diff --git "a/exp/log/log-train-2023-03-25-21-42-34-0" "b/exp/log/log-train-2023-03-25-21-42-34-0" new file mode 100644--- /dev/null +++ "b/exp/log/log-train-2023-03-25-21-42-34-0" @@ -0,0 +1,23215 @@ +2023-03-25 21:42:34,503 INFO [finetune.py:1046] (0/7) Training started +2023-03-25 21:42:34,504 INFO [finetune.py:1056] (0/7) Device: cuda:0 +2023-03-25 21:42:34,505 INFO [finetune.py:1065] (0/7) {'frame_shift_ms': 10.0, 'allowed_excess_duration_ratio': 0.1, 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'warm_step': 2000, 'env_info': {'k2-version': '1.23.4', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '62e404dd3f3a811d73e424199b3408e309c06e1a', 'k2-git-date': 'Mon Jan 30 02:26:16 2023', 'lhotse-version': '1.12.0.dev+git.3ccfeb7.clean', 'torch-version': '1.13.0', 'torch-cuda-available': True, 'torch-cuda-version': '11.7', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': 'd74822d-dirty', 'icefall-git-date': 'Tue Mar 21 21:35:32 2023', 'icefall-path': '/home/lishaojie/icefall', 'k2-path': '/home/lishaojie/.conda/envs/env_lishaojie/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/home/lishaojie/.conda/envs/env_lishaojie/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'cnc533', 'IP address': '127.0.1.1'}, 'world_size': 7, 'master_port': 18181, 'tensorboard': True, 'num_epochs': 30, 'start_epoch': 1, 'start_batch': 0, 'exp_dir': PosixPath('pruned_transducer_stateless7_streaming/exp1'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'base_lr': 0.004, 'lr_batches': 100000.0, 'lr_epochs': 100.0, 'context_size': 2, 'prune_range': 5, 'lm_scale': 0.25, 'am_scale': 0.0, 'simple_loss_scale': 0.5, 'seed': 42, 'print_diagnostics': False, 'inf_check': False, 'save_every_n': 2000, 'keep_last_k': 30, 'average_period': 200, 'use_fp16': True, 'num_encoder_layers': '2,4,3,2,4', 'feedforward_dims': '1024,1024,2048,2048,1024', 'nhead': '8,8,8,8,8', 'encoder_dims': '384,384,384,384,384', 'attention_dims': '192,192,192,192,192', 'encoder_unmasked_dims': '256,256,256,256,256', 'zipformer_downsampling_factors': '1,2,4,8,2', 'cnn_module_kernels': '31,31,31,31,31', 'decoder_dim': 512, 'joiner_dim': 512, 'do_finetune': True, 'init_modules': 'encoder', 'finetune_ckpt': '/home/lishaojie/icefall/egs/commonvoice/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/epoch-30.pt', 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 200, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'blank_id': 0, 'vocab_size': 500} +2023-03-25 21:42:34,506 INFO [finetune.py:1067] (0/7) About to create model +2023-03-25 21:42:34,857 INFO [zipformer.py:405] (0/7) At encoder stack 4, which has downsampling_factor=2, we will combine the outputs of layers 1 and 3, with downsampling_factors=2 and 8. +2023-03-25 21:42:34,866 INFO [finetune.py:1071] (0/7) Number of model parameters: 70369391 +2023-03-25 21:42:35,138 INFO [finetune.py:626] (0/7) Loading checkpoint from /home/lishaojie/icefall/egs/commonvoice/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/epoch-30.pt +2023-03-25 21:42:35,631 INFO [finetune.py:647] (0/7) Loading parameters starting with prefix encoder +2023-03-25 21:42:37,129 INFO [finetune.py:1093] (0/7) Using DDP +2023-03-25 21:42:37,687 INFO [commonvoice_fr.py:392] (0/7) About to get train cuts +2023-03-25 21:42:37,690 INFO [commonvoice_fr.py:218] (0/7) Enable MUSAN +2023-03-25 21:42:37,690 INFO [commonvoice_fr.py:219] (0/7) About to get Musan cuts +2023-03-25 21:42:39,374 INFO [commonvoice_fr.py:243] (0/7) Enable SpecAugment +2023-03-25 21:42:39,374 INFO [commonvoice_fr.py:244] (0/7) Time warp factor: 80 +2023-03-25 21:42:39,374 INFO [commonvoice_fr.py:254] (0/7) Num frame mask: 10 +2023-03-25 21:42:39,374 INFO [commonvoice_fr.py:267] (0/7) About to create train dataset +2023-03-25 21:42:39,374 INFO [commonvoice_fr.py:294] (0/7) Using DynamicBucketingSampler. +2023-03-25 21:42:41,952 INFO [commonvoice_fr.py:309] (0/7) About to create train dataloader +2023-03-25 21:42:41,953 INFO [commonvoice_fr.py:399] (0/7) About to get dev cuts +2023-03-25 21:42:41,954 INFO [commonvoice_fr.py:340] (0/7) About to create dev dataset +2023-03-25 21:42:42,363 INFO [commonvoice_fr.py:357] (0/7) About to create dev dataloader +2023-03-25 21:42:42,363 INFO [finetune.py:1289] (0/7) Sanity check -- see if any of the batches in epoch 1 would cause OOM. +2023-03-25 21:46:46,137 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5191MB +2023-03-25 21:46:46,827 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:46:48,912 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:46:49,576 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:46:50,267 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:46:50,962 INFO [finetune.py:1317] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:46:59,841 INFO [finetune.py:976] (0/7) Epoch 1, batch 0, loss[loss=7.482, simple_loss=6.785, pruned_loss=6.955, over 4828.00 frames. ], tot_loss[loss=7.482, simple_loss=6.785, pruned_loss=6.955, over 4828.00 frames. ], batch size: 25, lr: 2.00e-03, grad_scale: 2.0 +2023-03-25 21:46:59,843 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-25 21:47:14,952 INFO [finetune.py:1010] (0/7) Epoch 1, validation: loss=7.294, simple_loss=6.606, pruned_loss=6.863, over 2265189.00 frames. +2023-03-25 21:47:14,953 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 21:47:19,867 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:47:27,254 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=2.10 vs. limit=2.0 +2023-03-25 21:47:30,298 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:47:50,947 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=2.63 vs. limit=2.0 +2023-03-25 21:48:00,212 INFO [finetune.py:976] (0/7) Epoch 1, batch 50, loss[loss=3.199, simple_loss=2.988, pruned_loss=2.026, over 4893.00 frames. ], tot_loss[loss=4.335, simple_loss=3.894, pruned_loss=4.226, over 217796.55 frames. ], batch size: 32, lr: 2.20e-03, grad_scale: 0.000244140625 +2023-03-25 21:48:32,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:48:53,454 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.000244140625 +2023-03-25 21:48:53,454 INFO [finetune.py:976] (0/7) Epoch 1, batch 100, loss[loss=2.333, simple_loss=2.198, pruned_loss=1.313, over 4862.00 frames. ], tot_loss[loss=3.468, simple_loss=3.194, pruned_loss=2.669, over 380436.79 frames. ], batch size: 34, lr: 2.40e-03, grad_scale: 0.00048828125 +2023-03-25 21:49:02,131 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=3.11 vs. limit=2.0 +2023-03-25 21:49:13,205 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.539e+02 2.791e+03 6.484e+03 1.700e+04 1.722e+07, threshold=1.297e+04, percent-clipped=0.0 +2023-03-25 21:49:28,877 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 21:49:37,443 INFO [finetune.py:976] (0/7) Epoch 1, batch 150, loss[loss=1.733, simple_loss=1.571, pruned_loss=1.322, over 4771.00 frames. ], tot_loss[loss=2.853, simple_loss=2.635, pruned_loss=2.08, over 508368.86 frames. ], batch size: 28, lr: 2.60e-03, grad_scale: 0.00048828125 +2023-03-25 21:50:05,231 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-25 21:50:15,716 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.00048828125 +2023-03-25 21:50:15,717 INFO [finetune.py:976] (0/7) Epoch 1, batch 200, loss[loss=1.277, simple_loss=1.107, pruned_loss=1.188, over 4899.00 frames. ], tot_loss[loss=2.353, simple_loss=2.153, pruned_loss=1.787, over 608735.69 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 0.0009765625 +2023-03-25 21:50:17,850 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8069, 1.1097, 2.3376, 0.7416, 1.7454, 2.0803, 1.1350, 1.5817], + device='cuda:0'), covar=tensor([0.0734, 0.1145, 0.0537, 0.0972, 0.0770, 0.0900, 0.0802, 0.0681], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0081, 0.0071, 0.0073, 0.0090, 0.0076, 0.0084, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 21:50:29,335 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.018e+02 7.406e+02 1.293e+03 3.197e+03 6.754e+04, threshold=2.586e+03, percent-clipped=12.0 +2023-03-25 21:50:29,533 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=10.09 vs. limit=5.0 +2023-03-25 21:50:50,020 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3040, 1.8325, 1.6436, 2.1399, 2.3664, 2.3063, 2.1562, 1.5479], + device='cuda:0'), covar=tensor([0.0046, 0.0118, 0.0114, 0.0085, 0.0054, 0.0105, 0.0044, 0.0112], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0128, 0.0158, 0.0123, 0.0119, 0.0125, 0.0102, 0.0130], + device='cuda:0'), out_proj_covar=tensor([7.7769e-05, 1.0104e-04, 1.2878e-04, 9.7378e-05, 9.4777e-05, 9.4945e-05, + 7.9045e-05, 1.0199e-04], device='cuda:0') +2023-03-25 21:50:54,582 INFO [finetune.py:976] (0/7) Epoch 1, batch 250, loss[loss=1.115, simple_loss=0.9555, pruned_loss=1.031, over 4736.00 frames. ], tot_loss[loss=2.038, simple_loss=1.841, pruned_loss=1.62, over 686243.73 frames. ], batch size: 23, lr: 3.00e-03, grad_scale: 0.0009765625 +2023-03-25 21:50:55,803 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=8.96 vs. limit=5.0 +2023-03-25 21:50:56,280 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=13.40 vs. limit=5.0 +2023-03-25 21:51:26,477 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1555, 2.3914, 2.8886, 2.9205, 2.2856, 2.3395, 2.8314, 1.2035], + device='cuda:0'), covar=tensor([0.6082, 0.7831, 0.4233, 0.5515, 0.8895, 0.8594, 0.7315, 1.0588], + device='cuda:0'), in_proj_covar=tensor([0.0365, 0.0238, 0.0251, 0.0290, 0.0340, 0.0277, 0.0306, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 21:51:39,311 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-25 21:51:43,794 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=296.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:51:45,812 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=300.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 21:51:46,260 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.0009765625 +2023-03-25 21:51:46,260 INFO [finetune.py:976] (0/7) Epoch 1, batch 300, loss[loss=1.361, simple_loss=1.143, pruned_loss=1.291, over 4862.00 frames. ], tot_loss[loss=1.836, simple_loss=1.637, pruned_loss=1.515, over 747716.10 frames. ], batch size: 31, lr: 3.20e-03, grad_scale: 0.001953125 +2023-03-25 21:51:58,578 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.075e+01 5.781e+01 1.827e+02 5.788e+02 1.230e+04, threshold=3.655e+02, percent-clipped=4.0 +2023-03-25 21:52:39,148 INFO [finetune.py:976] (0/7) Epoch 1, batch 350, loss[loss=1.209, simple_loss=1.002, pruned_loss=1.141, over 4812.00 frames. ], tot_loss[loss=1.684, simple_loss=1.483, pruned_loss=1.429, over 791453.31 frames. ], batch size: 25, lr: 3.40e-03, grad_scale: 0.001953125 +2023-03-25 21:52:47,092 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=357.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:53:10,159 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=387.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:53:27,705 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.001953125 +2023-03-25 21:53:27,706 INFO [finetune.py:976] (0/7) Epoch 1, batch 400, loss[loss=1.09, simple_loss=0.8884, pruned_loss=1.037, over 4772.00 frames. ], tot_loss[loss=1.575, simple_loss=1.369, pruned_loss=1.366, over 828458.37 frames. ], batch size: 25, lr: 3.60e-03, grad_scale: 0.00390625 +2023-03-25 21:53:39,884 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.702e+01 2.277e+01 3.517e+01 1.113e+02 1.032e+03, threshold=7.035e+01, percent-clipped=3.0 +2023-03-25 21:53:51,258 WARNING [optim.py:389] (0/7) Scaling gradients by 0.06621765345335007, model_norm_threshold=70.34587860107422 +2023-03-25 21:53:51,346 INFO [optim.py:451] (0/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.67, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=7.539e+05, grad_sumsq = 2.933e+06, orig_rms_sq=2.571e-01 +2023-03-25 21:53:56,177 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.30 vs. limit=2.0 +2023-03-25 21:54:00,684 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=439.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 21:54:05,302 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=448.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:54:05,856 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=3.09 vs. limit=2.0 +2023-03-25 21:54:06,748 INFO [finetune.py:976] (0/7) Epoch 1, batch 450, loss[loss=1.137, simple_loss=0.9208, pruned_loss=1.058, over 4814.00 frames. ], tot_loss[loss=1.472, simple_loss=1.262, pruned_loss=1.297, over 856441.49 frames. ], batch size: 40, lr: 3.80e-03, grad_scale: 0.00390625 +2023-03-25 21:54:29,854 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2538, 1.5772, 2.4351, 2.7701, 2.0432, 2.3429, 1.0490, 2.0494], + device='cuda:0'), covar=tensor([0.1367, 0.1915, 0.1029, 0.2478, 0.1181, 0.2698, 0.1879, 0.1517], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0120, 0.0138, 0.0159, 0.0109, 0.0145, 0.0130, 0.0116], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-25 21:54:43,455 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.00390625 +2023-03-25 21:54:43,456 INFO [finetune.py:976] (0/7) Epoch 1, batch 500, loss[loss=1.165, simple_loss=0.9293, pruned_loss=1.082, over 4869.00 frames. ], tot_loss[loss=1.374, simple_loss=1.165, pruned_loss=1.223, over 876545.07 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.0078125 +2023-03-25 21:54:57,602 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.430e+01 1.676e+01 1.950e+01 4.114e+01 1.062e+03, threshold=3.899e+01, percent-clipped=11.0 +2023-03-25 21:55:28,674 INFO [finetune.py:976] (0/7) Epoch 1, batch 550, loss[loss=1.026, simple_loss=0.8164, pruned_loss=0.9261, over 4817.00 frames. ], tot_loss[loss=1.288, simple_loss=1.079, pruned_loss=1.154, over 896413.17 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 0.0078125 +2023-03-25 21:55:29,931 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=12.57 vs. limit=5.0 +2023-03-25 21:55:39,553 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=562.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:55:42,631 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=568.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:55:49,290 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-25 21:55:57,597 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=590.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:56:09,544 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=600.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:56:09,982 WARNING [finetune.py:966] (0/7) Grad scale is small: 0.0078125 +2023-03-25 21:56:09,982 INFO [finetune.py:976] (0/7) Epoch 1, batch 600, loss[loss=0.9982, simple_loss=0.7846, pruned_loss=0.8949, over 4768.00 frames. ], tot_loss[loss=1.229, simple_loss=1.017, pruned_loss=1.103, over 909590.78 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 0.015625 +2023-03-25 21:56:22,614 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.472e+01 1.758e+01 2.024e+01 2.271e+01 8.528e+01, threshold=4.048e+01, percent-clipped=5.0 +2023-03-25 21:56:27,296 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=623.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 21:56:36,086 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=629.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 21:56:54,357 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=648.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:56:55,842 INFO [finetune.py:976] (0/7) Epoch 1, batch 650, loss[loss=1.045, simple_loss=0.8161, pruned_loss=0.9198, over 4907.00 frames. ], tot_loss[loss=1.188, simple_loss=0.972, pruned_loss=1.065, over 919877.90 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.015625 +2023-03-25 21:56:55,927 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=651.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:56:56,433 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=652.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:56:57,513 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=41.40 vs. limit=5.0 +2023-03-25 21:57:02,306 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.23 vs. limit=2.0 +2023-03-25 21:57:31,101 INFO [finetune.py:976] (0/7) Epoch 1, batch 700, loss[loss=0.9225, simple_loss=0.7066, pruned_loss=0.8177, over 4796.00 frames. ], tot_loss[loss=1.151, simple_loss=0.9314, pruned_loss=1.028, over 927502.36 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 0.03125 +2023-03-25 21:57:33,811 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-25 21:57:38,289 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.810e+01 2.037e+01 2.232e+01 2.628e+01 5.516e+01, threshold=4.463e+01, percent-clipped=4.0 +2023-03-25 21:57:45,864 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-25 21:57:59,233 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=739.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:58:01,227 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=743.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 21:58:05,816 INFO [finetune.py:976] (0/7) Epoch 1, batch 750, loss[loss=1.045, simple_loss=0.8053, pruned_loss=0.8935, over 4783.00 frames. ], tot_loss[loss=1.12, simple_loss=0.8972, pruned_loss=0.9958, over 935368.95 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 0.03125 +2023-03-25 21:58:15,801 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3978, 1.8696, 0.9874, 1.9656, 2.2978, 1.1203, 1.3117, 1.8767], + device='cuda:0'), covar=tensor([0.0632, 0.1073, 0.1155, 0.0727, 0.1076, 0.0896, 0.1024, 0.0824], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0097, 0.0116, 0.0097, 0.0121, 0.0092, 0.0099, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 21:58:29,193 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=787.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:58:36,508 INFO [finetune.py:976] (0/7) Epoch 1, batch 800, loss[loss=1.006, simple_loss=0.7719, pruned_loss=0.844, over 4750.00 frames. ], tot_loss[loss=1.09, simple_loss=0.8646, pruned_loss=0.9624, over 939701.59 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 0.0625 +2023-03-25 21:58:45,189 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.059e+01 2.266e+01 2.508e+01 2.744e+01 4.199e+01, threshold=5.016e+01, percent-clipped=0.0 +2023-03-25 21:59:20,566 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=847.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:59:22,542 INFO [finetune.py:976] (0/7) Epoch 1, batch 850, loss[loss=0.8976, simple_loss=0.6715, pruned_loss=0.7639, over 4749.00 frames. ], tot_loss[loss=1.06, simple_loss=0.8327, pruned_loss=0.9281, over 943383.04 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 0.0625 +2023-03-25 21:59:43,768 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=2.09 vs. limit=2.0 +2023-03-25 22:00:12,170 INFO [finetune.py:976] (0/7) Epoch 1, batch 900, loss[loss=0.9119, simple_loss=0.6802, pruned_loss=0.7615, over 4771.00 frames. ], tot_loss[loss=1.03, simple_loss=0.8014, pruned_loss=0.8945, over 945719.11 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 0.125 +2023-03-25 22:00:15,803 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=908.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:00:25,569 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.101e+01 2.406e+01 2.575e+01 3.027e+01 5.726e+01, threshold=5.150e+01, percent-clipped=1.0 +2023-03-25 22:00:28,254 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=918.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:00:32,444 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=924.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 22:00:32,983 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8451, 2.0959, 1.7435, 2.4602, 2.2730, 2.1006, 1.6382, 1.4726], + device='cuda:0'), covar=tensor([0.0869, 0.0826, 0.1150, 0.0630, 0.0779, 0.0571, 0.1157, 0.0834], + device='cuda:0'), in_proj_covar=tensor([0.0257, 0.0239, 0.0229, 0.0209, 0.0272, 0.0204, 0.0235, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:00:53,627 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=946.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:00:56,170 INFO [finetune.py:976] (0/7) Epoch 1, batch 950, loss[loss=0.9857, simple_loss=0.7301, pruned_loss=0.813, over 4906.00 frames. ], tot_loss[loss=1.01, simple_loss=0.7787, pruned_loss=0.8686, over 948978.69 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.125 +2023-03-25 22:00:56,746 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=952.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:01:43,867 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1000.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:01:44,324 INFO [finetune.py:976] (0/7) Epoch 1, batch 1000, loss[loss=0.9956, simple_loss=0.7279, pruned_loss=0.8172, over 4894.00 frames. ], tot_loss[loss=1.006, simple_loss=0.7684, pruned_loss=0.8562, over 950248.18 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 0.25 +2023-03-25 22:01:58,296 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.382e+01 2.890e+01 3.153e+01 3.664e+01 7.462e+01, threshold=6.306e+01, percent-clipped=2.0 +2023-03-25 22:02:14,786 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=6.05 vs. limit=5.0 +2023-03-25 22:02:21,733 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1043.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:02:31,289 INFO [finetune.py:976] (0/7) Epoch 1, batch 1050, loss[loss=1.053, simple_loss=0.7738, pruned_loss=0.8419, over 4906.00 frames. ], tot_loss[loss=1.01, simple_loss=0.7641, pruned_loss=0.8499, over 950945.76 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.25 +2023-03-25 22:02:47,360 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.46 vs. limit=5.0 +2023-03-25 22:03:07,620 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1091.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:03:18,238 INFO [finetune.py:976] (0/7) Epoch 1, batch 1100, loss[loss=0.8809, simple_loss=0.6389, pruned_loss=0.7017, over 4805.00 frames. ], tot_loss[loss=1.007, simple_loss=0.756, pruned_loss=0.8386, over 953436.12 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 0.5 +2023-03-25 22:03:30,773 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 2.698e+01 3.337e+01 3.640e+01 4.251e+01 7.174e+01, threshold=7.279e+01, percent-clipped=4.0 +2023-03-25 22:03:36,654 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5737, 1.4792, 1.4649, 1.4981, 1.4306, 3.0529, 1.3649, 1.7434], + device='cuda:0'), covar=tensor([0.5885, 0.4437, 0.3193, 0.3805, 0.1879, 0.0795, 0.2714, 0.1274], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0121, 0.0128, 0.0125, 0.0111, 0.0099, 0.0095, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-25 22:04:04,833 INFO [finetune.py:976] (0/7) Epoch 1, batch 1150, loss[loss=0.8479, simple_loss=0.6106, pruned_loss=0.6684, over 4776.00 frames. ], tot_loss[loss=0.9967, simple_loss=0.7432, pruned_loss=0.8199, over 952515.19 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 0.5 +2023-03-25 22:04:05,999 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1153.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:04:46,444 INFO [finetune.py:976] (0/7) Epoch 1, batch 1200, loss[loss=0.8325, simple_loss=0.612, pruned_loss=0.6312, over 4750.00 frames. ], tot_loss[loss=0.9832, simple_loss=0.7301, pruned_loss=0.7975, over 953810.22 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:04:47,538 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1203.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:04:59,184 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 3.248e+01 4.460e+01 5.563e+01 6.854e+01 1.013e+02, threshold=1.113e+02, percent-clipped=20.0 +2023-03-25 22:04:59,290 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1214.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:05:01,634 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1218.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:05:05,627 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-25 22:05:06,577 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1224.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:05:08,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2997, 1.2966, 1.2353, 1.6833, 1.5497, 1.3170, 2.1041, 1.3692], + device='cuda:0'), covar=tensor([0.3975, 0.5367, 0.5077, 0.3423, 0.4118, 0.3023, 0.1615, 0.6314], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0225, 0.0264, 0.0296, 0.0248, 0.0201, 0.0211, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:05:21,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4502, 1.2279, 1.4712, 0.9999, 1.4979, 1.5178, 1.4245, 1.1222], + device='cuda:0'), covar=tensor([0.0569, 0.0971, 0.0757, 0.0821, 0.0723, 0.0604, 0.0657, 0.1601], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0133, 0.0130, 0.0119, 0.0107, 0.0129, 0.0135, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:05:24,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1246.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:05:26,689 INFO [finetune.py:976] (0/7) Epoch 1, batch 1250, loss[loss=0.9172, simple_loss=0.6759, pruned_loss=0.6829, over 4809.00 frames. ], tot_loss[loss=0.9618, simple_loss=0.713, pruned_loss=0.7683, over 954494.61 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:05:46,029 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1266.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:05:49,131 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1272.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:05:54,021 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.60 vs. limit=5.0 +2023-03-25 22:06:08,473 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1294.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:06:14,925 INFO [finetune.py:976] (0/7) Epoch 1, batch 1300, loss[loss=0.8893, simple_loss=0.6662, pruned_loss=0.6416, over 4867.00 frames. ], tot_loss[loss=0.9354, simple_loss=0.6936, pruned_loss=0.7348, over 955039.50 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:06:23,503 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 5.599e+01 8.403e+01 9.999e+01 1.262e+02 2.600e+02, threshold=2.000e+02, percent-clipped=40.0 +2023-03-25 22:06:57,097 INFO [finetune.py:976] (0/7) Epoch 1, batch 1350, loss[loss=0.7415, simple_loss=0.5583, pruned_loss=0.5251, over 4050.00 frames. ], tot_loss[loss=0.915, simple_loss=0.6801, pruned_loss=0.7056, over 955247.62 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:07:21,302 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1616, 1.8336, 1.2170, 2.1632, 1.7676, 1.5107, 1.6151, 2.3974], + device='cuda:0'), covar=tensor([1.1892, 1.0459, 1.1032, 1.0135, 1.0358, 1.0027, 1.5854, 0.3656], + device='cuda:0'), in_proj_covar=tensor([0.0299, 0.0297, 0.0256, 0.0341, 0.0283, 0.0240, 0.0299, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:07:47,967 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3801, 1.4410, 1.3877, 1.3619, 1.1636, 3.8117, 1.6006, 1.9466], + device='cuda:0'), covar=tensor([0.7587, 0.5011, 0.3774, 0.3788, 0.2578, 0.0363, 0.2870, 0.1604], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0124, 0.0121, 0.0106, 0.0096, 0.0090, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 22:07:50,278 INFO [finetune.py:976] (0/7) Epoch 1, batch 1400, loss[loss=0.8332, simple_loss=0.6446, pruned_loss=0.5672, over 4898.00 frames. ], tot_loss[loss=0.9005, simple_loss=0.6724, pruned_loss=0.6806, over 954903.42 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:07:58,270 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.277e+01 1.400e+02 1.610e+02 1.980e+02 2.974e+02, threshold=3.221e+02, percent-clipped=23.0 +2023-03-25 22:08:09,171 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1434.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:08:16,156 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-25 22:08:20,140 INFO [finetune.py:976] (0/7) Epoch 1, batch 1450, loss[loss=0.6772, simple_loss=0.5177, pruned_loss=0.4612, over 4188.00 frames. ], tot_loss[loss=0.8793, simple_loss=0.6607, pruned_loss=0.651, over 954915.56 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:08:36,825 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3506, 1.4509, 1.1723, 1.6142, 1.5825, 2.7727, 1.2853, 1.4021], + device='cuda:0'), covar=tensor([0.1099, 0.1921, 0.1302, 0.1209, 0.1816, 0.0323, 0.1804, 0.2118], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0080, 0.0071, 0.0073, 0.0090, 0.0076, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 22:08:47,666 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1495.