765] (4/8) Epoch 38, batch 1400, train_loss[loss=2.849, NarTop10Accuracy=0.754, over 6042.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7149, over 6032.75 frames. ], batch size: 11, lr: 2.23e-03
2024-08-06 22:59:22,854 INFO [trainer.py:765] (4/8) Epoch 38, batch 1500, train_loss[loss=3.474, NarTop10Accuracy=0.6243, over 6039.00 frames. ], tot_loss[loss=3.029, NarTop10Accuracy=0.7188, over 5959.78 frames. ], batch size: 50, lr: 2.23e-03
2024-08-06 22:59:50,644 INFO [trainer.py:765] (4/8) Epoch 38, batch 1600, train_loss[loss=3.417, NarTop10Accuracy=0.6434, over 7365.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7177, over 5934.91 frames. ], batch size: 23, lr: 2.23e-03
2024-08-06 23:00:17,315 INFO [trainer.py:765] (4/8) Epoch 38, batch 1700, train_loss[loss=2.993, NarTop10Accuracy=0.7326, over 6726.00 frames. ], tot_loss[loss=3.049, NarTop10Accuracy=0.7149, over 5933.25 frames. ], batch size: 14, lr: 2.23e-03
2024-08-06 23:00:43,764 INFO [trainer.py:765] (4/8) Epoch 38, batch 1800, train_loss[loss=3.243, NarTop10Accuracy=0.6749, over 6867.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7153, over 6000.59 frames. ], batch size: 22, lr: 2.23e-03
2024-08-06 23:01:10,192 INFO [trainer.py:765] (4/8) Epoch 38, batch 1900, train_loss[loss=3.477, NarTop10Accuracy=0.6315, over 6237.00 frames. ], tot_loss[loss=3.055, NarTop10Accuracy=0.7143, over 6031.98 frames. ], batch size: 50, lr: 2.23e-03
2024-08-06 23:01:35,681 INFO [trainer.py:765] (4/8) Epoch 38, batch 2000, train_loss[loss=3.302, NarTop10Accuracy=0.6634, over 6219.00 frames. ], tot_loss[loss=3.052, NarTop10Accuracy=0.7145, over 6001.05 frames. ], batch size: 50, lr: 2.23e-03
2024-08-06 23:02:01,050 INFO [trainer.py:765] (4/8) Epoch 38, batch 2100, train_loss[loss=2.835, NarTop10Accuracy=0.7541, over 3999.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7171, over 5987.19 frames. ], batch size: 4, lr: 2.23e-03
2024-08-06 23:02:26,314 INFO [trainer.py:765] (4/8) Epoch 38, batch 2200, train_loss[loss=2.838, NarTop10Accuracy=0.7512, over 7464.00 frames. ], tot_loss[loss=3.039, NarTop10Accuracy=0.7173, over 6041.19 frames. ], batch size: 31, lr: 2.23e-03
2024-08-06 23:02:51,419 INFO [trainer.py:765] (4/8) Epoch 38, batch 2300, train_loss[loss=2.624, NarTop10Accuracy=0.7938, over 5736.00 frames. ], tot_loss[loss=3.037, NarTop10Accuracy=0.7178, over 6038.02 frames. ], batch size: 9, lr: 2.22e-03
2024-08-06 23:03:16,348 INFO [trainer.py:765] (4/8) Epoch 38, batch 2400, train_loss[loss=2.684, NarTop10Accuracy=0.7881, over 5298.00 frames. ], tot_loss[loss=3.029, NarTop10Accuracy=0.719, over 5779.75 frames. ], batch size: 7, lr: 2.22e-03
2024-08-06 23:03:39,824 INFO [trainer.py:765] (4/8) Epoch 38, batch 2500, train_loss[loss=3.045, NarTop10Accuracy=0.7117, over 5280.00 frames. ], tot_loss[loss=3.004, NarTop10Accuracy=0.724, over 5484.26 frames. ], batch size: 7, lr: 2.22e-03
2024-08-06 23:03:59,757 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 23:04:58,941 INFO [trainer.py:765] (4/8) Epoch 39, batch 100, train_loss[loss=3.25, NarTop10Accuracy=0.6796, over 7557.00 frames. ], tot_loss[loss=2.982, NarTop10Accuracy=0.7295, over 2381.74 frames. ], batch size: 32, lr: 2.19e-03
2024-08-06 23:05:03,469 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 23:05:11,563 INFO [trainer.py:811] (4/8) Epoch 39, validation: loss=2.9, NarTop10Accuracy=0.7445, over 1905321.00 frames. 