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 22:08:51,790 INFO [finetune.py:976] (0/7) Epoch 1, batch 1500, loss[loss=0.6183, simple_loss=0.4868, pruned_loss=0.4051, over 4286.00 frames. ], tot_loss[loss=0.8531, simple_loss=0.6461, pruned_loss=0.6187, over 954076.62 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:08:52,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1503.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:09:02,818 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1509.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:09:05,944 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.968e+01 1.844e+02 2.293e+02 2.711e+02 4.587e+02, threshold=4.586e+02, percent-clipped=13.0 +2023-03-25 22:09:42,480 INFO [finetune.py:976] (0/7) Epoch 1, batch 1550, loss[loss=0.6468, simple_loss=0.5156, pruned_loss=0.415, over 4842.00 frames. ], tot_loss[loss=0.8213, simple_loss=0.6281, pruned_loss=0.5829, over 956409.60 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:09:42,537 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1551.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:09:52,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1566.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:10:21,704 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6142, 3.9750, 4.0220, 1.7861, 4.2134, 3.1153, 0.8794, 2.8743], + device='cuda:0'), covar=tensor([0.2349, 0.0972, 0.1149, 0.2996, 0.0645, 0.0777, 0.3882, 0.1185], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0147, 0.0156, 0.0123, 0.0146, 0.0111, 0.0136, 0.0114], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:10:22,840 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5815, 1.4220, 2.2750, 3.0617, 2.1832, 2.3222, 0.9020, 2.3623], + device='cuda:0'), covar=tensor([0.2506, 0.2171, 0.1427, 0.0852, 0.1246, 0.1705, 0.2441, 0.1211], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0115, 0.0128, 0.0150, 0.0104, 0.0136, 0.0122, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:10:33,777 INFO [finetune.py:976] (0/7) Epoch 1, batch 1600, loss[loss=0.6515, simple_loss=0.5277, pruned_loss=0.4084, over 4912.00 frames. ], tot_loss[loss=0.7853, simple_loss=0.6065, pruned_loss=0.5461, over 956753.67 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:10:40,881 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1611.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:10:42,942 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.367e+02 1.965e+02 2.441e+02 2.819e+02 5.041e+02, threshold=4.882e+02, percent-clipped=1.0 +2023-03-25 22:10:55,968 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1627.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:11:18,727 INFO [finetune.py:976] (0/7) Epoch 1, batch 1650, loss[loss=0.6149, simple_loss=0.5022, pruned_loss=0.3796, over 4893.00 frames. ], tot_loss[loss=0.7498, simple_loss=0.5847, pruned_loss=0.5113, over 956223.41 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:11:26,252 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-25 22:11:41,991 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1672.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:12:02,514 INFO [finetune.py:976] (0/7) Epoch 1, batch 1700, loss[loss=0.5251, simple_loss=0.4452, pruned_loss=0.3111, over 4822.00 frames. ], tot_loss[loss=0.7163, simple_loss=0.5648, pruned_loss=0.4787, over 955547.72 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:12:14,553 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 2.187e+02 2.736e+02 3.197e+02 8.210e+02, threshold=5.471e+02, percent-clipped=2.0 +2023-03-25 22:12:53,678 INFO [finetune.py:976] (0/7) Epoch 1, batch 1750, loss[loss=0.6399, simple_loss=0.5435, pruned_loss=0.3765, over 4811.00 frames. ], tot_loss[loss=0.6986, simple_loss=0.5572, pruned_loss=0.4571, over 956827.92 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:13:29,965 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1790.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:13:36,051 INFO [finetune.py:976] (0/7) Epoch 1, batch 1800, loss[loss=0.5412, simple_loss=0.4712, pruned_loss=0.3099, over 4884.00 frames. ], tot_loss[loss=0.6851, simple_loss=0.5525, pruned_loss=0.4396, over 957853.63 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:13:40,469 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1809.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:13:43,023 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 2.215e+02 2.629e+02 3.291e+02 5.990e+02, threshold=5.258e+02, percent-clipped=1.0 +2023-03-25 22:13:59,211 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1838.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:06,691 INFO [finetune.py:976] (0/7) Epoch 1, batch 1850, loss[loss=0.4293, simple_loss=0.3788, pruned_loss=0.242, over 4760.00 frames. ], tot_loss[loss=0.6671, simple_loss=0.5437, pruned_loss=0.4203, over 957772.35 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:14:10,079 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1857.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:14:10,121 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1857.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:17,819 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1870.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:51,509 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1899.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:14:52,519 INFO [finetune.py:976] (0/7) Epoch 1, batch 1900, loss[loss=0.5716, simple_loss=0.5016, pruned_loss=0.3227, over 4846.00 frames. ], tot_loss[loss=0.6498, simple_loss=0.5351, pruned_loss=0.4025, over 957980.87 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:15:03,949 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.381e+02 2.208e+02 2.560e+02 3.227e+02 6.450e+02, threshold=5.121e+02, percent-clipped=1.0 +2023-03-25 22:15:11,547 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1918.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:15:14,230 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1922.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:15:20,201 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1931.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:15:34,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5759, 1.4541, 2.3648, 3.3047, 2.1920, 2.3964, 1.0579, 2.4285], + device='cuda:0'), covar=tensor([0.1832, 0.1510, 0.1060, 0.0414, 0.0853, 0.1239, 0.1911, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0115, 0.0128, 0.0150, 0.0103, 0.0137, 0.0120, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:15:37,069 INFO [finetune.py:976] (0/7) Epoch 1, batch 1950, loss[loss=0.5522, simple_loss=0.4809, pruned_loss=0.3127, over 4905.00 frames. ], tot_loss[loss=0.6283, simple_loss=0.5224, pruned_loss=0.3832, over 958651.39 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:15:43,342 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=7.94 vs. limit=5.0 +2023-03-25 22:15:46,009 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1967.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:16:07,081 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=12.42 vs. limit=5.0 +2023-03-25 22:16:11,135 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-2000.pt +2023-03-25 22:16:12,900 INFO [finetune.py:976] (0/7) Epoch 1, batch 2000, loss[loss=0.4579, simple_loss=0.4189, pruned_loss=0.2484, over 4864.00 frames. ], tot_loss[loss=0.606, simple_loss=0.5088, pruned_loss=0.3643, over 958895.07 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 4.0 +2023-03-25 22:16:22,847 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.403e+02 2.183e+02 2.758e+02 3.285e+02 7.843e+02, threshold=5.515e+02, percent-clipped=1.0 +2023-03-25 22:16:57,264 INFO [finetune.py:976] (0/7) Epoch 1, batch 2050, loss[loss=0.4846, simple_loss=0.4408, pruned_loss=0.2641, over 4826.00 frames. ], tot_loss[loss=0.5801, simple_loss=0.493, pruned_loss=0.3435, over 956592.39 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:17:31,370 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2090.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:17:41,737 INFO [finetune.py:976] (0/7) Epoch 1, batch 2100, loss[loss=0.611, simple_loss=0.5095, pruned_loss=0.3562, over 4115.00 frames. ], tot_loss[loss=0.5633, simple_loss=0.4846, pruned_loss=0.3287, over 956254.92 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:17:50,750 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7372, 1.6458, 1.0851, 1.7458, 1.8438, 1.4752, 1.3558, 1.9262], + device='cuda:0'), covar=tensor([0.2153, 0.2090, 0.2195, 0.1366, 0.3113, 0.1942, 0.1692, 0.1869], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0094, 0.0109, 0.0091, 0.0121, 0.0088, 0.0095, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:17:55,063 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.336e+02 2.022e+02 2.484e+02 2.961e+02 6.695e+02, threshold=4.968e+02, percent-clipped=1.0 +2023-03-25 22:18:04,908 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-25 22:18:12,041 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2138.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:18:29,607 INFO [finetune.py:976] (0/7) Epoch 1, batch 2150, loss[loss=0.546, simple_loss=0.5031, pruned_loss=0.2945, over 4912.00 frames. ], tot_loss[loss=0.5573, simple_loss=0.4852, pruned_loss=0.3207, over 955858.96 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:19:03,494 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2194.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:08,497 INFO [finetune.py:976] (0/7) Epoch 1, batch 2200, loss[loss=0.481, simple_loss=0.4521, pruned_loss=0.2549, over 4925.00 frames. ], tot_loss[loss=0.5471, simple_loss=0.482, pruned_loss=0.3107, over 955984.97 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:19:17,021 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2213.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:17,476 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.568e+02 2.355e+02 2.819e+02 3.325e+02 5.172e+02, threshold=5.637e+02, percent-clipped=1.0 +2023-03-25 22:19:22,617 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2222.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:19:28,118 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2226.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:57,062 INFO [finetune.py:976] (0/7) Epoch 1, batch 2250, loss[loss=0.5474, simple_loss=0.5013, pruned_loss=0.2967, over 4799.00 frames. ], tot_loss[loss=0.5358, simple_loss=0.4767, pruned_loss=0.3011, over 955068.40 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:20:18,393 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2267.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:20:20,119 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2270.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:21:00,801 INFO [finetune.py:976] (0/7) Epoch 1, batch 2300, loss[loss=0.4312, simple_loss=0.4156, pruned_loss=0.2234, over 4838.00 frames. ], tot_loss[loss=0.5235, simple_loss=0.4703, pruned_loss=0.2912, over 954962.86 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:21:15,882 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.417e+02 2.050e+02 2.425e+02 2.921e+02 4.362e+02, threshold=4.850e+02, percent-clipped=0.0 +2023-03-25 22:21:21,996 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2315.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:21:37,369 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2340.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:21:54,803 INFO [finetune.py:976] (0/7) Epoch 1, batch 2350, loss[loss=0.413, simple_loss=0.3934, pruned_loss=0.2163, over 4834.00 frames. ], tot_loss[loss=0.506, simple_loss=0.4593, pruned_loss=0.2786, over 955495.06 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:22:57,440 INFO [finetune.py:976] (0/7) Epoch 1, batch 2400, loss[loss=0.3924, simple_loss=0.3794, pruned_loss=0.2026, over 4691.00 frames. ], tot_loss[loss=0.4941, simple_loss=0.4517, pruned_loss=0.27, over 953942.32 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:22:57,569 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2401.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:23:05,867 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.408e+02 1.953e+02 2.427e+02 2.971e+02 6.309e+02, threshold=4.853e+02, percent-clipped=1.0 +2023-03-25 22:23:32,012 INFO [finetune.py:976] (0/7) Epoch 1, batch 2450, loss[loss=0.5056, simple_loss=0.4533, pruned_loss=0.279, over 4846.00 frames. ], tot_loss[loss=0.4827, simple_loss=0.4443, pruned_loss=0.2619, over 954491.44 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:24:21,678 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2494.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:24:25,638 INFO [finetune.py:976] (0/7) Epoch 1, batch 2500, loss[loss=0.4545, simple_loss=0.4352, pruned_loss=0.2369, over 4820.00 frames. ], tot_loss[loss=0.48, simple_loss=0.4444, pruned_loss=0.2588, over 956330.74 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:24:34,877 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2513.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:24:35,373 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.353e+02 2.241e+02 2.593e+02 3.079e+02 4.323e+02, threshold=5.185e+02, percent-clipped=0.0 +2023-03-25 22:24:38,301 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8140, 1.9135, 1.5562, 1.3577, 2.4214, 2.2118, 1.9215, 1.7087], + device='cuda:0'), covar=tensor([0.0787, 0.0679, 0.0890, 0.1033, 0.0396, 0.0688, 0.0744, 0.1277], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0134, 0.0133, 0.0121, 0.0109, 0.0131, 0.0137, 0.0169], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:24:44,452 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2526.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:24:54,544 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2542.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:25:00,193 INFO [finetune.py:976] (0/7) Epoch 1, batch 2550, loss[loss=0.4933, simple_loss=0.4789, pruned_loss=0.2538, over 4924.00 frames. ], tot_loss[loss=0.4769, simple_loss=0.445, pruned_loss=0.2552, over 956091.01 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:25:09,687 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2561.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:25:18,669 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2574.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:25:48,339 INFO [finetune.py:976] (0/7) Epoch 1, batch 2600, loss[loss=0.3401, simple_loss=0.3373, pruned_loss=0.1714, over 4103.00 frames. ], tot_loss[loss=0.4737, simple_loss=0.4446, pruned_loss=0.252, over 953759.10 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:25:55,826 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.651e+02 2.205e+02 2.587e+02 2.996e+02 4.228e+02, threshold=5.174e+02, percent-clipped=0.0 +2023-03-25 22:25:55,966 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3731, 1.0058, 0.9894, 0.7750, 1.3174, 1.5936, 1.3652, 0.9225], + device='cuda:0'), covar=tensor([0.0309, 0.0450, 0.0741, 0.0529, 0.0310, 0.0252, 0.0365, 0.0511], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0119, 0.0142, 0.0116, 0.0108, 0.0108, 0.0092, 0.0118], + device='cuda:0'), out_proj_covar=tensor([7.0104e-05, 9.3946e-05, 1.1525e-04, 9.2429e-05, 8.6135e-05, 8.1581e-05, + 7.1539e-05, 9.2942e-05], device='cuda:0') +2023-03-25 22:26:15,513 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.89 vs. limit=5.0 +2023-03-25 22:26:19,959 INFO [finetune.py:976] (0/7) Epoch 1, batch 2650, loss[loss=0.48, simple_loss=0.4615, pruned_loss=0.2493, over 4777.00 frames. ], tot_loss[loss=0.4689, simple_loss=0.4423, pruned_loss=0.2482, over 952485.89 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:26:54,439 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2685.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:27:06,859 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2696.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:27:15,135 INFO [finetune.py:976] (0/7) Epoch 1, batch 2700, loss[loss=0.4352, simple_loss=0.4126, pruned_loss=0.2289, over 4766.00 frames. ], tot_loss[loss=0.4567, simple_loss=0.4348, pruned_loss=0.2397, over 951219.07 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:27:26,642 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8484, 1.4419, 1.0277, 1.8157, 2.0979, 1.4260, 1.4399, 1.9028], + device='cuda:0'), covar=tensor([0.1815, 0.2063, 0.2199, 0.1258, 0.2266, 0.1972, 0.1430, 0.1911], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0092, 0.0109, 0.0088, 0.0119, 0.0088, 0.0093, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:27:28,252 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.346e+02 2.127e+02 2.493e+02 3.058e+02 5.200e+02, threshold=4.985e+02, percent-clipped=1.0 +2023-03-25 22:28:14,525 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2746.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:28:17,278 INFO [finetune.py:976] (0/7) Epoch 1, batch 2750, loss[loss=0.4299, simple_loss=0.4151, pruned_loss=0.2224, over 4835.00 frames. ], tot_loss[loss=0.4478, simple_loss=0.4284, pruned_loss=0.2339, over 952759.31 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:28:58,715 INFO [finetune.py:976] (0/7) Epoch 1, batch 2800, loss[loss=0.3498, simple_loss=0.3557, pruned_loss=0.172, over 4809.00 frames. ], tot_loss[loss=0.4375, simple_loss=0.4207, pruned_loss=0.2274, over 953987.30 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:29:06,644 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.489e+02 2.264e+02 2.537e+02 3.001e+02 5.007e+02, threshold=5.073e+02, percent-clipped=1.0 +2023-03-25 22:29:12,561 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2824.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:29:23,186 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3615, 1.1641, 1.2374, 1.3126, 1.8440, 1.2487, 0.9135, 1.0270], + device='cuda:0'), covar=tensor([0.4110, 0.4540, 0.3677, 0.3806, 0.3436, 0.2780, 0.5788, 0.3559], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0202, 0.0190, 0.0176, 0.0224, 0.0175, 0.0199, 0.0176], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:29:40,284 INFO [finetune.py:976] (0/7) Epoch 1, batch 2850, loss[loss=0.4244, simple_loss=0.4168, pruned_loss=0.216, over 4835.00 frames. ], tot_loss[loss=0.4312, simple_loss=0.4165, pruned_loss=0.2231, over 953880.70 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:29:42,777 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-25 22:30:03,455 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2885.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:30:15,602 INFO [finetune.py:976] (0/7) Epoch 1, batch 2900, loss[loss=0.3393, simple_loss=0.3457, pruned_loss=0.1665, over 3859.00 frames. ], tot_loss[loss=0.4304, simple_loss=0.4178, pruned_loss=0.2216, over 954373.50 frames. ], batch size: 16, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:30:18,071 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0869, 1.8586, 1.7594, 0.7548, 1.9001, 2.3085, 1.6565, 1.7720], + device='cuda:0'), covar=tensor([0.0919, 0.0403, 0.0376, 0.0723, 0.0415, 0.0193, 0.0400, 0.0468], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0149, 0.0116, 0.0126, 0.0126, 0.0119, 0.0144, 0.0152], + device='cuda:0'), out_proj_covar=tensor([9.5204e-05, 1.1133e-04, 8.5189e-05, 9.2618e-05, 9.1421e-05, 8.8400e-05, + 1.0767e-04, 1.1304e-04], device='cuda:0') +2023-03-25 22:30:23,164 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.426e+02 2.100e+02 2.461e+02 2.914e+02 6.574e+02, threshold=4.923e+02, percent-clipped=3.0 +2023-03-25 22:30:37,167 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2937.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:30:50,912 INFO [finetune.py:976] (0/7) Epoch 1, batch 2950, loss[loss=0.4048, simple_loss=0.4104, pruned_loss=0.1996, over 4898.00 frames. ], tot_loss[loss=0.4283, simple_loss=0.4187, pruned_loss=0.2191, over 953193.40 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:31:31,734 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2996.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:31:31,780 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3653, 1.5659, 1.1542, 1.6194, 1.5888, 1.2645, 2.2474, 1.4449], + device='cuda:0'), covar=tensor([0.3391, 0.5074, 0.5583, 0.5464, 0.3682, 0.2941, 0.2800, 0.4872], + device='cuda:0'), in_proj_covar=tensor([0.0160, 0.0188, 0.0225, 0.0243, 0.0202, 0.0172, 0.0178, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-25 22:31:32,908 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2998.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:31:35,184 INFO [finetune.py:976] (0/7) Epoch 1, batch 3000, loss[loss=0.4595, simple_loss=0.4521, pruned_loss=0.2335, over 4855.00 frames. ], tot_loss[loss=0.4249, simple_loss=0.4166, pruned_loss=0.2167, over 952093.30 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:31:35,186 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-25 22:31:56,386 INFO [finetune.py:1010] (0/7) Epoch 1, validation: loss=0.4228, simple_loss=0.4589, pruned_loss=0.1933, over 2265189.00 frames. +2023-03-25 22:31:56,386 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 5672MB +2023-03-25 22:32:17,068 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.364e+02 2.092e+02 2.490e+02 2.940e+02 5.162e+02, threshold=4.980e+02, percent-clipped=2.0 +2023-03-25 22:32:25,823 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3019.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:32:39,246 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3041.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:32:40,979 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3044.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:32:44,999 INFO [finetune.py:976] (0/7) Epoch 1, batch 3050, loss[loss=0.4654, simple_loss=0.4415, pruned_loss=0.2446, over 4166.00 frames. ], tot_loss[loss=0.4237, simple_loss=0.4173, pruned_loss=0.2151, over 952309.83 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:33:09,757 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3080.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:33:38,269 INFO [finetune.py:976] (0/7) Epoch 1, batch 3100, loss[loss=0.4881, simple_loss=0.453, pruned_loss=0.2617, over 4899.00 frames. ], tot_loss[loss=0.4185, simple_loss=0.4134, pruned_loss=0.2118, over 951866.94 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:33:51,986 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.371e+02 2.009e+02 2.458e+02 3.052e+02 4.298e+02, threshold=4.916e+02, percent-clipped=0.0 +2023-03-25 22:34:39,799 INFO [finetune.py:976] (0/7) Epoch 1, batch 3150, loss[loss=0.3714, simple_loss=0.3868, pruned_loss=0.1781, over 4785.00 frames. ], tot_loss[loss=0.4094, simple_loss=0.4064, pruned_loss=0.2062, over 953847.85 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:34:57,756 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-25 22:35:10,570 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0418, 1.5004, 2.1466, 1.5239, 1.9776, 1.9390, 1.5615, 2.2082], + device='cuda:0'), covar=tensor([0.1814, 0.2118, 0.1535, 0.2186, 0.1078, 0.1796, 0.2726, 0.1220], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0194, 0.0191, 0.0179, 0.0161, 0.0201, 0.0201, 0.0178], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:35:11,099 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3180.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:35:39,621 INFO [finetune.py:976] (0/7) Epoch 1, batch 3200, loss[loss=0.4429, simple_loss=0.4365, pruned_loss=0.2247, over 4841.00 frames. ], tot_loss[loss=0.4001, simple_loss=0.3992, pruned_loss=0.2005, over 953034.92 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:35:40,959 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-25 22:35:52,730 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.973e+02 2.320e+02 2.787e+02 5.091e+02, threshold=4.641e+02, percent-clipped=1.0 +2023-03-25 22:36:29,140 INFO [finetune.py:976] (0/7) Epoch 1, batch 3250, loss[loss=0.4297, simple_loss=0.4304, pruned_loss=0.2145, over 4839.00 frames. ], tot_loss[loss=0.3982, simple_loss=0.3981, pruned_loss=0.1992, over 953298.01 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:36:29,769 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4747, 1.6922, 1.1213, 2.0128, 1.6912, 1.2632, 2.6828, 1.5078], + device='cuda:0'), covar=tensor([0.3206, 0.5232, 0.4704, 0.4751, 0.3102, 0.2664, 0.2081, 0.4284], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0187, 0.0224, 0.0241, 0.0201, 0.0171, 0.0178, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-25 22:36:39,341 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2793, 1.8411, 1.9596, 0.8097, 1.9833, 1.8474, 1.5054, 2.0348], + device='cuda:0'), covar=tensor([0.0591, 0.1080, 0.1261, 0.2406, 0.0939, 0.1725, 0.2131, 0.0910], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0162, 0.0177, 0.0162, 0.0179, 0.0178, 0.0186, 0.0173], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:37:12,255 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3288.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:37:19,995 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3293.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:37:24,559 INFO [finetune.py:976] (0/7) Epoch 1, batch 3300, loss[loss=0.3621, simple_loss=0.3616, pruned_loss=0.1814, over 4372.00 frames. ], tot_loss[loss=0.4034, simple_loss=0.4033, pruned_loss=0.2017, over 954069.75 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:37:32,704 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.553e+02 2.210e+02 2.512e+02 3.057e+02 4.555e+02, threshold=5.024e+02, percent-clipped=0.0 +2023-03-25 22:37:43,408 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-25 22:38:03,254 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3341.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:38:13,340 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3349.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 22:38:14,398 INFO [finetune.py:976] (0/7) Epoch 1, batch 3350, loss[loss=0.4454, simple_loss=0.4283, pruned_loss=0.2312, over 4801.00 frames. ], tot_loss[loss=0.4032, simple_loss=0.4045, pruned_loss=0.2009, over 953402.24 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:38:36,510 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-25 22:38:44,432 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3375.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:38:46,697 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3926, 1.6373, 1.4297, 1.6575, 0.8545, 3.4584, 1.1210, 1.7132], + device='cuda:0'), covar=tensor([0.3628, 0.2307, 0.2304, 0.2285, 0.2376, 0.0176, 0.3027, 0.1579], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0101, 0.0109, 0.0107, 0.0099, 0.0086, 0.0085, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 22:38:59,537 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3389.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:39:06,750 INFO [finetune.py:976] (0/7) Epoch 1, batch 3400, loss[loss=0.3697, simple_loss=0.3854, pruned_loss=0.177, over 4843.00 frames. ], tot_loss[loss=0.3999, simple_loss=0.4028, pruned_loss=0.1985, over 953507.16 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:39:14,429 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-25 22:39:20,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.307e+02 1.988e+02 2.392e+02 2.720e+02 4.202e+02, threshold=4.784e+02, percent-clipped=0.0 +2023-03-25 22:39:47,477 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-25 22:39:50,839 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-25 22:40:08,846 INFO [finetune.py:976] (0/7) Epoch 1, batch 3450, loss[loss=0.4317, simple_loss=0.4307, pruned_loss=0.2164, over 4830.00 frames. ], tot_loss[loss=0.3983, simple_loss=0.4025, pruned_loss=0.197, over 952736.06 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:40:11,261 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6029, 1.3538, 1.8566, 3.1398, 2.2535, 2.1842, 0.8748, 2.3829], + device='cuda:0'), covar=tensor([0.1896, 0.1755, 0.1420, 0.0492, 0.0885, 0.1439, 0.2056, 0.0758], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0114, 0.0131, 0.0152, 0.0101, 0.0139, 0.0122, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:40:40,364 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3480.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:40:44,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2026, 1.0630, 0.7989, 0.9961, 1.0487, 0.9434, 0.9585, 1.5339], + device='cuda:0'), covar=tensor([11.8276, 12.1489, 10.5734, 16.5622, 9.3896, 6.9067, 13.0256, 3.7701], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0226, 0.0206, 0.0264, 0.0220, 0.0190, 0.0226, 0.0169], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 22:41:03,281 INFO [finetune.py:976] (0/7) Epoch 1, batch 3500, loss[loss=0.3808, simple_loss=0.3807, pruned_loss=0.1904, over 4864.00 frames. ], tot_loss[loss=0.3916, simple_loss=0.3971, pruned_loss=0.193, over 954364.11 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:41:12,686 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9082, 1.5606, 2.1942, 1.5047, 1.7761, 1.9700, 1.6276, 2.2326], + device='cuda:0'), covar=tensor([0.1962, 0.2214, 0.1686, 0.2187, 0.1311, 0.1907, 0.2368, 0.1348], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0197, 0.0192, 0.0181, 0.0164, 0.0205, 0.0202, 0.0180], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:41:14,916 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 2.271e+02 2.832e+02 3.824e+02 1.123e+03, threshold=5.664e+02, percent-clipped=12.0 +2023-03-25 22:41:30,605 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3528.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:41:48,771 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3546.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:41:57,309 INFO [finetune.py:976] (0/7) Epoch 1, batch 3550, loss[loss=0.3551, simple_loss=0.3671, pruned_loss=0.1715, over 4823.00 frames. ], tot_loss[loss=0.3875, simple_loss=0.3929, pruned_loss=0.1911, over 955456.42 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:42:39,660 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3593.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:42:50,744 INFO [finetune.py:976] (0/7) Epoch 1, batch 3600, loss[loss=0.396, simple_loss=0.3974, pruned_loss=0.1973, over 4930.00 frames. ], tot_loss[loss=0.3838, simple_loss=0.3896, pruned_loss=0.189, over 954260.02 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:42:59,651 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3607.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:43:01,321 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9175, 4.2336, 4.1877, 2.2898, 4.4212, 3.3583, 0.8323, 3.0983], + device='cuda:0'), covar=tensor([0.2432, 0.1523, 0.1254, 0.3019, 0.0744, 0.0779, 0.4423, 0.1301], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0153, 0.0159, 0.0124, 0.0149, 0.0114, 0.0140, 0.0116], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:43:09,362 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.842e+02 2.557e+02 2.866e+02 3.769e+02 9.044e+02, threshold=5.732e+02, percent-clipped=5.0 +2023-03-25 22:43:12,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1687, 1.1915, 1.3521, 1.0041, 1.0340, 1.2709, 1.2181, 1.4986], + device='cuda:0'), covar=tensor([0.1888, 0.2457, 0.1653, 0.1762, 0.1207, 0.1563, 0.2872, 0.1014], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0197, 0.0193, 0.0182, 0.0164, 0.0206, 0.0203, 0.0181], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:43:44,098 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3641.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:43:44,177 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3665, 0.8632, 1.2058, 0.9689, 1.0106, 1.0049, 1.0495, 1.0647], + device='cuda:0'), covar=tensor([ 5.8874, 11.9803, 6.4421, 8.9243, 10.0162, 5.7976, 12.5393, 6.8513], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0248, 0.0234, 0.0266, 0.0250, 0.0215, 0.0279, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 22:43:46,913 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3644.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:43:52,544 INFO [finetune.py:976] (0/7) Epoch 1, batch 3650, loss[loss=0.319, simple_loss=0.3551, pruned_loss=0.1415, over 4753.00 frames. ], tot_loss[loss=0.3828, simple_loss=0.3899, pruned_loss=0.1879, over 954841.35 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:44:15,661 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-25 22:44:20,298 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3675.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:44:21,473 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3676.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:44:31,151 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.8437, 0.6176, 0.7832, 0.6440, 0.4832, 0.5442, 0.6487, 0.7042], + device='cuda:0'), covar=tensor([ 6.7902, 11.1865, 6.1147, 10.1352, 10.9748, 6.8550, 13.1974, 5.7928], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0245, 0.0231, 0.0262, 0.0247, 0.0212, 0.0275, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 22:44:40,051 INFO [finetune.py:976] (0/7) Epoch 1, batch 3700, loss[loss=0.4184, simple_loss=0.4283, pruned_loss=0.2043, over 4813.00 frames. ], tot_loss[loss=0.383, simple_loss=0.3919, pruned_loss=0.187, over 954153.37 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:44:52,809 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.548e+02 2.567e+02 2.980e+02 3.536e+02 5.905e+02, threshold=5.959e+02, percent-clipped=1.0 +2023-03-25 22:45:09,735 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3723.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:45:23,612 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3737.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:45:35,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8498, 1.5479, 2.1661, 1.5133, 1.8201, 1.9290, 1.5912, 2.2283], + device='cuda:0'), covar=tensor([0.1881, 0.2193, 0.1505, 0.2192, 0.1111, 0.1892, 0.2394, 0.1181], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0197, 0.0193, 0.0183, 0.0165, 0.0207, 0.0204, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:45:43,735 INFO [finetune.py:976] (0/7) Epoch 1, batch 3750, loss[loss=0.4365, simple_loss=0.4281, pruned_loss=0.2225, over 4917.00 frames. ], tot_loss[loss=0.3829, simple_loss=0.3933, pruned_loss=0.1862, over 955741.16 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:45:52,449 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.88 vs. limit=5.0 +2023-03-25 22:46:33,632 INFO [finetune.py:976] (0/7) Epoch 1, batch 3800, loss[loss=0.4411, simple_loss=0.4319, pruned_loss=0.2252, over 4787.00 frames. ], tot_loss[loss=0.3827, simple_loss=0.3938, pruned_loss=0.1858, over 955947.66 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:46:34,898 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6492, 1.7447, 1.5443, 1.0306, 2.0670, 1.8346, 1.7082, 1.6123], + device='cuda:0'), covar=tensor([0.0906, 0.0700, 0.0971, 0.1209, 0.0449, 0.0857, 0.0922, 0.1193], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0131, 0.0133, 0.0122, 0.0107, 0.0132, 0.0138, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:46:35,023 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.93 vs. limit=5.0 +2023-03-25 22:46:47,077 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.580e+02 2.141e+02 2.900e+02 3.620e+02 1.043e+03, threshold=5.800e+02, percent-clipped=4.0 +2023-03-25 22:46:53,725 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-25 22:47:22,187 INFO [finetune.py:976] (0/7) Epoch 1, batch 3850, loss[loss=0.38, simple_loss=0.3986, pruned_loss=0.1807, over 4806.00 frames. ], tot_loss[loss=0.3782, simple_loss=0.3908, pruned_loss=0.1828, over 956113.65 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:47:45,152 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-25 22:47:53,457 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-25 22:48:07,747 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1694, 1.3899, 0.8092, 1.9717, 2.2547, 1.6581, 1.5308, 2.0076], + device='cuda:0'), covar=tensor([0.1527, 0.2019, 0.2471, 0.1179, 0.2380, 0.2141, 0.1371, 0.1780], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0094, 0.0112, 0.0090, 0.0122, 0.0092, 0.0095, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:48:09,988 INFO [finetune.py:976] (0/7) Epoch 1, batch 3900, loss[loss=0.3234, simple_loss=0.3534, pruned_loss=0.1467, over 4865.00 frames. ], tot_loss[loss=0.3713, simple_loss=0.3844, pruned_loss=0.1791, over 956575.34 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:48:10,642 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3902.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:48:20,243 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.580e+02 2.264e+02 2.673e+02 3.196e+02 5.181e+02, threshold=5.346e+02, percent-clipped=0.0 +2023-03-25 22:48:30,351 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3926.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:48:50,596 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3944.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:48:54,666 INFO [finetune.py:976] (0/7) Epoch 1, batch 3950, loss[loss=0.3704, simple_loss=0.3835, pruned_loss=0.1786, over 4874.00 frames. ], tot_loss[loss=0.3647, simple_loss=0.3787, pruned_loss=0.1753, over 955436.35 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:49:44,630 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1071, 1.7261, 1.5862, 0.5772, 1.5694, 1.8607, 1.5532, 1.8016], + device='cuda:0'), covar=tensor([0.0727, 0.0920, 0.1393, 0.2224, 0.1118, 0.2091, 0.2228, 0.0841], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0168, 0.0182, 0.0167, 0.0186, 0.0185, 0.0191, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:49:45,853 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3987.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:49:53,753 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3992.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:50:04,082 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-4000.pt +2023-03-25 22:50:05,838 INFO [finetune.py:976] (0/7) Epoch 1, batch 4000, loss[loss=0.3835, simple_loss=0.3941, pruned_loss=0.1864, over 4816.00 frames. ], tot_loss[loss=0.3633, simple_loss=0.3774, pruned_loss=0.1745, over 956572.78 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:50:18,329 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 2.072e+02 2.562e+02 2.941e+02 5.028e+02, threshold=5.123e+02, percent-clipped=0.0 +2023-03-25 22:50:31,331 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4032.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:50:50,289 INFO [finetune.py:976] (0/7) Epoch 1, batch 4050, loss[loss=0.3829, simple_loss=0.4099, pruned_loss=0.178, over 4826.00 frames. ], tot_loss[loss=0.3659, simple_loss=0.3811, pruned_loss=0.1754, over 956746.06 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:51:08,861 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1146, 1.8014, 2.5307, 1.4284, 2.1322, 2.1785, 1.8635, 2.5691], + device='cuda:0'), covar=tensor([0.2216, 0.2477, 0.1906, 0.3028, 0.1381, 0.2276, 0.2674, 0.1282], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0198, 0.0195, 0.0184, 0.0167, 0.0209, 0.0204, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:51:46,312 INFO [finetune.py:976] (0/7) Epoch 1, batch 4100, loss[loss=0.3285, simple_loss=0.3661, pruned_loss=0.1455, over 4813.00 frames. ], tot_loss[loss=0.367, simple_loss=0.3832, pruned_loss=0.1754, over 957137.00 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:52:00,028 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.481e+02 2.010e+02 2.495e+02 2.957e+02 5.246e+02, threshold=4.990e+02, percent-clipped=1.0 +2023-03-25 22:52:03,634 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5894, 1.3834, 1.9993, 3.2039, 2.2403, 2.3388, 0.5180, 2.5249], + device='cuda:0'), covar=tensor([0.2146, 0.1954, 0.1680, 0.0711, 0.1045, 0.1771, 0.2630, 0.0857], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0154, 0.0103, 0.0140, 0.0124, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:52:33,094 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-25 22:52:34,833 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4141.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:52:43,038 INFO [finetune.py:976] (0/7) Epoch 1, batch 4150, loss[loss=0.3358, simple_loss=0.3786, pruned_loss=0.1465, over 4849.00 frames. ], tot_loss[loss=0.3652, simple_loss=0.3822, pruned_loss=0.1741, over 956498.41 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:52:43,134 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8400, 1.6443, 2.2888, 3.4773, 2.6178, 2.4610, 1.0201, 2.6894], + device='cuda:0'), covar=tensor([0.1796, 0.1551, 0.1216, 0.0508, 0.0766, 0.1504, 0.1950, 0.0714], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0154, 0.0102, 0.0140, 0.0124, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 22:52:49,726 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-25 22:53:24,779 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1829, 1.0061, 0.7763, 0.7480, 0.9740, 0.9063, 0.9353, 1.6289], + device='cuda:0'), covar=tensor([4.0192, 4.2549, 3.6062, 5.8258, 3.0144, 2.4023, 4.2357, 1.2635], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0203, 0.0187, 0.0237, 0.0198, 0.0172, 0.0202, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 22:53:43,097 INFO [finetune.py:976] (0/7) Epoch 1, batch 4200, loss[loss=0.3253, simple_loss=0.3586, pruned_loss=0.146, over 4809.00 frames. ], tot_loss[loss=0.3634, simple_loss=0.3819, pruned_loss=0.1725, over 955607.75 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:53:43,808 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4202.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:53:43,834 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4202.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:53:44,462 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0672, 0.8524, 0.8998, 0.2246, 0.6834, 1.0862, 1.0670, 0.9661], + device='cuda:0'), covar=tensor([0.1242, 0.0779, 0.0529, 0.0836, 0.0495, 0.0394, 0.0420, 0.0621], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0141, 0.0110, 0.0119, 0.0119, 0.0111, 0.0135, 0.0140], + device='cuda:0'), out_proj_covar=tensor([8.9132e-05, 1.0526e-04, 8.0978e-05, 8.7672e-05, 8.6468e-05, 8.2123e-05, + 1.0070e-04, 1.0407e-04], device='cuda:0') +2023-03-25 22:54:02,773 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.292e+02 2.089e+02 2.432e+02 2.936e+02 5.530e+02, threshold=4.864e+02, percent-clipped=1.0 +2023-03-25 22:54:03,126 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-25 22:54:44,471 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4250.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:54:45,013 INFO [finetune.py:976] (0/7) Epoch 1, batch 4250, loss[loss=0.3304, simple_loss=0.3571, pruned_loss=0.1519, over 4891.00 frames. ], tot_loss[loss=0.358, simple_loss=0.3771, pruned_loss=0.1695, over 955158.59 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:54:54,864 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5155, 1.7105, 1.7237, 1.9640, 1.7470, 3.9404, 1.3540, 1.9211], + device='cuda:0'), covar=tensor([0.1111, 0.1760, 0.1421, 0.1136, 0.1609, 0.0174, 0.1593, 0.1763], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0076, 0.0072, 0.0075, 0.0088, 0.0076, 0.0082, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 22:55:23,143 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4282.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:55:24,425 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-25 22:55:27,315 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4288.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:55:28,560 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6734, 1.7428, 1.9552, 1.0919, 1.6802, 2.0495, 1.7550, 1.6355], + device='cuda:0'), covar=tensor([0.0849, 0.0646, 0.0367, 0.0592, 0.0384, 0.0409, 0.0354, 0.0543], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0141, 0.0110, 0.0120, 0.0120, 0.0111, 0.0135, 0.0140], + device='cuda:0'), out_proj_covar=tensor([8.8917e-05, 1.0504e-04, 8.1066e-05, 8.7784e-05, 8.6788e-05, 8.1975e-05, + 1.0067e-04, 1.0377e-04], device='cuda:0') +2023-03-25 22:55:35,146 INFO [finetune.py:976] (0/7) Epoch 1, batch 4300, loss[loss=0.2965, simple_loss=0.3402, pruned_loss=0.1264, over 4748.00 frames. ], tot_loss[loss=0.3532, simple_loss=0.3725, pruned_loss=0.167, over 954143.85 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:55:45,449 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.425e+02 1.950e+02 2.267e+02 2.860e+02 4.056e+02, threshold=4.534e+02, percent-clipped=0.0 +2023-03-25 22:55:51,019 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-25 22:55:52,114 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4903, 1.2653, 1.3856, 1.5340, 1.7106, 1.4113, 0.8831, 1.1938], + device='cuda:0'), covar=tensor([0.2781, 0.2842, 0.2326, 0.2046, 0.2348, 0.1669, 0.3677, 0.2050], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0192, 0.0179, 0.0166, 0.0213, 0.0164, 0.0190, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 22:56:13,415 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4332.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:56:35,734 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4349.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:56:36,849 INFO [finetune.py:976] (0/7) Epoch 1, batch 4350, loss[loss=0.3739, simple_loss=0.383, pruned_loss=0.1824, over 4748.00 frames. ], tot_loss[loss=0.346, simple_loss=0.3663, pruned_loss=0.1628, over 955311.75 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:57:06,043 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-25 22:57:17,016 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4380.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:57:44,938 INFO [finetune.py:976] (0/7) Epoch 1, batch 4400, loss[loss=0.3026, simple_loss=0.3451, pruned_loss=0.1301, over 4785.00 frames. ], tot_loss[loss=0.3452, simple_loss=0.3655, pruned_loss=0.1625, over 953191.87 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:57:57,029 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.978e+02 2.430e+02 2.895e+02 4.966e+02, threshold=4.860e+02, percent-clipped=1.0 +2023-03-25 22:58:28,110 INFO [finetune.py:976] (0/7) Epoch 1, batch 4450, loss[loss=0.3314, simple_loss=0.3594, pruned_loss=0.1517, over 4188.00 frames. ], tot_loss[loss=0.3502, simple_loss=0.3702, pruned_loss=0.1651, over 952915.72 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:59:09,860 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4497.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:59:12,225 INFO [finetune.py:976] (0/7) Epoch 1, batch 4500, loss[loss=0.3364, simple_loss=0.365, pruned_loss=0.1539, over 4824.00 frames. ], tot_loss[loss=0.3499, simple_loss=0.3714, pruned_loss=0.1642, over 954229.26 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:59:29,322 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 2.111e+02 2.516e+02 2.889e+02 5.762e+02, threshold=5.032e+02, percent-clipped=1.0 +2023-03-25 22:59:41,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.9915, 0.6570, 0.8536, 0.7490, 0.6550, 0.6396, 0.7360, 0.8500], + device='cuda:0'), covar=tensor([4.1837, 8.5957, 5.1023, 7.1424, 7.7021, 5.0502, 9.5490, 4.8044], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0230, 0.0215, 0.0244, 0.0227, 0.0198, 0.0253, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 22:59:44,722 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3258, 1.1557, 0.8471, 1.0519, 1.1314, 1.0399, 1.0607, 1.8789], + device='cuda:0'), covar=tensor([4.9602, 4.9151, 4.4732, 7.0513, 3.9611, 3.1787, 5.2017, 1.4707], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0204, 0.0188, 0.0238, 0.0199, 0.0172, 0.0203, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:00:15,079 INFO [finetune.py:976] (0/7) Epoch 1, batch 4550, loss[loss=0.3401, simple_loss=0.3856, pruned_loss=0.1473, over 4896.00 frames. ], tot_loss[loss=0.3483, simple_loss=0.3705, pruned_loss=0.163, over 955082.22 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:00:55,802 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4582.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:00:58,193 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5586, 1.6410, 1.5738, 1.6630, 0.8099, 3.1019, 1.0990, 1.7567], + device='cuda:0'), covar=tensor([0.3397, 0.2333, 0.2066, 0.2151, 0.2341, 0.0233, 0.3119, 0.1542], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0104, 0.0111, 0.0110, 0.0104, 0.0090, 0.0090, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:01:14,503 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.4309, 4.6200, 4.8886, 5.2628, 5.0568, 4.8720, 5.4918, 1.6907], + device='cuda:0'), covar=tensor([0.0578, 0.0776, 0.0577, 0.0707, 0.1187, 0.1094, 0.0490, 0.5101], + device='cuda:0'), in_proj_covar=tensor([0.0371, 0.0246, 0.0269, 0.0297, 0.0350, 0.0291, 0.0313, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:01:25,592 INFO [finetune.py:976] (0/7) Epoch 1, batch 4600, loss[loss=0.3656, simple_loss=0.3856, pruned_loss=0.1728, over 4783.00 frames. ], tot_loss[loss=0.3458, simple_loss=0.3688, pruned_loss=0.1614, over 956405.12 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:01:38,110 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 2.082e+02 2.456e+02 3.064e+02 5.977e+02, threshold=4.911e+02, percent-clipped=1.0 +2023-03-25 23:01:48,194 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-25 23:01:59,342 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4630.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:02:18,003 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4644.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:02:28,981 INFO [finetune.py:976] (0/7) Epoch 1, batch 4650, loss[loss=0.2911, simple_loss=0.3263, pruned_loss=0.128, over 4861.00 frames. ], tot_loss[loss=0.3422, simple_loss=0.3657, pruned_loss=0.1593, over 957181.36 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:02:58,370 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2927, 1.0936, 0.9071, 0.8899, 1.0627, 0.9680, 1.0281, 1.8322], + device='cuda:0'), covar=tensor([4.6589, 4.6137, 3.8799, 6.5043, 3.6612, 3.0055, 4.9815, 1.4256], + device='cuda:0'), in_proj_covar=tensor([0.0214, 0.0205, 0.0189, 0.0238, 0.0200, 0.0172, 0.0204, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:03:02,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0556, 1.7009, 2.3370, 3.7562, 2.9767, 2.6499, 0.8331, 3.1618], + device='cuda:0'), covar=tensor([0.1833, 0.1632, 0.1470, 0.0466, 0.0758, 0.1539, 0.2231, 0.0618], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0155, 0.0102, 0.0141, 0.0126, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 23:03:06,337 INFO [finetune.py:976] (0/7) Epoch 1, batch 4700, loss[loss=0.262, simple_loss=0.2973, pruned_loss=0.1134, over 4919.00 frames. ], tot_loss[loss=0.3349, simple_loss=0.3598, pruned_loss=0.155, over 959305.44 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:03:09,203 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9108, 4.6845, 4.5789, 2.5941, 4.7508, 3.4417, 0.9688, 3.4639], + device='cuda:0'), covar=tensor([0.2479, 0.1353, 0.1237, 0.3171, 0.0710, 0.0910, 0.5034, 0.1435], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0157, 0.0161, 0.0125, 0.0151, 0.0116, 0.0144, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-25 23:03:20,653 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.870e+02 2.224e+02 2.796e+02 5.273e+02, threshold=4.448e+02, percent-clipped=2.0 +2023-03-25 23:03:49,669 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8741, 3.3806, 3.5274, 3.7567, 3.5909, 3.4191, 3.9743, 1.3473], + device='cuda:0'), covar=tensor([0.0859, 0.0903, 0.0871, 0.1059, 0.1406, 0.1424, 0.0774, 0.4861], + device='cuda:0'), in_proj_covar=tensor([0.0371, 0.0246, 0.0268, 0.0297, 0.0349, 0.0290, 0.0313, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:03:59,445 INFO [finetune.py:976] (0/7) Epoch 1, batch 4750, loss[loss=0.309, simple_loss=0.3384, pruned_loss=0.1398, over 4790.00 frames. ], tot_loss[loss=0.331, simple_loss=0.3563, pruned_loss=0.1529, over 958117.92 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:04:01,468 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-25 23:04:28,261 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-03-25 23:04:36,391 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4797.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:04:39,215 INFO [finetune.py:976] (0/7) Epoch 1, batch 4800, loss[loss=0.3327, simple_loss=0.3357, pruned_loss=0.1648, over 3946.00 frames. ], tot_loss[loss=0.3343, simple_loss=0.3595, pruned_loss=0.1545, over 955217.37 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:04:46,451 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-25 23:04:56,835 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.406e+02 2.096e+02 2.556e+02 3.186e+02 5.883e+02, threshold=5.111e+02, percent-clipped=4.0 +2023-03-25 23:04:56,943 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4920, 1.3183, 2.0517, 3.0026, 2.2245, 2.1503, 1.0749, 2.3853], + device='cuda:0'), covar=tensor([0.1857, 0.1703, 0.1204, 0.0519, 0.0783, 0.2133, 0.1834, 0.0704], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0157, 0.0103, 0.0142, 0.0126, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 23:05:08,479 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.8302, 0.2704, 0.8334, 0.5565, 0.6307, 0.6171, 0.5344, 0.7875], + device='cuda:0'), covar=tensor([3.7071, 7.6260, 4.8752, 5.7007, 6.8312, 4.1270, 8.1633, 4.4114], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0233, 0.0218, 0.0247, 0.0229, 0.0201, 0.0256, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:05:22,870 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.48 vs. limit=2.0 +2023-03-25 23:05:32,154 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4845.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:05:41,985 INFO [finetune.py:976] (0/7) Epoch 1, batch 4850, loss[loss=0.2607, simple_loss=0.3035, pruned_loss=0.109, over 4747.00 frames. ], tot_loss[loss=0.3364, simple_loss=0.3628, pruned_loss=0.155, over 954790.25 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:06:28,850 INFO [finetune.py:976] (0/7) Epoch 1, batch 4900, loss[loss=0.3413, simple_loss=0.3746, pruned_loss=0.154, over 4798.00 frames. ], tot_loss[loss=0.3363, simple_loss=0.363, pruned_loss=0.1548, over 955552.13 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:06:45,502 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.521e+02 2.056e+02 2.408e+02 2.893e+02 5.886e+02, threshold=4.817e+02, percent-clipped=2.0 +2023-03-25 23:07:15,535 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4944.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:07:19,810 INFO [finetune.py:976] (0/7) Epoch 1, batch 4950, loss[loss=0.3428, simple_loss=0.3651, pruned_loss=0.1603, over 4778.00 frames. ], tot_loss[loss=0.337, simple_loss=0.3639, pruned_loss=0.1551, over 955436.89 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:07:37,627 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4739, 1.3057, 1.8329, 2.7743, 1.9992, 1.9892, 1.0398, 2.2194], + device='cuda:0'), covar=tensor([0.1821, 0.1616, 0.1206, 0.0556, 0.0846, 0.1558, 0.1654, 0.0690], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0155, 0.0102, 0.0141, 0.0125, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 23:07:38,250 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3913, 1.2715, 1.5480, 2.3901, 1.7899, 1.9906, 0.8769, 1.9645], + device='cuda:0'), covar=tensor([0.1869, 0.1576, 0.1206, 0.0713, 0.0881, 0.1251, 0.1669, 0.0837], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0155, 0.0102, 0.0141, 0.0125, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 23:08:11,783 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4992.