2024-08-06 23:05:11,564 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 23:05:12,137 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.911e+02 2.316e+02 2.500e+02 2.688e+02 4.683e+02, threshold=5.001e+02, percent-clipped=0.0
2024-08-06 23:05:40,163 INFO [trainer.py:765] (4/8) Epoch 39, batch 200, train_loss[loss=2.776, NarTop10Accuracy=0.773, over 6729.00 frames. ], tot_loss[loss=2.991, NarTop10Accuracy=0.7274, over 3856.35 frames. ], batch size: 17, lr: 2.19e-03
2024-08-06 23:06:17,293 INFO [trainer.py:765] (4/8) Epoch 39, batch 300, train_loss[loss=3.1, NarTop10Accuracy=0.7136, over 6801.00 frames. ], tot_loss[loss=2.988, NarTop10Accuracy=0.7277, over 4658.06 frames. ], batch size: 22, lr: 2.19e-03
2024-08-06 23:06:48,275 INFO [trainer.py:765] (4/8) Epoch 39, batch 400, train_loss[loss=2.736, NarTop10Accuracy=0.7748, over 5199.00 frames. ], tot_loss[loss=2.987, NarTop10Accuracy=0.7278, over 5109.90 frames. ], batch size: 7, lr: 2.19e-03
2024-08-06 23:07:19,174 INFO [trainer.py:765] (4/8) Epoch 39, batch 500, train_loss[loss=3.329, NarTop10Accuracy=0.6564, over 6060.00 frames. ], tot_loss[loss=2.994, NarTop10Accuracy=0.7261, over 5408.21 frames. ], batch size: 11, lr: 2.19e-03
2024-08-06 23:07:52,563 INFO [trainer.py:765] (4/8) Epoch 39, batch 600, train_loss[loss=2.702, NarTop10Accuracy=0.7824, over 5757.00 frames. ], tot_loss[loss=3.011, NarTop10Accuracy=0.7229, over 5667.27 frames. ], batch size: 9, lr: 2.19e-03
2024-08-06 23:08:33,694 INFO [trainer.py:765] (4/8) Epoch 39, batch 700, train_loss[loss=3.187, NarTop10Accuracy=0.6878, over 5085.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7206, over 5735.75 frames. ], batch size: 6, lr: 2.18e-03
2024-08-06 23:09:05,861 INFO [trainer.py:765] (4/8) Epoch 39, batch 800, train_loss[loss=2.653, NarTop10Accuracy=0.7938, over 4350.00 frames. ], tot_loss[loss=3.027, NarTop10Accuracy=0.72, over 5777.27 frames. ], batch size: 5, lr: 2.18e-03
2024-08-06 23:09:38,865 INFO [trainer.py:765] (4/8) Epoch 39, batch 900, train_loss[loss=3.358, NarTop10Accuracy=0.6549, over 6180.00 frames. ], tot_loss[loss=3.024, NarTop10Accuracy=0.7204, over 5794.32 frames. ], batch size: 13, lr: 2.18e-03
2024-08-06 23:10:18,460 INFO [trainer.py:765] (4/8) Epoch 39, batch 1000, train_loss[loss=2.909, NarTop10Accuracy=0.7347, over 6672.00 frames. ], tot_loss[loss=3.015, NarTop10Accuracy=0.7221, over 5911.58 frames. ], batch size: 14, lr: 2.18e-03
2024-08-06 23:10:53,934 INFO [trainer.py:765] (4/8) Epoch 39, batch 1100, train_loss[loss=2.748, NarTop10Accuracy=0.7813, over 6825.00 frames. ], tot_loss[loss=3.033, NarTop10Accuracy=0.7189, over 5949.91 frames. ], batch size: 17, lr: 2.18e-03