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:08:22,746 INFO [finetune.py:976] (0/7) Epoch 1, batch 5000, loss[loss=0.3623, simple_loss=0.3792, pruned_loss=0.1728, over 4846.00 frames. ], tot_loss[loss=0.3333, simple_loss=0.3609, pruned_loss=0.1529, over 956333.57 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:08:32,727 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 2.131e+02 2.461e+02 3.038e+02 5.796e+02, threshold=4.923e+02, percent-clipped=4.0 +2023-03-25 23:08:49,849 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-25 23:09:20,248 INFO [finetune.py:976] (0/7) Epoch 1, batch 5050, loss[loss=0.2696, simple_loss=0.3167, pruned_loss=0.1112, over 4789.00 frames. ], tot_loss[loss=0.3296, simple_loss=0.3573, pruned_loss=0.151, over 958251.33 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:09:20,974 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6127, 1.7388, 1.6257, 1.0480, 1.9885, 1.7366, 1.6666, 1.6030], + device='cuda:0'), covar=tensor([0.0854, 0.0728, 0.0859, 0.1163, 0.0500, 0.0915, 0.0917, 0.1196], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0129, 0.0134, 0.0122, 0.0106, 0.0132, 0.0138, 0.0158], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:09:25,145 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4859, 1.7393, 1.6610, 1.9554, 1.7808, 3.5589, 1.5039, 1.8363], + device='cuda:0'), covar=tensor([0.1198, 0.1694, 0.1243, 0.1140, 0.1596, 0.0241, 0.1555, 0.1766], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0077, 0.0073, 0.0076, 0.0089, 0.0077, 0.0082, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:09:51,475 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6939, 1.7865, 1.7664, 1.2058, 2.0134, 1.8935, 1.7921, 1.6186], + device='cuda:0'), covar=tensor([0.0729, 0.0660, 0.0700, 0.1024, 0.0451, 0.0727, 0.0770, 0.1188], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0129, 0.0134, 0.0122, 0.0106, 0.0132, 0.0138, 0.0158], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:10:01,168 INFO [finetune.py:976] (0/7) Epoch 1, batch 5100, loss[loss=0.2928, simple_loss=0.3153, pruned_loss=0.1352, over 4747.00 frames. ], tot_loss[loss=0.3239, simple_loss=0.3519, pruned_loss=0.148, over 956870.80 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:10:09,454 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.918e+02 2.379e+02 2.966e+02 8.444e+02, threshold=4.758e+02, percent-clipped=2.0 +2023-03-25 23:10:34,835 INFO [finetune.py:976] (0/7) Epoch 1, batch 5150, loss[loss=0.2852, simple_loss=0.3332, pruned_loss=0.1187, over 4791.00 frames. ], tot_loss[loss=0.3229, simple_loss=0.351, pruned_loss=0.1474, over 955759.21 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:10:41,599 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-25 23:11:14,837 INFO [finetune.py:976] (0/7) Epoch 1, batch 5200, loss[loss=0.3071, simple_loss=0.3551, pruned_loss=0.1295, over 4935.00 frames. ], tot_loss[loss=0.3244, simple_loss=0.3535, pruned_loss=0.1477, over 954960.17 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:11:24,773 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.601e+02 2.261e+02 2.570e+02 3.078e+02 5.221e+02, threshold=5.140e+02, percent-clipped=2.0 +2023-03-25 23:12:06,968 INFO [finetune.py:976] (0/7) Epoch 1, batch 5250, loss[loss=0.3758, simple_loss=0.4047, pruned_loss=0.1735, over 4803.00 frames. ], tot_loss[loss=0.3267, simple_loss=0.3568, pruned_loss=0.1483, over 952585.43 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:12:07,773 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-25 23:12:14,754 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-25 23:12:55,129 INFO [finetune.py:976] (0/7) Epoch 1, batch 5300, loss[loss=0.3146, simple_loss=0.3539, pruned_loss=0.1377, over 4811.00 frames. ], tot_loss[loss=0.3279, simple_loss=0.3584, pruned_loss=0.1487, over 953990.79 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:13:05,584 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-25 23:13:08,327 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.321e+02 2.032e+02 2.465e+02 2.907e+02 4.480e+02, threshold=4.930e+02, percent-clipped=0.0 +2023-03-25 23:13:39,600 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6271, 3.8134, 3.6426, 1.8424, 3.9628, 2.9425, 0.9385, 2.6490], + device='cuda:0'), covar=tensor([0.2574, 0.1426, 0.1517, 0.3339, 0.0919, 0.0911, 0.4540, 0.1321], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0158, 0.0162, 0.0125, 0.0152, 0.0116, 0.0143, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-25 23:13:49,132 INFO [finetune.py:976] (0/7) Epoch 1, batch 5350, loss[loss=0.3357, simple_loss=0.3569, pruned_loss=0.1572, over 4745.00 frames. ], tot_loss[loss=0.3247, simple_loss=0.356, pruned_loss=0.1467, over 951501.48 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:14:48,111 INFO [finetune.py:976] (0/7) Epoch 1, batch 5400, loss[loss=0.3071, simple_loss=0.3371, pruned_loss=0.1385, over 4818.00 frames. ], tot_loss[loss=0.3223, simple_loss=0.3533, pruned_loss=0.1456, over 953927.88 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:14:48,240 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0680, 0.9494, 0.9811, 0.3296, 0.7506, 1.1399, 1.1644, 1.0211], + device='cuda:0'), covar=tensor([0.1058, 0.0633, 0.0421, 0.0758, 0.0523, 0.0470, 0.0374, 0.0587], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0144, 0.0111, 0.0122, 0.0122, 0.0112, 0.0137, 0.0139], + device='cuda:0'), out_proj_covar=tensor([9.0820e-05, 1.0758e-04, 8.1612e-05, 8.9575e-05, 8.8542e-05, 8.2382e-05, + 1.0254e-04, 1.0347e-04], device='cuda:0') +2023-03-25 23:14:55,970 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.226e+02 1.939e+02 2.339e+02 2.729e+02 4.650e+02, threshold=4.678e+02, percent-clipped=0.0 +2023-03-25 23:15:01,743 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3705, 2.9076, 2.0399, 1.7597, 3.1849, 2.9919, 2.6120, 2.4225], + device='cuda:0'), covar=tensor([0.0917, 0.0564, 0.1105, 0.1258, 0.0315, 0.0875, 0.0967, 0.1059], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0131, 0.0136, 0.0125, 0.0107, 0.0135, 0.0140, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:15:11,999 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-25 23:15:39,348 INFO [finetune.py:976] (0/7) Epoch 1, batch 5450, loss[loss=0.2667, simple_loss=0.2885, pruned_loss=0.1225, over 4784.00 frames. ], tot_loss[loss=0.3164, simple_loss=0.3474, pruned_loss=0.1427, over 955631.26 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:15:47,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9132, 4.0695, 3.9333, 2.1321, 4.2260, 3.1656, 1.1726, 3.0270], + device='cuda:0'), covar=tensor([0.2336, 0.1477, 0.1689, 0.3340, 0.0799, 0.0942, 0.4332, 0.1390], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0157, 0.0161, 0.0125, 0.0151, 0.0115, 0.0142, 0.0117], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-25 23:16:04,233 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-25 23:16:20,038 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0431, 0.8809, 1.0157, 0.4649, 0.7862, 1.1623, 1.1419, 1.0478], + device='cuda:0'), covar=tensor([0.1005, 0.0579, 0.0409, 0.0638, 0.0482, 0.0448, 0.0426, 0.0567], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0145, 0.0112, 0.0123, 0.0123, 0.0112, 0.0137, 0.0140], + device='cuda:0'), out_proj_covar=tensor([9.1279e-05, 1.0802e-04, 8.1847e-05, 8.9993e-05, 8.9002e-05, 8.2516e-05, + 1.0276e-04, 1.0359e-04], device='cuda:0') +2023-03-25 23:16:21,230 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7429, 1.5054, 2.0842, 1.3920, 1.7888, 1.9678, 1.5446, 2.1208], + device='cuda:0'), covar=tensor([0.1925, 0.2403, 0.1697, 0.2138, 0.1162, 0.1745, 0.2988, 0.1119], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0201, 0.0200, 0.0190, 0.0172, 0.0217, 0.0209, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:16:30,507 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5351, 1.3408, 1.0633, 1.1831, 1.2945, 1.1441, 1.1952, 2.1312], + device='cuda:0'), covar=tensor([4.1013, 4.0710, 3.3462, 5.5293, 3.1436, 2.3720, 4.2047, 1.0763], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0212, 0.0195, 0.0247, 0.0207, 0.0177, 0.0211, 0.0158], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:16:31,588 INFO [finetune.py:976] (0/7) Epoch 1, batch 5500, loss[loss=0.2731, simple_loss=0.3124, pruned_loss=0.1169, over 4911.00 frames. ], tot_loss[loss=0.3115, simple_loss=0.3427, pruned_loss=0.1402, over 955137.75 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:16:42,085 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2424, 0.3754, 1.1738, 0.8481, 0.8811, 0.8697, 0.8064, 1.1515], + device='cuda:0'), covar=tensor([3.0552, 6.7662, 4.1399, 4.9006, 5.4954, 3.4002, 6.5790, 4.0413], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0240, 0.0225, 0.0254, 0.0234, 0.0206, 0.0262, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:16:45,961 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.302e+02 2.026e+02 2.277e+02 2.875e+02 1.009e+03, threshold=4.553e+02, percent-clipped=5.0 +2023-03-25 23:17:20,559 INFO [finetune.py:976] (0/7) Epoch 1, batch 5550, loss[loss=0.3523, simple_loss=0.3872, pruned_loss=0.1587, over 4907.00 frames. ], tot_loss[loss=0.3118, simple_loss=0.3429, pruned_loss=0.1403, over 956040.45 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:17:30,716 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-25 23:18:01,090 INFO [finetune.py:976] (0/7) Epoch 1, batch 5600, loss[loss=0.3359, simple_loss=0.3635, pruned_loss=0.1542, over 4827.00 frames. ], tot_loss[loss=0.3151, simple_loss=0.3472, pruned_loss=0.1415, over 954940.98 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:18:19,453 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.468e+02 1.839e+02 2.287e+02 2.793e+02 4.099e+02, threshold=4.573e+02, percent-clipped=0.0 +2023-03-25 23:18:59,388 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5650.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:18:59,875 INFO [finetune.py:976] (0/7) Epoch 1, batch 5650, loss[loss=0.3996, simple_loss=0.4051, pruned_loss=0.1971, over 4057.00 frames. ], tot_loss[loss=0.3181, simple_loss=0.3504, pruned_loss=0.1429, over 954619.16 frames. ], batch size: 66, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:19:14,882 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-25 23:19:35,525 INFO [finetune.py:976] (0/7) Epoch 1, batch 5700, loss[loss=0.2229, simple_loss=0.2615, pruned_loss=0.09216, over 4617.00 frames. ], tot_loss[loss=0.3138, simple_loss=0.3446, pruned_loss=0.1415, over 934367.61 frames. ], batch size: 20, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:19:40,321 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1624, 2.7515, 2.6236, 2.5813, 2.5962, 4.7504, 2.2729, 2.8181], + device='cuda:0'), covar=tensor([0.0903, 0.1154, 0.0883, 0.0882, 0.1201, 0.0171, 0.1088, 0.1255], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0077, 0.0074, 0.0076, 0.0090, 0.0078, 0.0083, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:19:41,519 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5711.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:19:43,179 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.818e+02 2.245e+02 2.685e+02 4.321e+02, threshold=4.489e+02, percent-clipped=0.0 +2023-03-25 23:19:45,667 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1878, 2.1171, 2.4500, 1.0922, 2.3708, 2.8429, 2.1269, 2.1661], + device='cuda:0'), covar=tensor([0.1040, 0.0793, 0.0496, 0.0868, 0.0632, 0.0399, 0.0524, 0.0639], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0144, 0.0111, 0.0121, 0.0122, 0.0111, 0.0136, 0.0139], + device='cuda:0'), out_proj_covar=tensor([9.0462e-05, 1.0711e-04, 8.1569e-05, 8.9201e-05, 8.8613e-05, 8.1674e-05, + 1.0179e-04, 1.0289e-04], device='cuda:0') +2023-03-25 23:19:55,554 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.08 vs. limit=5.0 +2023-03-25 23:19:57,403 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-1.pt +2023-03-25 23:20:08,411 INFO [finetune.py:976] (0/7) Epoch 2, batch 0, loss[loss=0.368, simple_loss=0.3816, pruned_loss=0.1772, over 4837.00 frames. ], tot_loss[loss=0.368, simple_loss=0.3816, pruned_loss=0.1772, over 4837.00 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:20:08,412 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-25 23:20:25,000 INFO [finetune.py:1010] (0/7) Epoch 2, validation: loss=0.2224, simple_loss=0.2847, pruned_loss=0.08, over 2265189.00 frames. +2023-03-25 23:20:25,000 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6193MB +2023-03-25 23:20:47,068 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-25 23:20:56,829 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5755.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:21:22,830 INFO [finetune.py:976] (0/7) Epoch 2, batch 50, loss[loss=0.2688, simple_loss=0.325, pruned_loss=0.1063, over 4818.00 frames. ], tot_loss[loss=0.3197, simple_loss=0.3502, pruned_loss=0.1445, over 214140.87 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:21:54,356 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.870e+02 2.317e+02 2.912e+02 7.564e+02, threshold=4.633e+02, percent-clipped=3.0 +2023-03-25 23:21:55,671 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5816.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:22:11,072 INFO [finetune.py:976] (0/7) Epoch 2, batch 100, loss[loss=0.3139, simple_loss=0.3455, pruned_loss=0.1412, over 4867.00 frames. ], tot_loss[loss=0.3089, simple_loss=0.3406, pruned_loss=0.1385, over 379154.57 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:22:35,167 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5858.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:22:37,556 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-25 23:22:49,652 INFO [finetune.py:976] (0/7) Epoch 2, batch 150, loss[loss=0.2168, simple_loss=0.2592, pruned_loss=0.08716, over 4711.00 frames. ], tot_loss[loss=0.3014, simple_loss=0.3332, pruned_loss=0.1348, over 507994.01 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:23:18,197 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.883e+02 2.329e+02 2.858e+02 5.160e+02, threshold=4.657e+02, percent-clipped=2.0 +2023-03-25 23:23:21,840 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5919.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:23:28,108 INFO [finetune.py:976] (0/7) Epoch 2, batch 200, loss[loss=0.2883, simple_loss=0.3261, pruned_loss=0.1252, over 4867.00 frames. ], tot_loss[loss=0.3003, simple_loss=0.3322, pruned_loss=0.1341, over 607768.53 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:23:31,323 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-25 23:23:45,578 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0907, 1.7165, 2.5943, 1.4283, 2.2244, 2.0634, 1.6735, 2.3576], + device='cuda:0'), covar=tensor([0.1908, 0.2232, 0.1989, 0.3045, 0.1140, 0.2246, 0.2797, 0.1412], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0202, 0.0200, 0.0191, 0.0173, 0.0218, 0.0209, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:24:01,222 INFO [finetune.py:976] (0/7) Epoch 2, batch 250, loss[loss=0.2471, simple_loss=0.2984, pruned_loss=0.09793, over 4900.00 frames. ], tot_loss[loss=0.3044, simple_loss=0.337, pruned_loss=0.136, over 685009.13 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:24:20,400 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5990.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:24:31,870 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-6000.pt +2023-03-25 23:24:42,064 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6006.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:24:48,587 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.968e+02 2.365e+02 2.842e+02 7.361e+02, threshold=4.731e+02, percent-clipped=2.0 +2023-03-25 23:25:01,945 INFO [finetune.py:976] (0/7) Epoch 2, batch 300, loss[loss=0.298, simple_loss=0.3283, pruned_loss=0.1339, over 4771.00 frames. ], tot_loss[loss=0.3079, simple_loss=0.3406, pruned_loss=0.1376, over 744957.48 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:25:23,414 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6043.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:25:33,840 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6051.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:11,335 INFO [finetune.py:976] (0/7) Epoch 2, batch 350, loss[loss=0.2935, simple_loss=0.3364, pruned_loss=0.1253, over 4923.00 frames. ], tot_loss[loss=0.3122, simple_loss=0.3456, pruned_loss=0.1394, over 792799.34 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 32.0 +2023-03-25 23:26:29,578 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5615, 1.4592, 2.1796, 3.1853, 2.3193, 2.2553, 1.1128, 2.5786], + device='cuda:0'), covar=tensor([0.2003, 0.1824, 0.1321, 0.0606, 0.0859, 0.1647, 0.1916, 0.0771], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0118, 0.0137, 0.0159, 0.0104, 0.0144, 0.0128, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-25 23:26:34,996 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6104.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:39,163 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6111.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:41,011 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3531, 2.8351, 1.8726, 1.7367, 3.0538, 2.7174, 2.4019, 2.4447], + device='cuda:0'), covar=tensor([0.0896, 0.0530, 0.1126, 0.1156, 0.0354, 0.0917, 0.0977, 0.0957], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0137, 0.0126, 0.0107, 0.0136, 0.0142, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:26:41,481 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.275e+02 2.097e+02 2.536e+02 2.955e+02 5.135e+02, threshold=5.071e+02, percent-clipped=1.0 +2023-03-25 23:26:49,262 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8842, 4.2528, 4.4227, 4.7102, 4.5627, 4.3953, 5.0001, 1.4679], + device='cuda:0'), covar=tensor([0.0700, 0.0780, 0.0601, 0.0816, 0.1323, 0.1193, 0.0512, 0.5383], + device='cuda:0'), in_proj_covar=tensor([0.0371, 0.0246, 0.0270, 0.0298, 0.0350, 0.0291, 0.0312, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:26:59,642 INFO [finetune.py:976] (0/7) Epoch 2, batch 400, loss[loss=0.3004, simple_loss=0.3472, pruned_loss=0.1268, over 4884.00 frames. ], tot_loss[loss=0.3134, simple_loss=0.3476, pruned_loss=0.1396, over 825995.16 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:27:09,961 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6135.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:27:24,770 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6761, 1.4898, 1.2483, 1.4498, 1.4971, 1.3592, 1.3985, 2.5006], + device='cuda:0'), covar=tensor([3.1508, 3.9089, 2.7207, 4.7092, 2.7205, 1.8884, 3.5099, 0.8812], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0217, 0.0198, 0.0252, 0.0210, 0.0179, 0.0215, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:27:49,725 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6170.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:27:59,427 INFO [finetune.py:976] (0/7) Epoch 2, batch 450, loss[loss=0.3318, simple_loss=0.3665, pruned_loss=0.1486, over 4915.00 frames. ], tot_loss[loss=0.3114, simple_loss=0.346, pruned_loss=0.1384, over 855159.84 frames. ], batch size: 46, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:28:10,825 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1896, 1.5324, 1.1556, 1.5544, 1.4910, 2.9129, 1.3193, 1.5619], + device='cuda:0'), covar=tensor([0.1228, 0.1765, 0.1485, 0.1211, 0.1719, 0.0324, 0.1550, 0.1859], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0078, 0.0075, 0.0077, 0.0090, 0.0078, 0.0083, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:28:14,495 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6196.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:25,613 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6203.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:33,037 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6211.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:34,811 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6214.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:35,341 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.932e+02 2.244e+02 2.718e+02 3.817e+02, threshold=4.487e+02, percent-clipped=0.0 +2023-03-25 23:28:45,038 INFO [finetune.py:976] (0/7) Epoch 2, batch 500, loss[loss=0.2486, simple_loss=0.2916, pruned_loss=0.1028, over 4804.00 frames. ], tot_loss[loss=0.3072, simple_loss=0.3423, pruned_loss=0.1361, over 877122.11 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:28:52,709 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6231.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:54,735 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-25 23:29:24,969 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5069, 2.2101, 1.9872, 1.0331, 2.0863, 2.0078, 1.6413, 1.9847], + device='cuda:0'), covar=tensor([0.0800, 0.1085, 0.2142, 0.2679, 0.1622, 0.2181, 0.2223, 0.1443], + device='cuda:0'), in_proj_covar=tensor([0.0160, 0.0180, 0.0193, 0.0178, 0.0201, 0.0199, 0.0203, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:29:25,552 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6264.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:26,762 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3256, 1.5670, 1.3917, 1.6943, 1.4431, 3.2113, 1.2556, 1.5553], + device='cuda:0'), covar=tensor([0.1136, 0.1636, 0.1293, 0.1142, 0.1746, 0.0277, 0.1539, 0.1816], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0078, 0.0074, 0.0077, 0.0090, 0.0078, 0.0083, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:29:26,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6266.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:35,118 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6272.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:35,723 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6173, 1.5997, 1.2905, 1.2521, 1.8184, 2.0048, 1.6662, 1.2623], + device='cuda:0'), covar=tensor([0.0284, 0.0436, 0.0557, 0.0444, 0.0246, 0.0247, 0.0329, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0110, 0.0129, 0.0109, 0.0102, 0.0097, 0.0086, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.4009e-05, 8.6722e-05, 1.0429e-04, 8.6761e-05, 8.1336e-05, 7.2414e-05, + 6.6485e-05, 8.3949e-05], device='cuda:0') +2023-03-25 23:29:38,744 INFO [finetune.py:976] (0/7) Epoch 2, batch 550, loss[loss=0.2876, simple_loss=0.3269, pruned_loss=0.1241, over 4760.00 frames. ], tot_loss[loss=0.3028, simple_loss=0.3381, pruned_loss=0.1337, over 896336.78 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:29:58,226 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3102, 1.0845, 0.9731, 0.7628, 1.0707, 1.0112, 1.0455, 1.8032], + device='cuda:0'), covar=tensor([2.6031, 2.2164, 2.0149, 3.0052, 1.8394, 1.4290, 2.2884, 0.7174], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0217, 0.0198, 0.0253, 0.0210, 0.0179, 0.0215, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-25 23:30:17,690 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6306.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:30:29,435 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.346e+02 1.947e+02 2.353e+02 2.718e+02 5.175e+02, threshold=4.705e+02, percent-clipped=1.0 +2023-03-25 23:30:41,993 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6327.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:30:48,005 INFO [finetune.py:976] (0/7) Epoch 2, batch 600, loss[loss=0.3039, simple_loss=0.3446, pruned_loss=0.1316, over 4876.00 frames. ], tot_loss[loss=0.3027, simple_loss=0.3378, pruned_loss=0.1338, over 911679.84 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:31:07,304 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6346.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:12,641 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6354.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:31:23,953 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.68 vs. limit=2.0 +2023-03-25 23:31:28,047 INFO [finetune.py:976] (0/7) Epoch 2, batch 650, loss[loss=0.2724, simple_loss=0.3266, pruned_loss=0.1091, over 4751.00 frames. ], tot_loss[loss=0.3043, simple_loss=0.3397, pruned_loss=0.1344, over 922024.67 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:31:37,792 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5326, 1.6048, 1.4865, 1.7335, 1.0211, 3.2577, 1.2158, 1.6971], + device='cuda:0'), covar=tensor([0.3615, 0.2456, 0.2245, 0.2235, 0.2157, 0.0238, 0.3181, 0.1594], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0107, 0.0113, 0.0114, 0.0108, 0.0092, 0.0095, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-25 23:31:42,384 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6399.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:51,330 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6411.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:53,613 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.435e+02 2.011e+02 2.373e+02 2.999e+02 4.783e+02, threshold=4.746e+02, percent-clipped=1.0 +2023-03-25 23:32:01,499 INFO [finetune.py:976] (0/7) Epoch 2, batch 700, loss[loss=0.3088, simple_loss=0.3503, pruned_loss=0.1336, over 4770.00 frames. ], tot_loss[loss=0.3056, simple_loss=0.3414, pruned_loss=0.1349, over 929431.61 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:32:02,223 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0149, 2.1329, 1.9399, 1.2891, 2.3935, 2.2036, 2.0204, 1.9160], + device='cuda:0'), covar=tensor([0.0656, 0.0532, 0.0654, 0.0927, 0.0456, 0.0594, 0.0698, 0.0951], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0133, 0.0139, 0.0128, 0.0109, 0.0137, 0.0144, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:32:22,199 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6459.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:24,554 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6462.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:27,956 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6467.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:34,494 INFO [finetune.py:976] (0/7) Epoch 2, batch 750, loss[loss=0.3667, simple_loss=0.3969, pruned_loss=0.1682, over 4718.00 frames. ], tot_loss[loss=0.3044, simple_loss=0.3408, pruned_loss=0.134, over 934503.41 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:32:42,489 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6491.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:58,343 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6514.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:58,813 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.464e+02 1.998e+02 2.267e+02 2.688e+02 5.596e+02, threshold=4.534e+02, percent-clipped=2.0 +2023-03-25 23:33:04,284 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6523.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:33:06,024 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6526.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:07,177 INFO [finetune.py:976] (0/7) Epoch 2, batch 800, loss[loss=0.2618, simple_loss=0.3097, pruned_loss=0.107, over 4805.00 frames. ], tot_loss[loss=0.3028, simple_loss=0.34, pruned_loss=0.1328, over 939933.88 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:33:07,425 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6528.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:40,448 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6559.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:42,339 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6562.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:46,044 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6567.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:58,058 INFO [finetune.py:976] (0/7) Epoch 2, batch 850, loss[loss=0.2694, simple_loss=0.3201, pruned_loss=0.1094, over 4819.00 frames. ], tot_loss[loss=0.2998, simple_loss=0.3368, pruned_loss=0.1314, over 942554.21 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:34:16,610 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8700, 2.3144, 1.9516, 1.4890, 2.5299, 3.2961, 2.7777, 2.2707], + device='cuda:0'), covar=tensor([0.0150, 0.0486, 0.0462, 0.0526, 0.0236, 0.0154, 0.0198, 0.0354], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0111, 0.0130, 0.0111, 0.0103, 0.0098, 0.0087, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.4448e-05, 8.7838e-05, 1.0549e-04, 8.7859e-05, 8.1819e-05, 7.3046e-05, + 6.7276e-05, 8.4652e-05], device='cuda:0') +2023-03-25 23:34:37,095 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.789e+02 2.218e+02 2.697e+02 5.451e+02, threshold=4.436e+02, percent-clipped=1.0 +2023-03-25 23:34:42,614 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6622.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:34:46,286 INFO [finetune.py:976] (0/7) Epoch 2, batch 900, loss[loss=0.2821, simple_loss=0.3009, pruned_loss=0.1317, over 4249.00 frames. ], tot_loss[loss=0.2945, simple_loss=0.3321, pruned_loss=0.1284, over 944548.33 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:34:47,091 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6629.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:34:48,447 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-25 23:35:02,216 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7438, 1.7599, 1.6127, 1.7564, 0.9220, 3.7252, 1.4256, 1.9165], + device='cuda:0'), covar=tensor([0.3560, 0.2570, 0.2130, 0.2231, 0.2247, 0.0155, 0.2782, 0.1557], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0107, 0.0113, 0.0114, 0.0108, 0.0092, 0.0095, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-25 23:35:02,805 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6646.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:04,818 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-03-25 23:35:25,038 INFO [finetune.py:976] (0/7) Epoch 2, batch 950, loss[loss=0.2805, simple_loss=0.3252, pruned_loss=0.1179, over 4904.00 frames. ], tot_loss[loss=0.2908, simple_loss=0.3283, pruned_loss=0.1266, over 945749.09 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:35:32,448 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6690.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:34,866 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6694.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:34,912 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4472, 1.5612, 1.3892, 1.6926, 1.5477, 3.2565, 1.2994, 1.5991], + device='cuda:0'), covar=tensor([0.1099, 0.1751, 0.1323, 0.1119, 0.1784, 0.0269, 0.1647, 0.1927], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0078, 0.0074, 0.0077, 0.0090, 0.0079, 0.0083, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-25 23:35:37,912 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6699.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:48,888 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.729e+02 2.077e+02 2.706e+02 4.933e+02, threshold=4.155e+02, percent-clipped=1.0 +2023-03-25 23:35:49,078 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-25 23:35:51,490 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6435, 1.5186, 1.3087, 1.2229, 1.8598, 1.9342, 1.6627, 1.2863], + device='cuda:0'), covar=tensor([0.0320, 0.0410, 0.0536, 0.0440, 0.0262, 0.0358, 0.0305, 0.0481], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0112, 0.0131, 0.0112, 0.0103, 0.0098, 0.0088, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.4752e-05, 8.8619e-05, 1.0617e-04, 8.8526e-05, 8.2312e-05, 7.3352e-05, + 6.7945e-05, 8.5421e-05], device='cuda:0') +2023-03-25 23:36:03,623 INFO [finetune.py:976] (0/7) Epoch 2, batch 1000, loss[loss=0.3348, simple_loss=0.3472, pruned_loss=0.1611, over 4083.00 frames. ], tot_loss[loss=0.2943, simple_loss=0.3317, pruned_loss=0.1284, over 947664.61 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:36:22,930 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6747.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:01,246 INFO [finetune.py:976] (0/7) Epoch 2, batch 1050, loss[loss=0.341, simple_loss=0.3745, pruned_loss=0.1537, over 4858.00 frames. ], tot_loss[loss=0.2981, simple_loss=0.3359, pruned_loss=0.1302, over 948908.31 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:37:09,199 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6791.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:17,701 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-25 23:37:29,664 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.307e+02 2.050e+02 2.547e+02 2.958e+02 5.414e+02, threshold=5.095e+02, percent-clipped=8.0 +2023-03-25 23:37:37,310 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6818.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:37:40,266 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6823.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:47,836 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6826.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:48,963 INFO [finetune.py:976] (0/7) Epoch 2, batch 1100, loss[loss=0.2909, simple_loss=0.3346, pruned_loss=0.1236, over 4814.00 frames. ], tot_loss[loss=0.2994, simple_loss=0.3379, pruned_loss=0.1305, over 950898.23 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:37:56,676 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6839.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:56,759 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6602, 1.6440, 1.2425, 1.8705, 1.8505, 1.2984, 2.3610, 1.6479], + device='cuda:0'), covar=tensor([0.3027, 0.7687, 0.6311, 0.6294, 0.3982, 0.3086, 0.4945, 0.4388], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0189, 0.0232, 0.0244, 0.0204, 0.0176, 0.0192, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:38:14,089 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6859.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:20,759 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4934, 1.5541, 1.6607, 0.9249, 1.5685, 1.8641, 1.7719, 1.5315], + device='cuda:0'), covar=tensor([0.1084, 0.0588, 0.0400, 0.0723, 0.0392, 0.0519, 0.0345, 0.0608], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0147, 0.0112, 0.0124, 0.0124, 0.0113, 0.0140, 0.0139], + device='cuda:0'), out_proj_covar=tensor([9.2532e-05, 1.1002e-04, 8.2338e-05, 9.1494e-05, 8.9923e-05, 8.3623e-05, + 1.0451e-04, 1.0305e-04], device='cuda:0') +2023-03-25 23:38:23,766 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6867.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:28,916 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6874.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:31,316 INFO [finetune.py:976] (0/7) Epoch 2, batch 1150, loss[loss=0.2606, simple_loss=0.2995, pruned_loss=0.1109, over 4743.00 frames. ], tot_loss[loss=0.3009, simple_loss=0.3394, pruned_loss=0.1312, over 952981.63 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:38:31,436 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6878.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:38:52,179 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6907.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:53,370 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6909.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:56,974 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 2.014e+02 2.348e+02 2.865e+02 5.036e+02, threshold=4.696e+02, percent-clipped=0.0 +2023-03-25 23:38:57,045 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6915.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:07,707 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6922.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:17,351 INFO [finetune.py:976] (0/7) Epoch 2, batch 1200, loss[loss=0.2945, simple_loss=0.3214, pruned_loss=0.1338, over 4779.00 frames. ], tot_loss[loss=0.2997, simple_loss=0.3381, pruned_loss=0.1307, over 953829.61 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:39:31,062 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6939.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:39:49,900 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:49,967 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:56,178 INFO [finetune.py:976] (0/7) Epoch 2, batch 1250, loss[loss=0.2752, simple_loss=0.3073, pruned_loss=0.1216, over 4905.00 frames. ], tot_loss[loss=0.2941, simple_loss=0.3326, pruned_loss=0.1278, over 955427.97 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:40:00,601 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6985.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:40:04,641 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-03-25 23:40:23,990 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.918e+02 2.412e+02 2.798e+02 4.765e+02, threshold=4.825e+02, percent-clipped=1.0 +2023-03-25 23:40:39,102 INFO [finetune.py:976] (0/7) Epoch 2, batch 1300, loss[loss=0.2982, simple_loss=0.3269, pruned_loss=0.1347, over 4829.00 frames. ], tot_loss[loss=0.2884, simple_loss=0.3274, pruned_loss=0.1247, over 957222.20 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:40:52,828 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7048.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:41:17,562 INFO [finetune.py:976] (0/7) Epoch 2, batch 1350, loss[loss=0.2776, simple_loss=0.3329, pruned_loss=0.1112, over 4835.00 frames. ], tot_loss[loss=0.2861, simple_loss=0.3256, pruned_loss=0.1233, over 957366.57 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:41:41,171 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7109.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 23:41:47,302 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.456e+02 1.876e+02 2.208e+02 2.586e+02 5.614e+02, threshold=4.416e+02, percent-clipped=5.0 +2023-03-25 23:41:49,188 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7118.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:41:52,287 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7123.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:41:55,449 INFO [finetune.py:976] (0/7) Epoch 2, batch 1400, loss[loss=0.3211, simple_loss=0.359, pruned_loss=0.1415, over 4805.00 frames. ], tot_loss[loss=0.2894, simple_loss=0.3292, pruned_loss=0.1248, over 954785.30 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:42:31,278 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7163.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:33,025 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7166.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:36,028 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7171.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:36,684 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9545, 1.3348, 1.0022, 1.7422, 2.0967, 1.5392, 1.6313, 1.8020], + device='cuda:0'), covar=tensor([0.1595, 0.2272, 0.2280, 0.1303, 0.2280, 0.2121, 0.1412, 0.2090], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0097, 0.0116, 0.0093, 0.0124, 0.0096, 0.0099, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-25 23:42:40,187 INFO [finetune.py:976] (0/7) Epoch 2, batch 1450, loss[loss=0.3973, simple_loss=0.4075, pruned_loss=0.1936, over 4207.00 frames. ], tot_loss[loss=0.2925, simple_loss=0.3324, pruned_loss=0.1262, over 955096.69 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:43:29,959 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.933e+02 2.198e+02 2.731e+02 4.077e+02, threshold=4.395e+02, percent-clipped=0.0 +2023-03-25 23:43:40,408 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7224.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:43:42,757 INFO [finetune.py:976] (0/7) Epoch 2, batch 1500, loss[loss=0.3434, simple_loss=0.3903, pruned_loss=0.1483, over 4896.00 frames. ], tot_loss[loss=0.295, simple_loss=0.3347, pruned_loss=0.1277, over 956552.58 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:43:51,988 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7234.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:44:36,526 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7265.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:44:47,928 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7275.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:44:49,673 INFO [finetune.py:976] (0/7) Epoch 2, batch 1550, loss[loss=0.2449, simple_loss=0.2872, pruned_loss=0.1013, over 4708.00 frames. ], tot_loss[loss=0.2955, simple_loss=0.335, pruned_loss=0.128, over 953711.22 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:44:56,178 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4256, 1.2184, 1.2564, 1.4196, 1.4942, 1.3887, 0.6956, 1.2425], + device='cuda:0'), covar=tensor([0.2797, 0.2698, 0.2270, 0.2093, 0.2232, 0.1659, 0.3363, 0.2224], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0201, 0.0187, 0.0173, 0.0222, 0.0168, 0.0203, 0.0176], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:44:57,966 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7285.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:45:31,203 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.380e+02 1.989e+02 2.300e+02 2.720e+02 4.709e+02, threshold=4.600e+02, percent-clipped=4.0 +2023-03-25 23:45:38,832 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.08 vs. limit=5.0 +2023-03-25 23:45:39,140 INFO [finetune.py:976] (0/7) Epoch 2, batch 1600, loss[loss=0.2937, simple_loss=0.3336, pruned_loss=0.1269, over 4260.00 frames. ], tot_loss[loss=0.2943, simple_loss=0.3336, pruned_loss=0.1274, over 953833.49 frames. ], batch size: 66, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:45:42,235 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7333.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:45:44,174 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7336.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:46:35,387 INFO [finetune.py:976] (0/7) Epoch 2, batch 1650, loss[loss=0.2411, simple_loss=0.2802, pruned_loss=0.101, over 4835.00 frames. ], tot_loss[loss=0.2899, simple_loss=0.3291, pruned_loss=0.1254, over 955378.68 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:46:59,569 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7404.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:47:01,406 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7407.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:47:07,156 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.344e+02 1.965e+02 2.306e+02 2.769e+02 7.650e+02, threshold=4.611e+02, percent-clipped=3.0 +2023-03-25 23:47:15,137 INFO [finetune.py:976] (0/7) Epoch 2, batch 1700, loss[loss=0.2976, simple_loss=0.3393, pruned_loss=0.1279, over 4832.00 frames. ], tot_loss[loss=0.2853, simple_loss=0.3249, pruned_loss=0.1228, over 956399.98 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:47:48,714 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7468.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:48:00,184 INFO [finetune.py:976] (0/7) Epoch 2, batch 1750, loss[loss=0.2671, simple_loss=0.3217, pruned_loss=0.1063, over 4944.00 frames. ], tot_loss[loss=0.2899, simple_loss=0.329, pruned_loss=0.1254, over 956009.99 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:48:21,429 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4517, 1.4415, 1.4396, 1.6127, 0.8465, 3.1880, 1.0886, 1.6534], + device='cuda:0'), covar=tensor([0.3545, 0.2398, 0.2141, 0.2196, 0.2308, 0.0199, 0.3061, 0.1623], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0107, 0.0113, 0.0114, 0.0110, 0.0093, 0.0096, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-25 23:48:35,987 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7783, 1.5121, 2.2134, 3.0514, 2.2100, 2.3384, 1.3378, 2.4311], + device='cuda:0'), covar=tensor([0.1562, 0.1495, 0.1080, 0.0465, 0.0702, 0.1960, 0.1516, 0.0659], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0119, 0.0138, 0.0160, 0.0104, 0.0144, 0.0129, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-25 23:48:46,572 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.280e+02 2.019e+02 2.351e+02 2.794e+02 5.482e+02, threshold=4.701e+02, percent-clipped=1.0 +2023-03-25 23:48:51,438 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-25 23:48:53,182 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7519.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:49:03,493 INFO [finetune.py:976] (0/7) Epoch 2, batch 1800, loss[loss=0.352, simple_loss=0.3866, pruned_loss=0.1587, over 4866.00 frames. ], tot_loss[loss=0.295, simple_loss=0.3342, pruned_loss=0.1279, over 954653.81 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:49:03,624 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5650, 1.3586, 1.4370, 1.6010, 2.1114, 1.5403, 1.1518, 1.3122], + device='cuda:0'), covar=tensor([0.3213, 0.3251, 0.2576, 0.2596, 0.2753, 0.1720, 0.3802, 0.2529], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0202, 0.0188, 0.0174, 0.0224, 0.0169, 0.0205, 0.0178], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:49:12,368 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7534.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:49:34,903 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-25 23:49:35,855 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4732, 1.2070, 1.2669, 1.1236, 1.6406, 1.2708, 1.6218, 1.3790], + device='cuda:0'), covar=tensor([0.2837, 0.5497, 0.5715, 0.5323, 0.3799, 0.2911, 0.3636, 0.3969], + device='cuda:0'), in_proj_covar=tensor([0.0160, 0.0191, 0.0234, 0.0246, 0.0208, 0.0178, 0.0196, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:49:44,260 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7565.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:49:57,759 INFO [finetune.py:976] (0/7) Epoch 2, batch 1850, loss[loss=0.2676, simple_loss=0.3034, pruned_loss=0.1159, over 4837.00 frames. ], tot_loss[loss=0.2948, simple_loss=0.3345, pruned_loss=0.1275, over 957028.68 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:49:58,530 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4162, 1.2041, 1.1843, 1.0168, 1.5874, 1.1648, 1.5715, 1.3260], + device='cuda:0'), covar=tensor([0.3015, 0.5649, 0.5968, 0.5905, 0.3886, 0.3098, 0.4404, 0.4102], + device='cuda:0'), in_proj_covar=tensor([0.0161, 0.0192, 0.0234, 0.0247, 0.0208, 0.0179, 0.0197, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:50:05,101 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7582.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:50:41,637 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7613.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:50:48,710 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.368e+02 2.140e+02 2.552e+02 3.133e+02 4.516e+02, threshold=5.105e+02, percent-clipped=0.0 +2023-03-25 23:51:00,612 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7624.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:02,925 INFO [finetune.py:976] (0/7) Epoch 2, batch 1900, loss[loss=0.2979, simple_loss=0.3465, pruned_loss=0.1246, over 4883.00 frames. ], tot_loss[loss=0.294, simple_loss=0.3347, pruned_loss=0.1267, over 956025.64 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:51:09,860 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7631.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:41,826 INFO [finetune.py:976] (0/7) Epoch 2, batch 1950, loss[loss=0.2871, simple_loss=0.3281, pruned_loss=0.123, over 4818.00 frames. ], tot_loss[loss=0.292, simple_loss=0.3325, pruned_loss=0.1257, over 954719.10 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:51:47,781 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7685.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:59,368 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7704.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:52:07,437 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.355e+02 1.951e+02 2.306e+02 2.786e+02 5.176e+02, threshold=4.611e+02, percent-clipped=1.0 +2023-03-25 23:52:17,328 INFO [finetune.py:976] (0/7) Epoch 2, batch 2000, loss[loss=0.2758, simple_loss=0.3118, pruned_loss=0.1199, over 4826.00 frames. ], tot_loss[loss=0.2873, simple_loss=0.3279, pruned_loss=0.1233, over 957508.40 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:52:23,480 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3456, 2.1180, 2.5775, 1.5979, 2.5258, 2.4337, 2.1022, 2.5999], + device='cuda:0'), covar=tensor([0.2153, 0.2426, 0.2486, 0.3376, 0.1274, 0.2450, 0.2723, 0.1370], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0205, 0.0203, 0.0194, 0.0176, 0.0222, 0.0212, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:52:26,490 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7743.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:52:31,889 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7752.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:52:39,938 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7763.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:52:44,703 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9109, 1.7355, 1.8542, 1.8532, 2.2907, 1.8424, 1.6811, 1.6082], + device='cuda:0'), covar=tensor([0.2369, 0.2461, 0.1925, 0.1758, 0.2443, 0.1445, 0.3033, 0.1876], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0202, 0.0188, 0.0173, 0.0223, 0.0168, 0.0204, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:52:45,385 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-25 23:52:49,965 INFO [finetune.py:976] (0/7) Epoch 2, batch 2050, loss[loss=0.2209, simple_loss=0.2789, pruned_loss=0.08149, over 4885.00 frames. ], tot_loss[loss=0.2833, simple_loss=0.3238, pruned_loss=0.1214, over 955832.27 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:53:05,250 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7802.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:53:06,494 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7804.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:53:14,631 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.310e+02 1.893e+02 2.255e+02 2.795e+02 6.508e+02, threshold=4.510e+02, percent-clipped=3.0 +2023-03-25 23:53:17,155 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7819.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 23:53:26,209 INFO [finetune.py:976] (0/7) Epoch 2, batch 2100, loss[loss=0.3368, simple_loss=0.361, pruned_loss=0.1563, over 4717.00 frames. ], tot_loss[loss=0.2827, simple_loss=0.3232, pruned_loss=0.1211, over 954433.46 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:53:38,850 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8623, 4.6405, 4.4875, 2.6800, 4.7816, 3.5843, 0.8023, 3.1686], + device='cuda:0'), covar=tensor([0.2736, 0.1805, 0.1450, 0.3072, 0.0665, 0.0834, 0.5079, 0.1539], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0165, 0.0166, 0.0129, 0.0155, 0.0119, 0.0148, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-25 23:53:56,791 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7863.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:54:00,158 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7867.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:54:01,447 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5757, 2.2102, 2.2514, 1.4067, 2.3702, 2.0807, 1.7398, 2.1270], + device='cuda:0'), covar=tensor([0.1175, 0.1054, 0.1808, 0.2483, 0.1898, 0.2127, 0.2095, 0.1386], + device='cuda:0'), in_proj_covar=tensor([0.0162, 0.0186, 0.0198, 0.0181, 0.0207, 0.0203, 0.0208, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:54:08,357 INFO [finetune.py:976] (0/7) Epoch 2, batch 2150, loss[loss=0.3077, simple_loss=0.3546, pruned_loss=0.1304, over 4902.00 frames. ], tot_loss[loss=0.2848, simple_loss=0.3259, pruned_loss=0.1218, over 954022.39 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:54:17,034 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8350, 3.3886, 3.4977, 3.7647, 3.5767, 3.3943, 3.9153, 1.3760], + device='cuda:0'), covar=tensor([0.0795, 0.0773, 0.0736, 0.0839, 0.1246, 0.1329, 0.0787, 0.4515], + device='cuda:0'), in_proj_covar=tensor([0.0373, 0.0249, 0.0275, 0.0299, 0.0350, 0.0292, 0.0317, 0.0308], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:54:25,867 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8221, 1.5232, 2.1005, 1.4268, 2.0175, 1.9983, 1.6132, 2.1650], + device='cuda:0'), covar=tensor([0.1753, 0.2634, 0.1865, 0.2646, 0.1164, 0.2042, 0.2802, 0.1254], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0204, 0.0201, 0.0193, 0.0175, 0.0220, 0.0211, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:54:49,703 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.983e+02 2.392e+02 2.856e+02 4.131e+02, threshold=4.785e+02, percent-clipped=0.0 +2023-03-25 23:55:10,157 INFO [finetune.py:976] (0/7) Epoch 2, batch 2200, loss[loss=0.3104, simple_loss=0.3429, pruned_loss=0.139, over 4761.00 frames. ], tot_loss[loss=0.2877, simple_loss=0.3287, pruned_loss=0.1234, over 953650.75 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:55:12,775 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7931.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:56:12,552 INFO [finetune.py:976] (0/7) Epoch 2, batch 2250, loss[loss=0.3016, simple_loss=0.358, pruned_loss=0.1225, over 4822.00 frames. ], tot_loss[loss=0.2865, simple_loss=0.3281, pruned_loss=0.1224, over 952772.93 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:56:13,675 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7979.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:56:13,734 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1698, 1.9039, 1.5802, 0.7691, 1.7301, 1.6902, 1.3816, 1.7410], + device='cuda:0'), covar=tensor([0.0876, 0.1155, 0.2009, 0.2603, 0.1546, 0.2854, 0.2783, 0.1348], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0188, 0.0200, 0.0184, 0.0209, 0.0206, 0.0210, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-25 23:56:14,297 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7980.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:56:44,211 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-8000.pt +2023-03-25 23:57:05,131 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.321e+02 1.899e+02 2.269e+02 2.653e+02 4.132e+02, threshold=4.538e+02, percent-clipped=0.0 +2023-03-25 23:57:24,777 INFO [finetune.py:976] (0/7) Epoch 2, batch 2300, loss[loss=0.3132, simple_loss=0.3563, pruned_loss=0.1351, over 4786.00 frames. ], tot_loss[loss=0.2857, simple_loss=0.3283, pruned_loss=0.1215, over 953636.61 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:57:53,413 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8063.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:58:05,649 INFO [finetune.py:976] (0/7) Epoch 2, batch 2350, loss[loss=0.2314, simple_loss=0.2919, pruned_loss=0.0854, over 4764.00 frames. ], tot_loss[loss=0.2813, simple_loss=0.3243, pruned_loss=0.1192, over 954713.67 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:58:20,618 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8099.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:58:39,037 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8111.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:58:41,409 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.369e+02 1.949e+02 2.211e+02 2.649e+02 5.737e+02, threshold=4.422e+02, percent-clipped=2.0 +2023-03-25 23:59:01,109 INFO [finetune.py:976] (0/7) Epoch 2, batch 2400, loss[loss=0.2821, simple_loss=0.3243, pruned_loss=0.1199, over 4791.00 frames. ], tot_loss[loss=0.2773, simple_loss=0.3202, pruned_loss=0.1172, over 955476.89 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-03-25 23:59:27,534 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8158.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:59:39,736 INFO [finetune.py:976] (0/7) Epoch 2, batch 2450, loss[loss=0.2584, simple_loss=0.309, pruned_loss=0.1039, over 4930.00 frames. ], tot_loss[loss=0.2742, simple_loss=0.317, pruned_loss=0.1157, over 956015.94 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:00:11,268 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-26 00:00:20,950 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.358e+02 1.915e+02 2.165e+02 2.652e+02 4.234e+02, threshold=4.330e+02, percent-clipped=0.0 +2023-03-26 00:00:34,016 INFO [finetune.py:976] (0/7) Epoch 2, batch 2500, loss[loss=0.2741, simple_loss=0.3313, pruned_loss=0.1084, over 4737.00 frames. ], tot_loss[loss=0.2777, simple_loss=0.32, pruned_loss=0.1177, over 955710.71 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:01:11,702 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:01:30,123 INFO [finetune.py:976] (0/7) Epoch 2, batch 2550, loss[loss=0.2705, simple_loss=0.3297, pruned_loss=0.1056, over 4817.00 frames. ], tot_loss[loss=0.2818, simple_loss=0.3245, pruned_loss=0.1196, over 955030.80 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:01:31,459 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8280.