2024-08-06 23:11:27,822 INFO [trainer.py:765] (4/8) Epoch 39, batch 1200, train_loss[loss=2.883, NarTop10Accuracy=0.7485, over 7017.00 frames. ], tot_loss[loss=3.022, NarTop10Accuracy=0.7203, over 5956.34 frames. ], batch size: 31, lr: 2.18e-03
2024-08-06 23:12:07,252 INFO [trainer.py:765] (4/8) Epoch 39, batch 1300, train_loss[loss=2.903, NarTop10Accuracy=0.7553, over 5229.00 frames. ], tot_loss[loss=3.016, NarTop10Accuracy=0.7219, over 6024.80 frames. ], batch size: 6, lr: 2.18e-03
2024-08-06 23:12:39,301 INFO [trainer.py:765] (4/8) Epoch 39, batch 1400, train_loss[loss=3.088, NarTop10Accuracy=0.7108, over 6105.00 frames. ], tot_loss[loss=3.018, NarTop10Accuracy=0.7215, over 6033.38 frames. ], batch size: 11, lr: 2.18e-03
2024-08-06 23:13:09,756 INFO [trainer.py:765] (4/8) Epoch 39, batch 1500, train_loss[loss=3.626, NarTop10Accuracy=0.5982, over 6585.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7211, over 5961.23 frames. ], batch size: 53, lr: 2.18e-03
2024-08-06 23:13:37,586 INFO [trainer.py:765] (4/8) Epoch 39, batch 1600, train_loss[loss=2.884, NarTop10Accuracy=0.7564, over 7107.00 frames. ], tot_loss[loss=3.009, NarTop10Accuracy=0.7234, over 5944.35 frames. ], batch size: 22, lr: 2.17e-03
2024-08-06 23:14:04,219 INFO [trainer.py:765] (4/8) Epoch 39, batch 1700, train_loss[loss=3.396, NarTop10Accuracy=0.6466, over 6672.00 frames. ], tot_loss[loss=3.031, NarTop10Accuracy=0.7191, over 5911.76 frames. ], batch size: 14, lr: 2.17e-03
2024-08-06 23:14:30,767 INFO [trainer.py:765] (4/8) Epoch 39, batch 1800, train_loss[loss=2.769, NarTop10Accuracy=0.7785, over 7212.00 frames. ], tot_loss[loss=3.04, NarTop10Accuracy=0.7175, over 5969.11 frames. ], batch size: 22, lr: 2.17e-03
2024-08-06 23:14:57,179 INFO [trainer.py:765] (4/8) Epoch 39, batch 1900, train_loss[loss=2.98, NarTop10Accuracy=0.7312, over 5820.00 frames. ], tot_loss[loss=3.048, NarTop10Accuracy=0.7156, over 6026.19 frames. ], batch size: 50, lr: 2.17e-03
2024-08-06 23:15:22,750 INFO [trainer.py:765] (4/8) Epoch 39, batch 2000, train_loss[loss=3.241, NarTop10Accuracy=0.6778, over 6213.00 frames. ], tot_loss[loss=3.033, NarTop10Accuracy=0.7188, over 6010.64 frames. ], batch size: 50, lr: 2.17e-03
2024-08-06 23:15:48,060 INFO [trainer.py:765] (4/8) Epoch 39, batch 2100, train_loss[loss=3.329, NarTop10Accuracy=0.6588, over 3831.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7201, over 5994.50 frames. ], batch size: 4, lr: 2.17e-03
2024-08-06 23:15:51,871 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 23:16:02,156 INFO [trainer.py:811] (4/8) Epoch 39, validation: loss=2.85, NarTop10Accuracy=0.7552, over 1905321.00 frames. 