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:02:02,894 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.611e+02 2.134e+02 2.551e+02 3.097e+02 6.135e+02, threshold=5.102e+02, percent-clipped=4.0 +2023-03-26 00:02:03,743 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 00:02:09,752 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:02:10,867 INFO [finetune.py:976] (0/7) Epoch 2, batch 2600, loss[loss=0.2797, simple_loss=0.3262, pruned_loss=0.1166, over 4815.00 frames. ], tot_loss[loss=0.2831, simple_loss=0.326, pruned_loss=0.1201, over 954967.28 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:02:10,927 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8328.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:03:09,438 INFO [finetune.py:976] (0/7) Epoch 2, batch 2650, loss[loss=0.3053, simple_loss=0.3535, pruned_loss=0.1286, over 4899.00 frames. ], tot_loss[loss=0.2867, simple_loss=0.3289, pruned_loss=0.1223, over 952842.18 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:03:28,496 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:03:30,338 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1370, 1.1597, 1.3962, 1.0312, 1.0302, 1.2178, 1.1279, 1.4123], + device='cuda:0'), covar=tensor([0.1876, 0.2451, 0.1575, 0.1657, 0.1625, 0.1614, 0.3125, 0.1278], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0205, 0.0203, 0.0193, 0.0176, 0.0222, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:03:45,676 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.975e+02 2.270e+02 2.796e+02 5.066e+02, threshold=4.539e+02, percent-clipped=0.0 +2023-03-26 00:03:57,447 INFO [finetune.py:976] (0/7) Epoch 2, batch 2700, loss[loss=0.2851, simple_loss=0.3353, pruned_loss=0.1175, over 4854.00 frames. ], tot_loss[loss=0.2855, simple_loss=0.3277, pruned_loss=0.1216, over 954168.74 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:04:17,365 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:04:28,866 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:04:47,925 INFO [finetune.py:976] (0/7) Epoch 2, batch 2750, loss[loss=0.2652, simple_loss=0.3088, pruned_loss=0.1108, over 4884.00 frames. ], tot_loss[loss=0.2829, simple_loss=0.3242, pruned_loss=0.1208, over 954965.86 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:05:09,616 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:05:25,225 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.975e+02 2.311e+02 2.901e+02 5.278e+02, threshold=4.621e+02, percent-clipped=1.0 +2023-03-26 00:05:38,378 INFO [finetune.py:976] (0/7) Epoch 2, batch 2800, loss[loss=0.2301, simple_loss=0.2743, pruned_loss=0.09298, over 4740.00 frames. ], tot_loss[loss=0.2778, simple_loss=0.3189, pruned_loss=0.1183, over 956353.11 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:06:20,364 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7747, 1.6226, 2.0070, 1.3851, 1.7526, 1.9263, 1.5338, 2.0651], + device='cuda:0'), covar=tensor([0.1649, 0.2206, 0.1458, 0.2132, 0.1188, 0.1693, 0.2578, 0.1023], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0206, 0.0204, 0.0195, 0.0177, 0.0223, 0.0213, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:06:44,210 INFO [finetune.py:976] (0/7) Epoch 2, batch 2850, loss[loss=0.3629, simple_loss=0.3821, pruned_loss=0.1718, over 4813.00 frames. ], tot_loss[loss=0.2745, simple_loss=0.3162, pruned_loss=0.1163, over 955782.18 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:07:08,737 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.882e+02 2.289e+02 2.777e+02 4.779e+02, threshold=4.578e+02, percent-clipped=1.0 +2023-03-26 00:07:13,341 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8621.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:07:18,065 INFO [finetune.py:976] (0/7) Epoch 2, batch 2900, loss[loss=0.3078, simple_loss=0.3559, pruned_loss=0.1299, over 4905.00 frames. ], tot_loss[loss=0.277, simple_loss=0.3189, pruned_loss=0.1176, over 954908.32 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:07:31,490 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3666, 1.2653, 1.7413, 2.3875, 1.6824, 1.9736, 1.0145, 1.9473], + device='cuda:0'), covar=tensor([0.1885, 0.1577, 0.1082, 0.0628, 0.0885, 0.1265, 0.1516, 0.0758], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0120, 0.0140, 0.0163, 0.0105, 0.0146, 0.0133, 0.0109], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:07:37,452 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 00:07:51,241 INFO [finetune.py:976] (0/7) Epoch 2, batch 2950, loss[loss=0.3313, simple_loss=0.3342, pruned_loss=0.1643, over 4573.00 frames. ], tot_loss[loss=0.2807, simple_loss=0.3226, pruned_loss=0.1194, over 954217.65 frames. ], batch size: 20, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:08:17,805 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.871e+02 2.334e+02 2.881e+02 5.890e+02, threshold=4.669e+02, percent-clipped=2.0 +2023-03-26 00:08:27,821 INFO [finetune.py:976] (0/7) Epoch 2, batch 3000, loss[loss=0.3103, simple_loss=0.3512, pruned_loss=0.1347, over 4721.00 frames. ], tot_loss[loss=0.2841, simple_loss=0.326, pruned_loss=0.1211, over 955440.92 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:08:27,823 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 00:08:38,564 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8052, 3.3375, 3.4377, 3.6790, 3.5196, 3.3365, 3.8593, 1.4683], + device='cuda:0'), covar=tensor([0.0901, 0.0811, 0.0903, 0.1074, 0.1421, 0.1489, 0.0716, 0.4835], + device='cuda:0'), in_proj_covar=tensor([0.0372, 0.0248, 0.0276, 0.0300, 0.0349, 0.0291, 0.0316, 0.0306], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:08:43,567 INFO [finetune.py:1010] (0/7) Epoch 2, validation: loss=0.1956, simple_loss=0.2636, pruned_loss=0.06384, over 2265189.00 frames. +2023-03-26 00:08:43,568 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6193MB +2023-03-26 00:09:09,159 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:09:26,209 INFO [finetune.py:976] (0/7) Epoch 2, batch 3050, loss[loss=0.2958, simple_loss=0.3334, pruned_loss=0.1291, over 4745.00 frames. ], tot_loss[loss=0.2854, simple_loss=0.3277, pruned_loss=0.1216, over 954207.51 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:10:07,398 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.935e+02 2.296e+02 2.584e+02 4.666e+02, threshold=4.592e+02, percent-clipped=0.0 +2023-03-26 00:10:07,678 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 00:10:11,723 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:10:16,913 INFO [finetune.py:976] (0/7) Epoch 2, batch 3100, loss[loss=0.2147, simple_loss=0.2702, pruned_loss=0.07964, over 4802.00 frames. ], tot_loss[loss=0.2799, simple_loss=0.3229, pruned_loss=0.1185, over 953250.08 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:10:27,560 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3892, 1.8661, 2.6204, 4.0618, 2.9299, 2.8104, 0.6822, 3.3925], + device='cuda:0'), covar=tensor([0.1640, 0.1508, 0.1292, 0.0447, 0.0791, 0.1325, 0.2261, 0.0584], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0120, 0.0138, 0.0162, 0.0104, 0.0145, 0.0132, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:10:46,263 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8288, 1.6095, 2.0904, 3.2680, 2.4125, 2.4153, 0.9114, 2.5936], + device='cuda:0'), covar=tensor([0.1613, 0.1425, 0.1240, 0.0483, 0.0684, 0.1412, 0.1992, 0.0578], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0120, 0.0138, 0.0162, 0.0104, 0.0145, 0.0132, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:10:59,136 INFO [finetune.py:976] (0/7) Epoch 2, batch 3150, loss[loss=0.2454, simple_loss=0.2972, pruned_loss=0.09678, over 4917.00 frames. ], tot_loss[loss=0.278, simple_loss=0.3201, pruned_loss=0.1179, over 954781.69 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:11:28,139 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.838e+02 2.165e+02 2.835e+02 5.909e+02, threshold=4.329e+02, percent-clipped=1.0 +2023-03-26 00:11:33,401 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8921.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:11:43,241 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8927.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:11:43,730 INFO [finetune.py:976] (0/7) Epoch 2, batch 3200, loss[loss=0.3015, simple_loss=0.3308, pruned_loss=0.1361, over 4823.00 frames. ], tot_loss[loss=0.2725, simple_loss=0.3151, pruned_loss=0.115, over 956606.27 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:11:59,076 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9498, 1.1490, 0.9531, 1.6444, 2.2140, 1.1437, 1.3521, 1.7105], + device='cuda:0'), covar=tensor([0.1535, 0.2225, 0.2224, 0.1268, 0.1940, 0.2141, 0.1475, 0.1936], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0097, 0.0115, 0.0092, 0.0124, 0.0096, 0.0098, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 00:12:09,909 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6216, 3.9850, 4.1336, 4.4459, 4.2873, 4.0425, 4.6876, 1.4091], + device='cuda:0'), covar=tensor([0.0650, 0.0724, 0.0713, 0.0864, 0.1205, 0.1203, 0.0540, 0.5279], + device='cuda:0'), in_proj_covar=tensor([0.0370, 0.0246, 0.0274, 0.0298, 0.0346, 0.0290, 0.0314, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:12:25,335 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:12:30,759 INFO [finetune.py:976] (0/7) Epoch 2, batch 3250, loss[loss=0.3431, simple_loss=0.3673, pruned_loss=0.1594, over 4747.00 frames. ], tot_loss[loss=0.2731, simple_loss=0.3155, pruned_loss=0.1153, over 956727.33 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:12:36,302 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:12:37,536 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:00,580 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9013.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:03,774 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.894e+02 2.281e+02 2.922e+02 4.541e+02, threshold=4.561e+02, percent-clipped=3.0 +2023-03-26 00:13:11,706 INFO [finetune.py:976] (0/7) Epoch 2, batch 3300, loss[loss=0.2859, simple_loss=0.342, pruned_loss=0.1149, over 4933.00 frames. ], tot_loss[loss=0.2779, simple_loss=0.3209, pruned_loss=0.1175, over 957271.30 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:13:25,343 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9047.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:42,680 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:44,937 INFO [finetune.py:976] (0/7) Epoch 2, batch 3350, loss[loss=0.2949, simple_loss=0.3431, pruned_loss=0.1234, over 4747.00 frames. ], tot_loss[loss=0.2803, simple_loss=0.3237, pruned_loss=0.1184, over 955280.62 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:13:45,634 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4540, 3.8397, 4.0058, 4.3088, 4.1633, 3.8773, 4.5040, 1.4657], + device='cuda:0'), covar=tensor([0.0641, 0.0732, 0.0742, 0.0726, 0.1075, 0.1281, 0.0621, 0.4875], + device='cuda:0'), in_proj_covar=tensor([0.0371, 0.0247, 0.0275, 0.0299, 0.0348, 0.0290, 0.0314, 0.0306], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:13:51,625 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6118, 1.6510, 1.4662, 1.6084, 0.9821, 3.5250, 1.3732, 2.0138], + device='cuda:0'), covar=tensor([0.3701, 0.2374, 0.2303, 0.2482, 0.2244, 0.0180, 0.2982, 0.1482], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0108, 0.0113, 0.0115, 0.0111, 0.0094, 0.0098, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 00:13:59,869 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9034, 2.0997, 1.9070, 1.5165, 2.3613, 2.2339, 2.0526, 1.8379], + device='cuda:0'), covar=tensor([0.0636, 0.0454, 0.0688, 0.0854, 0.0362, 0.0539, 0.0607, 0.0949], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0132, 0.0142, 0.0130, 0.0108, 0.0139, 0.0146, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:14:17,308 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 00:14:20,188 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.430e+02 1.942e+02 2.307e+02 2.996e+02 6.023e+02, threshold=4.614e+02, percent-clipped=2.0 +2023-03-26 00:14:20,874 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:14:28,092 INFO [finetune.py:976] (0/7) Epoch 2, batch 3400, loss[loss=0.297, simple_loss=0.3387, pruned_loss=0.1276, over 4914.00 frames. ], tot_loss[loss=0.2822, simple_loss=0.326, pruned_loss=0.1192, over 956650.65 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:15:12,188 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6530, 1.4691, 1.6011, 1.7696, 1.8584, 1.7425, 1.0288, 1.4792], + device='cuda:0'), covar=tensor([0.2434, 0.2496, 0.2004, 0.1924, 0.2048, 0.1232, 0.3172, 0.2094], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0204, 0.0190, 0.0176, 0.0227, 0.0169, 0.0207, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:15:23,468 INFO [finetune.py:976] (0/7) Epoch 2, batch 3450, loss[loss=0.3215, simple_loss=0.3512, pruned_loss=0.1459, over 4857.00 frames. ], tot_loss[loss=0.2803, simple_loss=0.3247, pruned_loss=0.118, over 956241.44 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:15:59,478 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.266e+02 1.993e+02 2.349e+02 2.850e+02 4.291e+02, threshold=4.698e+02, percent-clipped=0.0 +2023-03-26 00:16:12,592 INFO [finetune.py:976] (0/7) Epoch 2, batch 3500, loss[loss=0.1823, simple_loss=0.2341, pruned_loss=0.06521, over 4756.00 frames. ], tot_loss[loss=0.2767, simple_loss=0.3209, pruned_loss=0.1163, over 955650.20 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:17:13,577 INFO [finetune.py:976] (0/7) Epoch 2, batch 3550, loss[loss=0.2045, simple_loss=0.2638, pruned_loss=0.07257, over 4783.00 frames. ], tot_loss[loss=0.2725, simple_loss=0.3171, pruned_loss=0.114, over 956097.64 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:17:16,742 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9283.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:17:53,394 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.769e+02 2.188e+02 2.766e+02 5.069e+02, threshold=4.376e+02, percent-clipped=2.0 +2023-03-26 00:18:09,362 INFO [finetune.py:976] (0/7) Epoch 2, batch 3600, loss[loss=0.2901, simple_loss=0.3241, pruned_loss=0.1281, over 4823.00 frames. ], tot_loss[loss=0.2701, simple_loss=0.3138, pruned_loss=0.1132, over 953700.61 frames. ], batch size: 41, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:18:22,989 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9342.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:18:45,304 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9369.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:18:51,740 INFO [finetune.py:976] (0/7) Epoch 2, batch 3650, loss[loss=0.2704, simple_loss=0.3346, pruned_loss=0.1031, over 4855.00 frames. ], tot_loss[loss=0.2706, simple_loss=0.315, pruned_loss=0.1132, over 951111.02 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:19:24,498 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9414.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:19:24,980 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 1.902e+02 2.269e+02 2.850e+02 5.426e+02, threshold=4.539e+02, percent-clipped=4.0 +2023-03-26 00:19:31,403 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9416.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:19:45,283 INFO [finetune.py:976] (0/7) Epoch 2, batch 3700, loss[loss=0.3124, simple_loss=0.3612, pruned_loss=0.1318, over 4822.00 frames. ], tot_loss[loss=0.2729, simple_loss=0.3185, pruned_loss=0.1136, over 952956.20 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:20:15,355 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:20:17,102 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3800, 1.4525, 1.2561, 1.5836, 1.6235, 3.0247, 1.3612, 1.6089], + device='cuda:0'), covar=tensor([0.1077, 0.1784, 0.1237, 0.1072, 0.1632, 0.0268, 0.1491, 0.1721], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0080, 0.0076, 0.0079, 0.0092, 0.0081, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 00:20:21,638 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5662, 3.8453, 3.7772, 1.9871, 3.9967, 2.9077, 0.7288, 2.5790], + device='cuda:0'), covar=tensor([0.2583, 0.1497, 0.1403, 0.3097, 0.0889, 0.0948, 0.4554, 0.1487], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0166, 0.0165, 0.0128, 0.0156, 0.0120, 0.0147, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 00:20:23,958 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:20:26,041 INFO [finetune.py:976] (0/7) Epoch 2, batch 3750, loss[loss=0.3231, simple_loss=0.3566, pruned_loss=0.1448, over 4834.00 frames. ], tot_loss[loss=0.2752, simple_loss=0.3204, pruned_loss=0.115, over 952571.52 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:20:40,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8816, 1.4485, 1.5852, 1.5509, 1.4142, 1.4526, 1.4815, 1.5972], + device='cuda:0'), covar=tensor([1.5513, 2.5711, 1.7815, 2.3286, 2.4632, 1.7482, 3.1083, 1.6188], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0258, 0.0246, 0.0270, 0.0246, 0.0219, 0.0280, 0.0217], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 00:20:40,943 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 00:20:55,390 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.357e+02 1.893e+02 2.405e+02 2.686e+02 6.929e+02, threshold=4.810e+02, percent-clipped=1.0 +2023-03-26 00:21:10,672 INFO [finetune.py:976] (0/7) Epoch 2, batch 3800, loss[loss=0.2719, simple_loss=0.3167, pruned_loss=0.1135, over 4784.00 frames. ], tot_loss[loss=0.276, simple_loss=0.3216, pruned_loss=0.1152, over 952293.23 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:21:54,265 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 00:22:03,070 INFO [finetune.py:976] (0/7) Epoch 2, batch 3850, loss[loss=0.2544, simple_loss=0.2968, pruned_loss=0.1061, over 4819.00 frames. ], tot_loss[loss=0.273, simple_loss=0.3186, pruned_loss=0.1137, over 953818.35 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:22:07,241 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9583.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:22:07,865 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:22:07,888 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4406, 1.2612, 1.1942, 1.1993, 1.5801, 1.2874, 1.6430, 1.3753], + device='cuda:0'), covar=tensor([0.3098, 0.5218, 0.6137, 0.5332, 0.4141, 0.3004, 0.4174, 0.3896], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0195, 0.0238, 0.0252, 0.0214, 0.0182, 0.0203, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:22:33,609 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.890e+02 2.331e+02 2.867e+02 5.576e+02, threshold=4.662e+02, percent-clipped=3.0 +2023-03-26 00:22:41,215 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-26 00:22:48,506 INFO [finetune.py:976] (0/7) Epoch 2, batch 3900, loss[loss=0.3108, simple_loss=0.3369, pruned_loss=0.1424, over 4852.00 frames. ], tot_loss[loss=0.2688, simple_loss=0.3143, pruned_loss=0.1117, over 954812.97 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:22:49,807 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1690, 1.9044, 2.2320, 0.9640, 2.2395, 2.5637, 2.0285, 2.0779], + device='cuda:0'), covar=tensor([0.1055, 0.1046, 0.0406, 0.0978, 0.0549, 0.0708, 0.0510, 0.0738], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0154, 0.0116, 0.0132, 0.0130, 0.0117, 0.0143, 0.0143], + device='cuda:0'), out_proj_covar=tensor([9.5643e-05, 1.1454e-04, 8.5167e-05, 9.7042e-05, 9.4134e-05, 8.6115e-05, + 1.0713e-04, 1.0573e-04], device='cuda:0') +2023-03-26 00:22:50,356 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9631.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:08,236 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9642.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:10,105 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9645.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:28,457 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:40,492 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:42,503 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 00:23:46,465 INFO [finetune.py:976] (0/7) Epoch 2, batch 3950, loss[loss=0.2525, simple_loss=0.3027, pruned_loss=0.1011, over 4912.00 frames. ], tot_loss[loss=0.2654, simple_loss=0.3107, pruned_loss=0.1101, over 955143.57 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:24:00,115 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9690.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:18,675 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-03-26 00:24:28,618 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.967e+02 2.302e+02 2.784e+02 7.100e+02, threshold=4.604e+02, percent-clipped=2.0 +2023-03-26 00:24:29,311 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9717.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:30,641 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9719.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:34,280 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.92 vs. limit=5.0 +2023-03-26 00:24:36,431 INFO [finetune.py:976] (0/7) Epoch 2, batch 4000, loss[loss=0.2966, simple_loss=0.3446, pruned_loss=0.1243, over 4817.00 frames. ], tot_loss[loss=0.2659, simple_loss=0.3105, pruned_loss=0.1107, over 956252.47 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:24:48,474 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:59,836 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7174, 1.5973, 1.1825, 1.7433, 1.6679, 1.4210, 2.0920, 1.6344], + device='cuda:0'), covar=tensor([0.2803, 0.5460, 0.5923, 0.5464, 0.4225, 0.2860, 0.6461, 0.3622], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0195, 0.0238, 0.0252, 0.0215, 0.0182, 0.0204, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:25:10,956 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:25:16,250 INFO [finetune.py:976] (0/7) Epoch 2, batch 4050, loss[loss=0.3158, simple_loss=0.3679, pruned_loss=0.1318, over 4805.00 frames. ], tot_loss[loss=0.2688, simple_loss=0.314, pruned_loss=0.1118, over 956278.77 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:25:19,834 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 00:25:24,722 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-26 00:25:34,019 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3801, 0.5511, 1.2443, 1.0844, 1.0882, 1.0373, 0.9605, 1.1287], + device='cuda:0'), covar=tensor([1.0597, 2.0525, 1.6035, 1.7645, 1.8550, 1.3902, 2.3316, 1.4071], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0256, 0.0246, 0.0269, 0.0245, 0.0218, 0.0278, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 00:25:37,121 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:25:44,425 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 1.934e+02 2.184e+02 2.585e+02 5.351e+02, threshold=4.368e+02, percent-clipped=2.0 +2023-03-26 00:25:57,343 INFO [finetune.py:976] (0/7) Epoch 2, batch 4100, loss[loss=0.2712, simple_loss=0.331, pruned_loss=0.1057, over 4909.00 frames. ], tot_loss[loss=0.2735, simple_loss=0.3193, pruned_loss=0.1139, over 956546.69 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:26:28,849 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0369, 1.3231, 0.9539, 1.8202, 2.3054, 1.4715, 1.7272, 1.9346], + device='cuda:0'), covar=tensor([0.1524, 0.2262, 0.2435, 0.1291, 0.1917, 0.2005, 0.1450, 0.2133], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0098, 0.0118, 0.0094, 0.0126, 0.0098, 0.0100, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 00:26:30,072 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4519, 1.4751, 1.6681, 0.9092, 1.4313, 1.8204, 1.6749, 1.4925], + device='cuda:0'), covar=tensor([0.1051, 0.0655, 0.0414, 0.0700, 0.0432, 0.0405, 0.0379, 0.0539], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0116, 0.0132, 0.0129, 0.0116, 0.0143, 0.0142], + device='cuda:0'), out_proj_covar=tensor([9.5709e-05, 1.1427e-04, 8.5035e-05, 9.6858e-05, 9.3964e-05, 8.5720e-05, + 1.0693e-04, 1.0564e-04], device='cuda:0') +2023-03-26 00:26:34,083 INFO [finetune.py:976] (0/7) Epoch 2, batch 4150, loss[loss=0.2617, simple_loss=0.2939, pruned_loss=0.1147, over 4077.00 frames. ], tot_loss[loss=0.2742, simple_loss=0.32, pruned_loss=0.1142, over 956160.25 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:27:05,051 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.789e+02 2.151e+02 2.605e+02 5.306e+02, threshold=4.302e+02, percent-clipped=2.0 +2023-03-26 00:27:17,337 INFO [finetune.py:976] (0/7) Epoch 2, batch 4200, loss[loss=0.2738, simple_loss=0.3179, pruned_loss=0.1149, over 4751.00 frames. ], tot_loss[loss=0.2728, simple_loss=0.3195, pruned_loss=0.113, over 956916.13 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:27:25,208 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:28:05,867 INFO [finetune.py:976] (0/7) Epoch 2, batch 4250, loss[loss=0.3409, simple_loss=0.3504, pruned_loss=0.1656, over 4238.00 frames. ], tot_loss[loss=0.2702, simple_loss=0.3167, pruned_loss=0.1118, over 956423.16 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:28:06,604 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1288, 1.7250, 2.4355, 1.6294, 2.2036, 2.2978, 1.8524, 2.5420], + device='cuda:0'), covar=tensor([0.1547, 0.2221, 0.1577, 0.2226, 0.1130, 0.1492, 0.2603, 0.0936], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0208, 0.0207, 0.0198, 0.0180, 0.0226, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:28:31,817 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-10000.pt +2023-03-26 00:28:41,979 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10014.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:28:47,608 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.798e+02 2.175e+02 2.768e+02 4.624e+02, threshold=4.351e+02, percent-clipped=3.0 +2023-03-26 00:28:50,809 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3971, 1.3826, 1.4792, 0.7957, 1.6252, 1.5170, 1.4495, 1.3249], + device='cuda:0'), covar=tensor([0.0741, 0.0821, 0.0751, 0.1109, 0.0641, 0.0838, 0.0786, 0.1367], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0131, 0.0142, 0.0129, 0.0108, 0.0139, 0.0145, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:29:00,128 INFO [finetune.py:976] (0/7) Epoch 2, batch 4300, loss[loss=0.2635, simple_loss=0.2972, pruned_loss=0.1149, over 4222.00 frames. ], tot_loss[loss=0.2687, simple_loss=0.3144, pruned_loss=0.1114, over 957651.45 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:29:02,275 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 00:29:12,696 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10039.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:29:36,678 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10062.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:29:46,290 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10070.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:29:52,923 INFO [finetune.py:976] (0/7) Epoch 2, batch 4350, loss[loss=0.2716, simple_loss=0.308, pruned_loss=0.1176, over 4729.00 frames. ], tot_loss[loss=0.2654, simple_loss=0.3109, pruned_loss=0.11, over 957227.97 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:29:58,528 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-03-26 00:29:58,591 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 00:29:59,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2910, 2.0366, 1.7795, 2.3181, 2.4099, 1.9625, 2.6409, 2.1271], + device='cuda:0'), covar=tensor([0.2247, 0.5043, 0.5115, 0.4586, 0.3079, 0.2344, 0.3995, 0.3561], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0195, 0.0237, 0.0252, 0.0214, 0.0181, 0.0204, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:30:06,136 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10099.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:30:07,342 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10100.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:30:25,353 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 1.930e+02 2.275e+02 2.717e+02 4.483e+02, threshold=4.550e+02, percent-clipped=1.0 +2023-03-26 00:30:27,179 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:30:34,451 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10123.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:30:37,373 INFO [finetune.py:976] (0/7) Epoch 2, batch 4400, loss[loss=0.283, simple_loss=0.3357, pruned_loss=0.1152, over 4906.00 frames. ], tot_loss[loss=0.2673, simple_loss=0.3118, pruned_loss=0.1115, over 954000.23 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:31:35,876 INFO [finetune.py:976] (0/7) Epoch 2, batch 4450, loss[loss=0.2973, simple_loss=0.3449, pruned_loss=0.1249, over 4812.00 frames. ], tot_loss[loss=0.2685, simple_loss=0.3138, pruned_loss=0.1116, over 953936.87 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:32:16,507 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2814, 1.4394, 1.2289, 1.5297, 1.5711, 3.0191, 1.2912, 1.5689], + device='cuda:0'), covar=tensor([0.1070, 0.1636, 0.1197, 0.1074, 0.1597, 0.0307, 0.1391, 0.1718], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0080, 0.0077, 0.0079, 0.0092, 0.0082, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 00:32:18,679 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 1.960e+02 2.424e+02 3.015e+02 7.276e+02, threshold=4.848e+02, percent-clipped=5.0 +2023-03-26 00:32:36,845 INFO [finetune.py:976] (0/7) Epoch 2, batch 4500, loss[loss=0.3065, simple_loss=0.3393, pruned_loss=0.1369, over 4735.00 frames. ], tot_loss[loss=0.2713, simple_loss=0.3163, pruned_loss=0.1132, over 950874.71 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:32:44,226 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:32:46,816 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-26 00:33:20,854 INFO [finetune.py:976] (0/7) Epoch 2, batch 4550, loss[loss=0.2582, simple_loss=0.315, pruned_loss=0.1007, over 4851.00 frames. ], tot_loss[loss=0.2717, simple_loss=0.3173, pruned_loss=0.1131, over 950501.40 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:33:32,439 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10288.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:33:59,118 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10314.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:34:00,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.803e+02 2.127e+02 2.576e+02 5.771e+02, threshold=4.255e+02, percent-clipped=1.0 +2023-03-26 00:34:01,549 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6084, 1.5543, 1.8151, 1.8658, 1.7024, 3.5712, 1.3547, 1.7803], + device='cuda:0'), covar=tensor([0.1013, 0.1755, 0.1122, 0.1088, 0.1590, 0.0232, 0.1477, 0.1731], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0080, 0.0077, 0.0079, 0.0092, 0.0082, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 00:34:14,432 INFO [finetune.py:976] (0/7) Epoch 2, batch 4600, loss[loss=0.2835, simple_loss=0.3123, pruned_loss=0.1274, over 4895.00 frames. ], tot_loss[loss=0.2715, simple_loss=0.3174, pruned_loss=0.1128, over 951771.83 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:34:49,715 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:01,343 INFO [finetune.py:976] (0/7) Epoch 2, batch 4650, loss[loss=0.2546, simple_loss=0.3096, pruned_loss=0.09983, over 4756.00 frames. ], tot_loss[loss=0.2671, simple_loss=0.3131, pruned_loss=0.1106, over 953706.56 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:35:12,308 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10395.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 00:35:14,743 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:21,431 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10409.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:25,532 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.369e+02 1.834e+02 2.117e+02 2.478e+02 4.313e+02, threshold=4.233e+02, percent-clipped=1.0 +2023-03-26 00:35:27,258 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10418.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:35:34,604 INFO [finetune.py:976] (0/7) Epoch 2, batch 4700, loss[loss=0.2207, simple_loss=0.2731, pruned_loss=0.08416, over 4800.00 frames. ], tot_loss[loss=0.2636, simple_loss=0.3092, pruned_loss=0.109, over 952241.95 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:35:34,879 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 00:35:46,700 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:50,996 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:36:01,080 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:36:07,252 INFO [finetune.py:976] (0/7) Epoch 2, batch 4750, loss[loss=0.217, simple_loss=0.2656, pruned_loss=0.08415, over 4721.00 frames. ], tot_loss[loss=0.2614, simple_loss=0.3067, pruned_loss=0.108, over 952127.23 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:36:36,938 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10515.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:36:42,344 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.733e+02 2.215e+02 2.656e+02 7.843e+02, threshold=4.429e+02, percent-clipped=2.0 +2023-03-26 00:36:50,652 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2729, 1.0902, 1.4562, 2.3020, 1.5614, 2.2126, 0.7893, 2.0306], + device='cuda:0'), covar=tensor([0.2148, 0.2322, 0.1549, 0.1070, 0.1188, 0.1268, 0.1985, 0.0828], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0121, 0.0140, 0.0166, 0.0107, 0.0149, 0.0132, 0.0109], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:36:51,176 INFO [finetune.py:976] (0/7) Epoch 2, batch 4800, loss[loss=0.3064, simple_loss=0.3523, pruned_loss=0.1302, over 4815.00 frames. ], tot_loss[loss=0.2635, simple_loss=0.309, pruned_loss=0.109, over 951492.18 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:37:18,485 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-03-26 00:37:26,734 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5949, 1.4330, 1.4395, 1.6721, 1.8946, 1.5824, 0.9708, 1.3561], + device='cuda:0'), covar=tensor([0.2811, 0.2773, 0.2344, 0.2125, 0.2316, 0.1456, 0.3495, 0.2256], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0206, 0.0192, 0.0178, 0.0228, 0.0170, 0.0208, 0.0181], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:37:43,276 INFO [finetune.py:976] (0/7) Epoch 2, batch 4850, loss[loss=0.2488, simple_loss=0.2984, pruned_loss=0.09965, over 4925.00 frames. ], tot_loss[loss=0.2678, simple_loss=0.3139, pruned_loss=0.1108, over 952133.11 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:38:04,562 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-26 00:38:22,410 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.945e+02 2.280e+02 2.839e+02 5.226e+02, threshold=4.560e+02, percent-clipped=1.0 +2023-03-26 00:38:28,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8704, 1.2872, 0.9539, 1.5713, 2.0748, 1.3700, 1.4290, 1.7309], + device='cuda:0'), covar=tensor([0.1530, 0.2225, 0.2341, 0.1357, 0.2170, 0.2202, 0.1546, 0.2014], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0098, 0.0118, 0.0094, 0.0126, 0.0098, 0.0100, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 00:38:33,268 INFO [finetune.py:976] (0/7) Epoch 2, batch 4900, loss[loss=0.3101, simple_loss=0.3413, pruned_loss=0.1395, over 4727.00 frames. ], tot_loss[loss=0.2689, simple_loss=0.3159, pruned_loss=0.1109, over 953747.77 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:39:15,445 INFO [finetune.py:976] (0/7) Epoch 2, batch 4950, loss[loss=0.2868, simple_loss=0.3318, pruned_loss=0.121, over 4898.00 frames. ], tot_loss[loss=0.269, simple_loss=0.3167, pruned_loss=0.1107, over 952489.88 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:39:26,676 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7508, 1.5151, 1.8515, 2.8347, 2.1231, 2.2947, 1.0806, 2.3790], + device='cuda:0'), covar=tensor([0.1768, 0.1578, 0.1324, 0.0726, 0.0890, 0.1246, 0.1770, 0.0729], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0138, 0.0163, 0.0105, 0.0146, 0.0130, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:39:27,271 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10695.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:39:28,480 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2106, 1.7980, 2.4858, 4.0894, 2.9885, 2.7100, 0.8342, 3.3900], + device='cuda:0'), covar=tensor([0.1831, 0.1600, 0.1455, 0.0409, 0.0729, 0.1480, 0.2203, 0.0522], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0138, 0.0163, 0.0105, 0.0146, 0.0130, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:39:40,469 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.382e+02 1.832e+02 2.144e+02 2.563e+02 4.788e+02, threshold=4.289e+02, percent-clipped=1.0 +2023-03-26 00:39:42,227 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10718.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:39:48,191 INFO [finetune.py:976] (0/7) Epoch 2, batch 5000, loss[loss=0.27, simple_loss=0.3067, pruned_loss=0.1166, over 4730.00 frames. ], tot_loss[loss=0.2649, simple_loss=0.3131, pruned_loss=0.1084, over 954431.19 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:39:53,164 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 00:39:59,860 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:00,531 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:13,893 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:14,499 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10766.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:40:19,870 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:25,517 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0190, 1.7542, 2.4035, 1.6442, 2.1630, 2.2789, 1.6692, 2.4565], + device='cuda:0'), covar=tensor([0.1679, 0.2161, 0.1530, 0.2338, 0.1010, 0.1564, 0.2852, 0.0991], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0208, 0.0206, 0.0199, 0.0181, 0.0227, 0.0216, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:40:25,991 INFO [finetune.py:976] (0/7) Epoch 2, batch 5050, loss[loss=0.2636, simple_loss=0.3056, pruned_loss=0.1109, over 4872.00 frames. ], tot_loss[loss=0.2625, simple_loss=0.3098, pruned_loss=0.1076, over 954736.89 frames. ], batch size: 34, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:40:28,684 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-03-26 00:40:39,663 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2437, 2.1148, 1.6207, 2.4122, 2.3249, 1.7451, 2.9531, 2.1440], + device='cuda:0'), covar=tensor([0.2621, 0.5894, 0.5930, 0.5344, 0.3628, 0.2657, 0.4001, 0.3987], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0196, 0.0238, 0.0252, 0.0216, 0.0182, 0.0205, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:40:48,608 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10805.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:52,130 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10810.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:40:54,539 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3339, 2.8866, 2.9966, 3.2487, 3.0880, 2.8928, 3.3786, 1.0425], + device='cuda:0'), covar=tensor([0.1071, 0.0999, 0.1243, 0.1229, 0.1620, 0.1702, 0.1076, 0.4828], + device='cuda:0'), in_proj_covar=tensor([0.0371, 0.0246, 0.0277, 0.0298, 0.0343, 0.0291, 0.0312, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:40:55,640 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.711e+02 2.025e+02 2.497e+02 4.623e+02, threshold=4.049e+02, percent-clipped=1.0 +2023-03-26 00:41:03,403 INFO [finetune.py:976] (0/7) Epoch 2, batch 5100, loss[loss=0.227, simple_loss=0.2785, pruned_loss=0.08775, over 4899.00 frames. ], tot_loss[loss=0.2581, simple_loss=0.3054, pruned_loss=0.1054, over 950453.54 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:41:07,740 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10835.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:41:28,216 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3637, 1.1983, 1.7108, 2.3316, 1.7125, 1.9534, 1.0217, 1.9352], + device='cuda:0'), covar=tensor([0.1775, 0.1761, 0.1135, 0.0727, 0.0915, 0.1389, 0.1549, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0120, 0.0140, 0.0164, 0.0106, 0.0147, 0.0131, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:41:38,259 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7023, 1.2836, 1.4359, 1.3266, 1.7760, 1.8576, 1.6547, 1.2687], + device='cuda:0'), covar=tensor([0.0272, 0.0471, 0.0512, 0.0377, 0.0264, 0.0377, 0.0270, 0.0471], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0111, 0.0132, 0.0112, 0.0101, 0.0097, 0.0086, 0.0106], + device='cuda:0'), out_proj_covar=tensor([6.3017e-05, 8.8023e-05, 1.0653e-04, 8.8477e-05, 8.0517e-05, 7.2124e-05, + 6.6649e-05, 8.3186e-05], device='cuda:0') +2023-03-26 00:41:47,384 INFO [finetune.py:976] (0/7) Epoch 2, batch 5150, loss[loss=0.2538, simple_loss=0.3045, pruned_loss=0.1016, over 4833.00 frames. ], tot_loss[loss=0.2592, simple_loss=0.3057, pruned_loss=0.1063, over 951260.11 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:41:54,240 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10885.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:41:57,730 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5963, 1.3736, 1.4081, 1.6797, 1.8017, 1.5803, 0.9664, 1.2975], + device='cuda:0'), covar=tensor([0.3060, 0.3100, 0.2612, 0.2200, 0.2445, 0.1554, 0.3796, 0.2463], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0207, 0.0193, 0.0178, 0.0229, 0.0170, 0.0210, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:42:25,076 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.747e+02 2.089e+02 2.521e+02 5.475e+02, threshold=4.178e+02, percent-clipped=1.0 +2023-03-26 00:42:37,755 INFO [finetune.py:976] (0/7) Epoch 2, batch 5200, loss[loss=0.2717, simple_loss=0.2897, pruned_loss=0.1269, over 4489.00 frames. ], tot_loss[loss=0.2612, simple_loss=0.3081, pruned_loss=0.1072, over 952137.39 frames. ], batch size: 19, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:43:00,796 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10946.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:43:29,722 INFO [finetune.py:976] (0/7) Epoch 2, batch 5250, loss[loss=0.2529, simple_loss=0.3043, pruned_loss=0.1007, over 4812.00 frames. ], tot_loss[loss=0.2645, simple_loss=0.3118, pruned_loss=0.1086, over 954010.73 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:44:03,163 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 2.014e+02 2.376e+02 2.957e+02 8.531e+02, threshold=4.753e+02, percent-clipped=3.0 +2023-03-26 00:44:10,959 INFO [finetune.py:976] (0/7) Epoch 2, batch 5300, loss[loss=0.2327, simple_loss=0.2908, pruned_loss=0.08733, over 4822.00 frames. ], tot_loss[loss=0.2661, simple_loss=0.3132, pruned_loss=0.1094, over 954169.60 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:44:38,215 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-26 00:44:44,429 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:44:58,486 INFO [finetune.py:976] (0/7) Epoch 2, batch 5350, loss[loss=0.2894, simple_loss=0.3284, pruned_loss=0.1252, over 4206.00 frames. ], tot_loss[loss=0.2654, simple_loss=0.3134, pruned_loss=0.1087, over 954729.82 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:45:15,532 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11100.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:45:23,036 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11110.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:45:25,461 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:45:27,191 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.217e+02 1.830e+02 2.178e+02 2.684e+02 7.389e+02, threshold=4.357e+02, percent-clipped=1.0 +2023-03-26 00:45:27,909 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6190, 3.5840, 3.5286, 1.8955, 3.8263, 2.8054, 0.6790, 2.5649], + device='cuda:0'), covar=tensor([0.2680, 0.1814, 0.1518, 0.3208, 0.0971, 0.1030, 0.4740, 0.1546], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0168, 0.0166, 0.0129, 0.0157, 0.0120, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 00:45:39,916 INFO [finetune.py:976] (0/7) Epoch 2, batch 5400, loss[loss=0.2869, simple_loss=0.3257, pruned_loss=0.124, over 4934.00 frames. ], tot_loss[loss=0.2622, simple_loss=0.3102, pruned_loss=0.1071, over 955911.40 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:45:41,697 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:45:50,208 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-03-26 00:46:05,247 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:46:22,034 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11174.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:46:24,381 INFO [finetune.py:976] (0/7) Epoch 2, batch 5450, loss[loss=0.2749, simple_loss=0.3112, pruned_loss=0.1193, over 4912.00 frames. ], tot_loss[loss=0.2583, simple_loss=0.3062, pruned_loss=0.1051, over 955828.53 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:46:55,233 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.719e+02 1.995e+02 2.396e+02 4.116e+02, threshold=3.991e+02, percent-clipped=0.0 +2023-03-26 00:47:11,561 INFO [finetune.py:976] (0/7) Epoch 2, batch 5500, loss[loss=0.2794, simple_loss=0.3274, pruned_loss=0.1157, over 4909.00 frames. ], tot_loss[loss=0.2558, simple_loss=0.3038, pruned_loss=0.1039, over 956472.65 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:47:21,373 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:47:25,453 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11241.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:48:00,495 INFO [finetune.py:976] (0/7) Epoch 2, batch 5550, loss[loss=0.2359, simple_loss=0.2889, pruned_loss=0.0914, over 4802.00 frames. ], tot_loss[loss=0.2563, simple_loss=0.304, pruned_loss=0.1043, over 954279.12 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:48:05,578 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1426, 2.0271, 1.7466, 1.0016, 1.9101, 1.8247, 1.6190, 1.9179], + device='cuda:0'), covar=tensor([0.0865, 0.0713, 0.1161, 0.1752, 0.1114, 0.1794, 0.1714, 0.0834], + device='cuda:0'), in_proj_covar=tensor([0.0162, 0.0191, 0.0200, 0.0184, 0.0210, 0.0206, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:48:30,779 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.394e+02 1.988e+02 2.263e+02 2.636e+02 5.646e+02, threshold=4.525e+02, percent-clipped=4.0 +2023-03-26 00:48:38,279 INFO [finetune.py:976] (0/7) Epoch 2, batch 5600, loss[loss=0.3133, simple_loss=0.3374, pruned_loss=0.1445, over 4150.00 frames. ], tot_loss[loss=0.2622, simple_loss=0.31, pruned_loss=0.1072, over 951799.64 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:48:44,293 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 00:48:59,304 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5955, 1.4777, 1.9193, 1.2804, 1.7044, 1.6732, 1.4569, 1.9536], + device='cuda:0'), covar=tensor([0.1469, 0.2280, 0.1595, 0.1955, 0.1049, 0.1549, 0.2830, 0.1018], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0207, 0.0205, 0.0197, 0.0180, 0.0226, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 00:49:19,620 INFO [finetune.py:976] (0/7) Epoch 2, batch 5650, loss[loss=0.2622, simple_loss=0.308, pruned_loss=0.1082, over 4760.00 frames. ], tot_loss[loss=0.2649, simple_loss=0.3135, pruned_loss=0.1082, over 953197.83 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:49:44,831 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:49:54,701 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.700e+02 2.043e+02 2.333e+02 3.537e+02, threshold=4.085e+02, percent-clipped=0.0 +2023-03-26 00:50:01,882 INFO [finetune.py:976] (0/7) Epoch 2, batch 5700, loss[loss=0.2815, simple_loss=0.2991, pruned_loss=0.1319, over 3931.00 frames. ], tot_loss[loss=0.2619, simple_loss=0.3092, pruned_loss=0.1073, over 936566.26 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:50:03,156 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11430.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:50:13,812 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:50:17,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7883, 1.4675, 1.5180, 1.4838, 1.8320, 1.9462, 1.6708, 1.3955], + device='cuda:0'), covar=tensor([0.0255, 0.0358, 0.0502, 0.0320, 0.0257, 0.0297, 0.0272, 0.0378], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0112, 0.0134, 0.0113, 0.0103, 0.0097, 0.0088, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.3810e-05, 8.8664e-05, 1.0834e-04, 8.9642e-05, 8.1678e-05, 7.2827e-05, + 6.7628e-05, 8.4081e-05], device='cuda:0') +2023-03-26 00:50:18,694 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-2.pt +2023-03-26 00:50:33,842 INFO [finetune.py:976] (0/7) Epoch 3, batch 0, loss[loss=0.2582, simple_loss=0.2995, pruned_loss=0.1084, over 4798.00 frames. ], tot_loss[loss=0.2582, simple_loss=0.2995, pruned_loss=0.1084, over 4798.00 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:50:33,843 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 00:50:55,330 INFO [finetune.py:1010] (0/7) Epoch 3, validation: loss=0.1864, simple_loss=0.2566, pruned_loss=0.05807, over 2265189.00 frames. +2023-03-26 00:50:55,331 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6259MB +2023-03-26 00:51:14,458 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 00:51:20,331 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:51:29,402 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3848, 1.5026, 1.5054, 1.7238, 1.5782, 3.1980, 1.2404, 1.5886], + device='cuda:0'), covar=tensor([0.1126, 0.1706, 0.1220, 0.1095, 0.1645, 0.0287, 0.1643, 0.1798], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0080, 0.0078, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 00:51:38,821 INFO [finetune.py:976] (0/7) Epoch 3, batch 50, loss[loss=0.2899, simple_loss=0.3301, pruned_loss=0.1248, over 4905.00 frames. ], tot_loss[loss=0.2668, simple_loss=0.3146, pruned_loss=0.1095, over 217674.11 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:51:45,838 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.779e+02 2.075e+02 2.495e+02 4.593e+02, threshold=4.151e+02, percent-clipped=1.0 +2023-03-26 00:51:54,922 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:51:56,195 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11532.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:52:01,992 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11541.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:52:03,920 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 00:52:12,048 INFO [finetune.py:976] (0/7) Epoch 3, batch 100, loss[loss=0.2442, simple_loss=0.2932, pruned_loss=0.09766, over 4910.00 frames. ], tot_loss[loss=0.2559, simple_loss=0.3041, pruned_loss=0.1039, over 382438.15 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:52:25,128 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-26 00:52:33,515 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11589.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:52:36,421 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:52:39,394 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5771, 1.4540, 2.0457, 3.3761, 2.2598, 2.3206, 1.0948, 2.5529], + device='cuda:0'), covar=tensor([0.1878, 0.1664, 0.1431, 0.0541, 0.0896, 0.1502, 0.1931, 0.0642], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0138, 0.0163, 0.0105, 0.0145, 0.0130, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 00:52:47,407 INFO [finetune.py:976] (0/7) Epoch 3, batch 150, loss[loss=0.2305, simple_loss=0.2557, pruned_loss=0.1026, over 4261.00 frames. ], tot_loss[loss=0.2524, simple_loss=0.2995, pruned_loss=0.1027, over 511023.10 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:53:00,294 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.308e+02 1.853e+02 2.230e+02 2.582e+02 4.758e+02, threshold=4.459e+02, percent-clipped=3.0 +2023-03-26 00:53:24,139 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:53:48,108 INFO [finetune.py:976] (0/7) Epoch 3, batch 200, loss[loss=0.2917, simple_loss=0.3427, pruned_loss=0.1204, over 4829.00 frames. ], tot_loss[loss=0.2477, simple_loss=0.2958, pruned_loss=0.09982, over 608912.39 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:54:15,078 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 00:54:23,871 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-26 00:54:35,491 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:54:37,306 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:54:44,848 INFO [finetune.py:976] (0/7) Epoch 3, batch 250, loss[loss=0.259, simple_loss=0.3136, pruned_loss=0.1022, over 4819.00 frames. ], tot_loss[loss=0.2503, simple_loss=0.2994, pruned_loss=0.1006, over 684537.79 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:55:02,879 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.255e+02 1.783e+02 2.124e+02 2.629e+02 6.988e+02, threshold=4.248e+02, percent-clipped=1.0 +2023-03-26 00:55:32,486 INFO [finetune.py:976] (0/7) Epoch 3, batch 300, loss[loss=0.2517, simple_loss=0.3104, pruned_loss=0.09651, over 4904.00 frames. ], tot_loss[loss=0.2542, simple_loss=0.3038, pruned_loss=0.1023, over 743976.93 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:55:36,034 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:13,667 INFO [finetune.py:976] (0/7) Epoch 3, batch 350, loss[loss=0.2799, simple_loss=0.3296, pruned_loss=0.1151, over 4856.00 frames. ], tot_loss[loss=0.2583, simple_loss=0.3078, pruned_loss=0.1044, over 792182.99 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:56:20,332 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.891e+02 2.265e+02 2.576e+02 3.939e+02, threshold=4.529e+02, percent-clipped=0.0 +2023-03-26 00:56:20,449 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:30,479 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11830.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:45,043 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11846.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:51,649 INFO [finetune.py:976] (0/7) Epoch 3, batch 400, loss[loss=0.2585, simple_loss=0.3187, pruned_loss=0.09914, over 4873.00 frames. ], tot_loss[loss=0.2585, simple_loss=0.3081, pruned_loss=0.1045, over 828393.51 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:57:20,631 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11877.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:21,178 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:33,117 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:55,601 INFO [finetune.py:976] (0/7) Epoch 3, batch 450, loss[loss=0.2054, simple_loss=0.2608, pruned_loss=0.07497, over 4814.00 frames. ], tot_loss[loss=0.2567, simple_loss=0.3062, pruned_loss=0.1036, over 857086.37 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:57:56,360 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11907.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:12,549 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.239e+02 1.814e+02 2.317e+02 2.793e+02 4.030e+02, threshold=4.633e+02, percent-clipped=0.0 +2023-03-26 00:58:15,070 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:39,225 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:50,438 INFO [finetune.py:976] (0/7) Epoch 3, batch 500, loss[loss=0.2378, simple_loss=0.2833, pruned_loss=0.09614, over 4836.00 frames. ], tot_loss[loss=0.2549, simple_loss=0.3038, pruned_loss=0.103, over 879418.64 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:59:19,977 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:27,084 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11984.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:31,816 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:42,237 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-12000.pt +2023-03-26 00:59:46,861 INFO [finetune.py:976] (0/7) Epoch 3, batch 550, loss[loss=0.2613, simple_loss=0.3022, pruned_loss=0.1102, over 4909.00 frames. ], tot_loss[loss=0.2514, simple_loss=0.3002, pruned_loss=0.1013, over 897844.22 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:59:49,268 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12009.