2024-08-06 23:16:02,156 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 23:16:02,645 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.940e+02 2.369e+02 2.530e+02 2.720e+02 6.127e+02, threshold=5.059e+02, percent-clipped=0.2
2024-08-06 23:16:23,652 INFO [trainer.py:765] (4/8) Epoch 39, batch 2200, train_loss[loss=3.174, NarTop10Accuracy=0.6927, over 7446.00 frames. ], tot_loss[loss=3.03, NarTop10Accuracy=0.719, over 6034.82 frames. ], batch size: 31, lr: 2.17e-03
2024-08-06 23:16:48,847 INFO [trainer.py:765] (4/8) Epoch 39, batch 2300, train_loss[loss=2.651, NarTop10Accuracy=0.7979, over 5820.00 frames. ], tot_loss[loss=3.041, NarTop10Accuracy=0.7169, over 6026.39 frames. ], batch size: 9, lr: 2.17e-03
2024-08-06 23:17:13,136 INFO [trainer.py:765] (4/8) Epoch 39, batch 2400, train_loss[loss=2.687, NarTop10Accuracy=0.7792, over 4977.00 frames. ], tot_loss[loss=3.013, NarTop10Accuracy=0.7226, over 5774.96 frames. ], batch size: 7, lr: 2.17e-03
2024-08-06 23:17:36,712 INFO [trainer.py:765] (4/8) Epoch 39, batch 2500, train_loss[loss=2.917, NarTop10Accuracy=0.7547, over 5175.00 frames. ], tot_loss[loss=2.99, NarTop10Accuracy=0.7271, over 5474.43 frames. ], batch size: 7, lr: 2.16e-03
2024-08-06 23:17:56,403 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 23:18:48,946 INFO [trainer.py:765] (4/8) Epoch 40, batch 100, train_loss[loss=2.963, NarTop10Accuracy=0.729, over 6978.00 frames. ], tot_loss[loss=3.003, NarTop10Accuracy=0.7239, over 2358.12 frames. ], batch size: 31, lr: 2.14e-03
2024-08-06 23:19:23,035 INFO [trainer.py:765] (4/8) Epoch 40, batch 200, train_loss[loss=2.831, NarTop10Accuracy=0.7614, over 6837.00 frames. ], tot_loss[loss=2.994, NarTop10Accuracy=0.7264, over 3850.55 frames. ], batch size: 17, lr: 2.13e-03
2024-08-06 23:19:57,187 INFO [trainer.py:765] (4/8) Epoch 40, batch 300, train_loss[loss=2.838, NarTop10Accuracy=0.7656, over 6990.00 frames. ], tot_loss[loss=3.014, NarTop10Accuracy=0.722, over 4663.71 frames. ], batch size: 22, lr: 2.13e-03
2024-08-06 23:20:30,182 INFO [trainer.py:765] (4/8) Epoch 40, batch 400, train_loss[loss=2.784, NarTop10Accuracy=0.7696, over 5073.00 frames. ], tot_loss[loss=3.015, NarTop10Accuracy=0.7225, over 5117.30 frames. ], batch size: 7, lr: 2.13e-03
2024-08-06 23:21:00,250 INFO [trainer.py:765] (4/8) Epoch 40, batch 500, train_loss[loss=2.649, NarTop10Accuracy=0.7964, over 6165.00 frames. ], tot_loss[loss=3.008, NarTop10Accuracy=0.7236, over 5396.49 frames. ], batch size: 11, lr: 2.13e-03
2024-08-06 23:21:34,881 INFO [trainer.py:765] (4/8) Epoch 40, batch 600, train_loss[loss=2.92, NarTop10Accuracy=0.7518, over 5856.00 frames. ], tot_loss[loss=2.998, NarTop10Accuracy=0.7255, over 5649.00 frames. ], batch size: 9, lr: 2.13e-03
2024-08-06 23:22:11,097 INFO [trainer.py:765] (4/8) Epoch 40, batch 700, train_loss[loss=3.109, NarTop10Accuracy=0.7039, over 5190.00 frames. ], tot_loss[loss=3.006, NarTop10Accuracy=0.7243, over 5709.88 frames. ], batch size: 6, lr: 2.13e-03
2024-08-06 23:22:44,753 INFO [trainer.py:765] (4/8) Epoch 40, batch 800, train_loss[loss=2.629, NarTop10Accuracy=0.8094, over 4944.00 frames. ], tot_loss[loss=3.017, NarTop10Accuracy=0.7218, over 5770.95 frames. ], batch size: 6, lr: 2.13e-03
2024-08-06 23:23:16,635 INFO [trainer.py:765] (4/8) Epoch 40, batch 900, train_loss[loss=3.329, NarTop10Accuracy=0.6489, over 6438.00 frames. ], tot_loss[loss=3.019, NarTop10Accuracy=0.7215, over 5794.28 frames. ], batch size: 13, lr: 2.13e-03
2024-08-06 23:23:55,591 INFO [trainer.py:765] (4/8) Epoch 40, batch 1000, train_loss[loss=3.427, NarTop10Accuracy=0.631, over 6711.00 frames. ], tot_loss[loss=3.024, NarTop10Accuracy=0.7201, over 5901.63 frames. ], batch size: 14, lr: 2.13e-03
2024-08-06 23:24:30,208 INFO [trainer.py:765] (4/8) Epoch 40, batch 1100, train_loss[loss=2.648, NarTop10Accuracy=0.7906, over 6789.00 frames. ], tot_loss[loss=3.023, NarTop10Accuracy=0.7206, over 5956.79 frames. ], batch size: 17, lr: 2.12e-03
2024-08-06 23:25:03,090 INFO [trainer.py:765] (4/8) Epoch 40, batch 1200, train_loss[loss=2.958, NarTop10Accuracy=0.7335, over 7242.00 frames. ], tot_loss[loss=3.016, NarTop10Accuracy=0.7221, over 5941.27 frames. ], batch size: 31, lr: 2.12e-03
2024-08-06 23:25:41,842 INFO [trainer.py:765] (4/8) Epoch 40, batch 1300, train_loss[loss=2.857, NarTop10Accuracy=0.749, over 5223.00 frames. ], tot_loss[loss=3.013, NarTop10Accuracy=0.7228, over 6019.29 frames. ], batch size: 6, lr: 2.12e-03
2024-08-06 23:26:13,384 INFO [trainer.py:765] (4/8) Epoch 40, batch 1400, train_loss[loss=2.756, NarTop10Accuracy=0.7748, over 6048.00 frames. ], tot_loss[loss=3.021, NarTop10Accuracy=0.7215, over 6017.88 frames. ], batch size: 11, lr: 2.12e-03
2024-08-06 23:26:43,377 INFO [trainer.py:765] (4/8) Epoch 40, batch 1500, train_loss[loss=3.335, NarTop10Accuracy=0.6556, over 6141.00 frames. ], tot_loss[loss=3.008, NarTop10Accuracy=0.724, over 5953.12 frames. ], batch size: 50, lr: 2.12e-03
2024-08-06 23:26:54,419 INFO [trainer.py:803] (4/8) Computing validation loss
2024-08-06 23:27:02,676 INFO [trainer.py:811] (4/8) Epoch 40, validation: loss=2.86, NarTop10Accuracy=0.7522, over 1905321.00 frames. 
2024-08-06 23:27:02,677 INFO [trainer.py:814] (4/8) Maximum memory allocated so far is 30368MB
2024-08-06 23:27:03,156 INFO [optim.py:386] (4/8) Clipping_scale=2.0, grad-norm quartiles 1.941e+02 2.329e+02 2.511e+02 2.723e+02 1.241e+03, threshold=5.022e+02, percent-clipped=0.2
2024-08-06 23:27:19,382 INFO [trainer.py:765] (4/8) Epoch 40, batch 1600, train_loss[loss=2.971, NarTop10Accuracy=0.7263, over 7011.00 frames. ], tot_loss[loss=3.02, NarTop10Accuracy=0.7215, over 5925.46 frames. ], batch size: 22, lr: 2.12e-03
2024-08-06 23:27:46,056 INFO [trainer.py:765] (4/8) Epoch 40, batch 1700, train_loss[loss=3.33, NarTop10Accuracy=0.6632, over 6609.00 frames. ], tot_loss[loss=3.013, NarTop10Accuracy=0.7224, over 5920.21 frames. ], batch size: 14, lr: 2.12e-03
2024-08-06 23:28:12,579 INFO [trainer.py:765] (4/8) Epoch 40, batch 1800, train_loss[loss=2.982, NarTop10Accuracy=0.7281, over 7161.00 frames. ], tot_loss[loss=2.991, NarTop10Accuracy=0.7265, over 5980.01 frames. ], batch size: 22, lr: 2.12e-03
2024-08-06 23:28:38,909 INFO [trainer.py:765] (4/8) Epoch 40, batch 1900, train_loss[loss=3.187, NarTop10Accuracy=0.6933, over 6129.00 frames. ], tot_loss[loss=3.001, NarTop10Accuracy=0.7252, over 6039.17 frames. ], batch size: 50, lr: 2.12e-03
2024-08-06 23:29:04,445 INFO [trainer.py:765] (4/8) Epoch 40, batch 2000, train_loss[loss=3.565, NarTop10Accuracy=0.5999, over 6261.00 frames. ], tot_loss[loss=3.004, NarTop10Accuracy=0.7243, over 6008.91 frames. ], batch size: 50, lr: 2.12e-03
2024-08-06 23:29:29,750 INFO [trainer.py:765] (4/8) Epoch 40, batch 2100, train_loss[loss=2.694, NarTop10Accuracy=0.7824, over 3882.00 frames. ], tot_loss[loss=3.012, NarTop10Accuracy=0.7227, over 5969.87 frames. ], batch size: 4, lr: 2.11e-03
2024-08-06 23:29:54,939 INFO [trainer.py:765] (4/8) Epoch 40, batch 2200, train_loss[loss=3.24, NarTop10Accuracy=0.6771, over 7335.00 frames. ], tot_loss[loss=3.026, NarTop10Accuracy=0.7199, over 5993.98 frames. ], batch size: 32, lr: 2.11e-03
2024-08-06 23:30:20,013 INFO [trainer.py:765] (4/8) Epoch 40, batch 2300, train_loss[loss=2.869, NarTop10Accuracy=0.7563, over 5808.00 frames. ], tot_loss[loss=3.031, NarTop10Accuracy=0.7192, over 5989.51 frames. ], batch size: 9, lr: 2.11e-03
2024-08-06 23:30:44,296 INFO [trainer.py:765] (4/8) Epoch 40, batch 2400, train_loss[loss=2.792, NarTop10Accuracy=0.7659, over 5118.00 frames. ], tot_loss[loss=3.025, NarTop10Accuracy=0.7203, over 5733.71 frames. ], batch size: 7, lr: 2.11e-03
2024-08-06 23:31:07,738 INFO [trainer.py:765] (4/8) Epoch 40, batch 2500, train_loss[loss=3.047, NarTop10Accuracy=0.7081, over 5169.00 frames. ], tot_loss[loss=2.987, NarTop10Accuracy=0.7276, over 5451.86 frames. ], batch size: 7, lr: 2.11e-03
2024-08-06 23:31:28,035 INFO [trainer.py:650] (4/8) Reaches end of dataloader.
2024-08-06 23:31:28,038 INFO [trainer.py:1069] (4/8) Done!