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:53,392 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.776e+02 2.066e+02 2.700e+02 4.009e+02, threshold=4.133e+02, percent-clipped=0.0 +2023-03-26 01:00:04,654 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-03-26 01:00:07,570 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5224, 1.5434, 1.4952, 1.5702, 0.9438, 2.9495, 1.0908, 1.6031], + device='cuda:0'), covar=tensor([0.3640, 0.2474, 0.2151, 0.2492, 0.2250, 0.0268, 0.2872, 0.1475], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0109, 0.0114, 0.0117, 0.0114, 0.0096, 0.0099, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 01:00:13,593 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12045.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:21,764 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:22,311 INFO [finetune.py:976] (0/7) Epoch 3, batch 600, loss[loss=0.2836, simple_loss=0.3279, pruned_loss=0.1197, over 4894.00 frames. ], tot_loss[loss=0.2528, simple_loss=0.3016, pruned_loss=0.102, over 910920.42 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 01:00:41,401 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:42,048 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1761, 1.9188, 2.1437, 0.9258, 2.1796, 2.4602, 1.9160, 2.0903], + device='cuda:0'), covar=tensor([0.0844, 0.0679, 0.0524, 0.0853, 0.0520, 0.0504, 0.0477, 0.0510], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0156, 0.0117, 0.0135, 0.0132, 0.0118, 0.0146, 0.0143], + device='cuda:0'), out_proj_covar=tensor([9.7573e-05, 1.1579e-04, 8.5631e-05, 9.9222e-05, 9.5680e-05, 8.7229e-05, + 1.0896e-04, 1.0614e-04], device='cuda:0') +2023-03-26 01:00:48,031 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 01:01:07,917 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 01:01:09,707 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 01:01:16,812 INFO [finetune.py:976] (0/7) Epoch 3, batch 650, loss[loss=0.2789, simple_loss=0.336, pruned_loss=0.1109, over 4842.00 frames. ], tot_loss[loss=0.2548, simple_loss=0.3046, pruned_loss=0.1025, over 920953.53 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 01:01:23,449 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.901e+02 2.250e+02 2.651e+02 5.885e+02, threshold=4.501e+02, percent-clipped=2.0 +2023-03-26 01:01:50,149 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12137.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:02,047 INFO [finetune.py:976] (0/7) Epoch 3, batch 700, loss[loss=0.2368, simple_loss=0.2913, pruned_loss=0.09112, over 4805.00 frames. ], tot_loss[loss=0.2572, simple_loss=0.3066, pruned_loss=0.1039, over 928234.96 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:02:02,759 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0509, 1.8045, 1.9937, 0.9888, 2.1064, 2.3807, 1.8358, 1.9350], + device='cuda:0'), covar=tensor([0.1169, 0.0862, 0.0641, 0.0914, 0.0528, 0.0544, 0.0610, 0.0686], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0157, 0.0118, 0.0136, 0.0133, 0.0119, 0.0147, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.8391e-05, 1.1709e-04, 8.6563e-05, 9.9923e-05, 9.6455e-05, 8.7789e-05, + 1.0979e-04, 1.0685e-04], device='cuda:0') +2023-03-26 01:02:12,356 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12172.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:24,037 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12188.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:39,280 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:43,052 INFO [finetune.py:976] (0/7) Epoch 3, batch 750, loss[loss=0.2425, simple_loss=0.3035, pruned_loss=0.09074, over 4828.00 frames. ], tot_loss[loss=0.2579, simple_loss=0.3078, pruned_loss=0.104, over 933776.70 frames. ], batch size: 30, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:02:54,948 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.858e+02 2.308e+02 2.738e+02 5.308e+02, threshold=4.616e+02, percent-clipped=1.0 +2023-03-26 01:03:08,384 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5781, 3.9812, 3.8437, 1.8987, 4.0910, 2.9644, 0.8083, 2.7557], + device='cuda:0'), covar=tensor([0.2419, 0.1584, 0.1572, 0.3156, 0.0966, 0.0999, 0.4547, 0.1410], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0169, 0.0167, 0.0130, 0.0157, 0.0122, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:03:11,348 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-03-26 01:03:14,094 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:03:30,122 INFO [finetune.py:976] (0/7) Epoch 3, batch 800, loss[loss=0.1942, simple_loss=0.2644, pruned_loss=0.06198, over 4845.00 frames. ], tot_loss[loss=0.258, simple_loss=0.3078, pruned_loss=0.1041, over 939017.04 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:03:42,870 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:03:53,101 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8124, 1.2004, 0.9391, 1.6303, 2.1292, 1.3918, 1.4757, 1.7592], + device='cuda:0'), covar=tensor([0.1629, 0.2369, 0.2315, 0.1308, 0.2270, 0.2092, 0.1563, 0.1938], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0098, 0.0117, 0.0094, 0.0125, 0.0097, 0.0100, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 01:04:03,344 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12291.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:11,111 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:12,239 INFO [finetune.py:976] (0/7) Epoch 3, batch 850, loss[loss=0.257, simple_loss=0.2938, pruned_loss=0.1101, over 4714.00 frames. ], tot_loss[loss=0.2563, simple_loss=0.3054, pruned_loss=0.1036, over 940882.55 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:04:19,496 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.828e+02 2.109e+02 2.576e+02 5.946e+02, threshold=4.217e+02, percent-clipped=1.0 +2023-03-26 01:04:25,766 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8160, 1.6599, 1.3411, 1.6905, 1.8970, 1.4694, 2.2481, 1.7591], + device='cuda:0'), covar=tensor([0.2503, 0.5017, 0.5670, 0.5018, 0.3615, 0.2605, 0.4328, 0.3475], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0196, 0.0238, 0.0253, 0.0218, 0.0183, 0.0206, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:04:46,217 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12339.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:46,837 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12340.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:06,369 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:06,917 INFO [finetune.py:976] (0/7) Epoch 3, batch 900, loss[loss=0.2147, simple_loss=0.2684, pruned_loss=0.08052, over 4803.00 frames. ], tot_loss[loss=0.2529, simple_loss=0.3022, pruned_loss=0.1018, over 945872.64 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:05:16,992 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2954, 2.9015, 3.0314, 3.2213, 3.0600, 2.8971, 3.3443, 1.0256], + device='cuda:0'), covar=tensor([0.0949, 0.0898, 0.1008, 0.0959, 0.1447, 0.1363, 0.0980, 0.4637], + device='cuda:0'), in_proj_covar=tensor([0.0368, 0.0245, 0.0274, 0.0297, 0.0341, 0.0286, 0.0311, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:05:30,240 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5682, 3.7169, 3.5778, 1.6115, 3.7444, 2.6754, 0.8868, 2.4706], + device='cuda:0'), covar=tensor([0.2406, 0.1965, 0.1658, 0.3823, 0.1113, 0.1217, 0.4784, 0.1619], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0168, 0.0167, 0.0129, 0.0157, 0.0122, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:05:43,766 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12403.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:45,595 INFO [finetune.py:976] (0/7) Epoch 3, batch 950, loss[loss=0.2677, simple_loss=0.2921, pruned_loss=0.1217, over 4368.00 frames. ], tot_loss[loss=0.2518, simple_loss=0.3007, pruned_loss=0.1015, over 947946.57 frames. ], batch size: 19, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:05:57,751 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.779e+02 2.123e+02 2.532e+02 4.452e+02, threshold=4.246e+02, percent-clipped=1.0 +2023-03-26 01:06:12,415 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:06:22,015 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12446.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:06:28,914 INFO [finetune.py:976] (0/7) Epoch 3, batch 1000, loss[loss=0.2606, simple_loss=0.2978, pruned_loss=0.1117, over 4686.00 frames. ], tot_loss[loss=0.2548, simple_loss=0.3035, pruned_loss=0.1031, over 948145.73 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:06:36,771 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4818, 1.6273, 1.9336, 1.8660, 1.6978, 3.7586, 1.3393, 1.7368], + device='cuda:0'), covar=tensor([0.1046, 0.1653, 0.1257, 0.1062, 0.1591, 0.0223, 0.1482, 0.1716], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:06:39,186 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:17,553 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:19,892 INFO [finetune.py:976] (0/7) Epoch 3, batch 1050, loss[loss=0.2337, simple_loss=0.2672, pruned_loss=0.1001, over 4411.00 frames. ], tot_loss[loss=0.2556, simple_loss=0.305, pruned_loss=0.1032, over 949118.02 frames. ], batch size: 19, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:07:20,629 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:31,080 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 2.008e+02 2.380e+02 2.733e+02 7.204e+02, threshold=4.759e+02, percent-clipped=3.0 +2023-03-26 01:07:38,101 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:08,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7135, 1.5484, 1.5811, 1.8951, 2.1933, 1.7761, 1.5126, 1.3179], + device='cuda:0'), covar=tensor([0.3087, 0.3110, 0.2553, 0.2441, 0.2983, 0.1661, 0.3548, 0.2618], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0207, 0.0194, 0.0179, 0.0229, 0.0171, 0.0210, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:08:10,384 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:17,959 INFO [finetune.py:976] (0/7) Epoch 3, batch 1100, loss[loss=0.2177, simple_loss=0.2819, pruned_loss=0.07671, over 4758.00 frames. ], tot_loss[loss=0.2567, simple_loss=0.3069, pruned_loss=0.1032, over 951181.86 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:08:35,450 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12576.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:44,777 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 01:08:57,756 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 01:08:58,294 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:59,394 INFO [finetune.py:976] (0/7) Epoch 3, batch 1150, loss[loss=0.245, simple_loss=0.301, pruned_loss=0.09452, over 4902.00 frames. ], tot_loss[loss=0.2574, simple_loss=0.3081, pruned_loss=0.1034, over 952418.24 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:09:11,543 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.712e+02 1.953e+02 2.432e+02 5.551e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 01:09:21,033 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:09:21,782 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-03-26 01:09:42,266 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:09:55,568 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12652.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:10:03,871 INFO [finetune.py:976] (0/7) Epoch 3, batch 1200, loss[loss=0.2745, simple_loss=0.3181, pruned_loss=0.1155, over 4856.00 frames. ], tot_loss[loss=0.2559, simple_loss=0.3065, pruned_loss=0.1027, over 953528.30 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:10:05,260 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7758, 1.1418, 1.5258, 1.4798, 1.3607, 1.3754, 1.3660, 1.4299], + device='cuda:0'), covar=tensor([1.1080, 1.8561, 1.3893, 1.5807, 1.6867, 1.2255, 2.0744, 1.2479], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0257, 0.0253, 0.0269, 0.0245, 0.0219, 0.0280, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:10:31,613 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12688.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:10:34,140 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7543, 0.9992, 1.5324, 1.4670, 1.3885, 1.3865, 1.3074, 1.4385], + device='cuda:0'), covar=tensor([1.0269, 1.8715, 1.5077, 1.6558, 1.7902, 1.2772, 2.0763, 1.3097], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0256, 0.0252, 0.0269, 0.0245, 0.0219, 0.0279, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:10:43,576 INFO [finetune.py:976] (0/7) Epoch 3, batch 1250, loss[loss=0.2541, simple_loss=0.3051, pruned_loss=0.1015, over 4819.00 frames. ], tot_loss[loss=0.2531, simple_loss=0.3036, pruned_loss=0.1014, over 954430.01 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:10:47,464 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 01:10:51,325 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.791e+02 2.152e+02 2.550e+02 3.946e+02, threshold=4.304e+02, percent-clipped=1.0 +2023-03-26 01:11:06,463 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:11:25,781 INFO [finetune.py:976] (0/7) Epoch 3, batch 1300, loss[loss=0.1928, simple_loss=0.2535, pruned_loss=0.0661, over 4763.00 frames. ], tot_loss[loss=0.2504, simple_loss=0.3006, pruned_loss=0.1002, over 954241.77 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:11:31,204 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:11:48,952 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12780.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:11:49,025 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5481, 1.2923, 1.2747, 1.4701, 1.7769, 1.4516, 0.8245, 1.3006], + device='cuda:0'), covar=tensor([0.2380, 0.2376, 0.2050, 0.1883, 0.1579, 0.1293, 0.2969, 0.2001], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0206, 0.0193, 0.0179, 0.0228, 0.0170, 0.0210, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:12:19,504 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:12:22,426 INFO [finetune.py:976] (0/7) Epoch 3, batch 1350, loss[loss=0.2178, simple_loss=0.2766, pruned_loss=0.07952, over 4770.00 frames. ], tot_loss[loss=0.2508, simple_loss=0.3008, pruned_loss=0.1004, over 953618.60 frames. ], batch size: 28, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:12:30,655 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 01:12:40,683 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.832e+02 2.182e+02 2.674e+02 3.468e+02, threshold=4.364e+02, percent-clipped=0.0 +2023-03-26 01:12:44,569 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 01:12:50,671 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12824.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:12:56,833 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 01:13:12,094 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5440, 1.5476, 1.7032, 1.9151, 1.5970, 3.4581, 1.2427, 1.6719], + device='cuda:0'), covar=tensor([0.1085, 0.1782, 0.1303, 0.1083, 0.1674, 0.0275, 0.1574, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:13:21,545 INFO [finetune.py:976] (0/7) Epoch 3, batch 1400, loss[loss=0.2445, simple_loss=0.2881, pruned_loss=0.1004, over 4881.00 frames. ], tot_loss[loss=0.2538, simple_loss=0.3043, pruned_loss=0.1016, over 955775.06 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:13:41,937 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3922, 3.3696, 3.2723, 1.5686, 3.4760, 2.5194, 0.8943, 2.2550], + device='cuda:0'), covar=tensor([0.2475, 0.1926, 0.1742, 0.3534, 0.1229, 0.1107, 0.4524, 0.1635], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0169, 0.0167, 0.0129, 0.0157, 0.0121, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:14:08,125 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6903, 1.5708, 1.2450, 1.4854, 1.4864, 1.3996, 1.4442, 2.2481], + device='cuda:0'), covar=tensor([1.3285, 1.2602, 1.0095, 1.3340, 1.0898, 0.7102, 1.2457, 0.4194], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0243, 0.0216, 0.0279, 0.0232, 0.0194, 0.0236, 0.0180], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:14:19,107 INFO [finetune.py:976] (0/7) Epoch 3, batch 1450, loss[loss=0.2853, simple_loss=0.3485, pruned_loss=0.1111, over 4867.00 frames. ], tot_loss[loss=0.2569, simple_loss=0.3074, pruned_loss=0.1032, over 955910.34 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:14:38,007 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.344e+02 1.849e+02 2.229e+02 2.789e+02 4.972e+02, threshold=4.459e+02, percent-clipped=1.0 +2023-03-26 01:15:12,661 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4277, 1.6106, 1.8700, 1.9002, 1.7998, 4.0944, 1.3991, 1.8337], + device='cuda:0'), covar=tensor([0.1096, 0.1738, 0.1330, 0.1109, 0.1540, 0.0210, 0.1464, 0.1736], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:15:13,772 INFO [finetune.py:976] (0/7) Epoch 3, batch 1500, loss[loss=0.2906, simple_loss=0.349, pruned_loss=0.1161, over 4846.00 frames. ], tot_loss[loss=0.2577, simple_loss=0.3081, pruned_loss=0.1037, over 955518.91 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:15:58,825 INFO [finetune.py:976] (0/7) Epoch 3, batch 1550, loss[loss=0.2619, simple_loss=0.3034, pruned_loss=0.1102, over 4748.00 frames. ], tot_loss[loss=0.2552, simple_loss=0.3064, pruned_loss=0.1021, over 956523.92 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:16:09,621 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.917e+01 1.840e+02 2.219e+02 2.788e+02 8.539e+02, threshold=4.437e+02, percent-clipped=2.0 +2023-03-26 01:16:44,395 INFO [finetune.py:976] (0/7) Epoch 3, batch 1600, loss[loss=0.2461, simple_loss=0.2956, pruned_loss=0.09833, over 4849.00 frames. ], tot_loss[loss=0.2547, simple_loss=0.3047, pruned_loss=0.1024, over 956782.25 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:17:22,876 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-26 01:17:41,223 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:17:43,655 INFO [finetune.py:976] (0/7) Epoch 3, batch 1650, loss[loss=0.2201, simple_loss=0.2651, pruned_loss=0.08751, over 4770.00 frames. ], tot_loss[loss=0.2509, simple_loss=0.3009, pruned_loss=0.1004, over 957079.08 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:17:55,149 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.264e+02 1.841e+02 2.109e+02 2.450e+02 4.226e+02, threshold=4.217e+02, percent-clipped=0.0 +2023-03-26 01:17:56,996 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=13119.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:18:07,778 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5712, 2.4068, 2.5481, 2.8124, 3.0589, 2.7709, 2.4010, 2.2892], + device='cuda:0'), covar=tensor([0.2170, 0.2024, 0.1610, 0.1530, 0.1838, 0.0999, 0.2362, 0.1761], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0207, 0.0194, 0.0180, 0.0230, 0.0171, 0.0211, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:18:19,026 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6659, 1.4640, 1.3266, 1.3466, 1.7269, 1.4531, 1.7076, 1.5855], + device='cuda:0'), covar=tensor([0.2184, 0.4223, 0.5047, 0.4065, 0.3498, 0.2387, 0.3670, 0.3034], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0198, 0.0240, 0.0256, 0.0221, 0.0185, 0.0210, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:18:32,131 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:18:41,626 INFO [finetune.py:976] (0/7) Epoch 3, batch 1700, loss[loss=0.2578, simple_loss=0.3125, pruned_loss=0.1016, over 4761.00 frames. ], tot_loss[loss=0.2502, simple_loss=0.2997, pruned_loss=0.1004, over 958606.13 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:18:42,461 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-03-26 01:18:48,319 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6002, 1.3841, 1.8025, 2.7863, 2.0295, 2.1065, 0.9111, 2.2383], + device='cuda:0'), covar=tensor([0.1669, 0.1606, 0.1269, 0.0632, 0.0831, 0.1450, 0.1812, 0.0669], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0119, 0.0137, 0.0163, 0.0104, 0.0144, 0.0129, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 01:18:54,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6930, 1.6217, 1.5964, 1.6351, 1.1395, 3.4379, 1.4240, 2.0161], + device='cuda:0'), covar=tensor([0.3410, 0.2247, 0.1942, 0.2165, 0.1897, 0.0170, 0.2766, 0.1328], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0110, 0.0114, 0.0118, 0.0115, 0.0096, 0.0100, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 01:19:22,367 INFO [finetune.py:976] (0/7) Epoch 3, batch 1750, loss[loss=0.2463, simple_loss=0.291, pruned_loss=0.1008, over 4871.00 frames. ], tot_loss[loss=0.2534, simple_loss=0.3025, pruned_loss=0.1022, over 955899.54 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:19:31,217 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.973e+02 2.314e+02 2.693e+02 6.749e+02, threshold=4.629e+02, percent-clipped=3.0 +2023-03-26 01:20:18,401 INFO [finetune.py:976] (0/7) Epoch 3, batch 1800, loss[loss=0.249, simple_loss=0.3056, pruned_loss=0.09615, over 4141.00 frames. ], tot_loss[loss=0.2557, simple_loss=0.306, pruned_loss=0.1027, over 955160.30 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:20:31,986 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 01:20:49,395 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-26 01:20:53,323 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5269, 1.5003, 1.7441, 1.8956, 1.6261, 3.4329, 1.3033, 1.6850], + device='cuda:0'), covar=tensor([0.1116, 0.1865, 0.1205, 0.1116, 0.1702, 0.0303, 0.1705, 0.1863], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0084, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:20:59,915 INFO [finetune.py:976] (0/7) Epoch 3, batch 1850, loss[loss=0.279, simple_loss=0.3245, pruned_loss=0.1167, over 4808.00 frames. ], tot_loss[loss=0.2576, simple_loss=0.3079, pruned_loss=0.1036, over 955753.73 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:21:08,031 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.691e+02 1.934e+02 2.488e+02 4.482e+02, threshold=3.868e+02, percent-clipped=0.0 +2023-03-26 01:21:25,137 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8452, 2.3729, 2.0996, 1.0547, 2.2467, 2.1910, 1.8183, 2.1804], + device='cuda:0'), covar=tensor([0.0632, 0.1055, 0.1626, 0.2585, 0.1499, 0.2068, 0.2276, 0.1065], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0197, 0.0203, 0.0189, 0.0214, 0.0209, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:21:46,506 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2479, 2.1869, 2.3305, 1.3708, 2.5028, 2.8046, 2.0942, 2.2906], + device='cuda:0'), covar=tensor([0.1280, 0.0701, 0.0686, 0.0647, 0.0605, 0.0577, 0.0663, 0.0677], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0157, 0.0119, 0.0136, 0.0133, 0.0120, 0.0146, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.7643e-05, 1.1715e-04, 8.6949e-05, 9.9716e-05, 9.6296e-05, 8.8262e-05, + 1.0918e-04, 1.0683e-04], device='cuda:0') +2023-03-26 01:21:50,069 INFO [finetune.py:976] (0/7) Epoch 3, batch 1900, loss[loss=0.2629, simple_loss=0.3304, pruned_loss=0.09763, over 4818.00 frames. ], tot_loss[loss=0.2561, simple_loss=0.3072, pruned_loss=0.1025, over 955465.26 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:21:59,734 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6106, 3.0322, 2.6607, 1.6563, 2.9319, 2.6370, 2.2965, 2.6013], + device='cuda:0'), covar=tensor([0.0684, 0.1161, 0.1842, 0.2910, 0.1709, 0.2110, 0.2258, 0.1346], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0197, 0.0203, 0.0188, 0.0214, 0.0209, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:22:40,358 INFO [finetune.py:976] (0/7) Epoch 3, batch 1950, loss[loss=0.2214, simple_loss=0.2654, pruned_loss=0.0887, over 3914.00 frames. ], tot_loss[loss=0.2524, simple_loss=0.3035, pruned_loss=0.1007, over 954755.03 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:22:45,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7807, 1.5124, 1.3461, 1.2433, 1.5053, 1.4862, 1.4518, 2.2245], + device='cuda:0'), covar=tensor([1.3610, 1.2166, 1.0009, 1.3084, 0.9774, 0.6931, 1.1070, 0.4373], + device='cuda:0'), in_proj_covar=tensor([0.0268, 0.0244, 0.0217, 0.0280, 0.0233, 0.0195, 0.0236, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:22:47,001 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.849e+02 2.191e+02 2.472e+02 6.030e+02, threshold=4.381e+02, percent-clipped=4.0 +2023-03-26 01:22:48,811 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13419.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:23:25,636 INFO [finetune.py:976] (0/7) Epoch 3, batch 2000, loss[loss=0.2198, simple_loss=0.2658, pruned_loss=0.08684, over 4028.00 frames. ], tot_loss[loss=0.2494, simple_loss=0.2998, pruned_loss=0.09954, over 955597.21 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:23:36,603 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:24:18,865 INFO [finetune.py:976] (0/7) Epoch 3, batch 2050, loss[loss=0.1707, simple_loss=0.2397, pruned_loss=0.05091, over 4788.00 frames. ], tot_loss[loss=0.2462, simple_loss=0.2963, pruned_loss=0.09803, over 955387.92 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:24:22,269 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2239, 1.7343, 2.6784, 1.7718, 2.4010, 2.3690, 2.0120, 2.5415], + device='cuda:0'), covar=tensor([0.1928, 0.2369, 0.1831, 0.2655, 0.1078, 0.1901, 0.2377, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0206, 0.0207, 0.0200, 0.0182, 0.0230, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:24:26,007 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 01:24:33,950 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.806e+02 2.206e+02 2.674e+02 5.377e+02, threshold=4.412e+02, percent-clipped=2.0 +2023-03-26 01:24:35,404 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 01:25:00,755 INFO [finetune.py:976] (0/7) Epoch 3, batch 2100, loss[loss=0.2542, simple_loss=0.3048, pruned_loss=0.1019, over 4910.00 frames. ], tot_loss[loss=0.2466, simple_loss=0.2964, pruned_loss=0.09841, over 954504.69 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:25:11,101 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2273, 2.8011, 2.7017, 1.2762, 2.9578, 2.1288, 0.6684, 1.7924], + device='cuda:0'), covar=tensor([0.2556, 0.2088, 0.1944, 0.3467, 0.1295, 0.1119, 0.4236, 0.1884], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0167, 0.0165, 0.0128, 0.0155, 0.0120, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:25:40,720 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8236, 3.3734, 3.4620, 3.7452, 3.5932, 3.3485, 3.8978, 1.2642], + device='cuda:0'), covar=tensor([0.0856, 0.0795, 0.0885, 0.0891, 0.1423, 0.1466, 0.0811, 0.4783], + device='cuda:0'), in_proj_covar=tensor([0.0367, 0.0244, 0.0275, 0.0294, 0.0341, 0.0286, 0.0310, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:25:45,307 INFO [finetune.py:976] (0/7) Epoch 3, batch 2150, loss[loss=0.265, simple_loss=0.3203, pruned_loss=0.1049, over 4840.00 frames. ], tot_loss[loss=0.2516, simple_loss=0.3013, pruned_loss=0.1009, over 956098.92 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:25:54,299 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3966, 2.0148, 1.5761, 0.6976, 1.9168, 1.8227, 1.5310, 1.7300], + device='cuda:0'), covar=tensor([0.0730, 0.0951, 0.1550, 0.2375, 0.1211, 0.2349, 0.2387, 0.1115], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0202, 0.0188, 0.0214, 0.0208, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:26:01,366 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.918e+02 2.256e+02 2.684e+02 5.304e+02, threshold=4.512e+02, percent-clipped=2.0 +2023-03-26 01:26:06,390 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3646, 1.3858, 1.5039, 0.8815, 1.3295, 1.6272, 1.6331, 1.3388], + device='cuda:0'), covar=tensor([0.1350, 0.0779, 0.0516, 0.0720, 0.0482, 0.0515, 0.0479, 0.0750], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0157, 0.0118, 0.0136, 0.0133, 0.0120, 0.0147, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.8134e-05, 1.1728e-04, 8.6577e-05, 9.9716e-05, 9.6310e-05, 8.8294e-05, + 1.0955e-04, 1.0677e-04], device='cuda:0') +2023-03-26 01:26:07,691 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 01:26:27,571 INFO [finetune.py:976] (0/7) Epoch 3, batch 2200, loss[loss=0.2449, simple_loss=0.3079, pruned_loss=0.09096, over 4811.00 frames. ], tot_loss[loss=0.2525, simple_loss=0.303, pruned_loss=0.101, over 956770.22 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:27:00,297 INFO [finetune.py:976] (0/7) Epoch 3, batch 2250, loss[loss=0.2533, simple_loss=0.3081, pruned_loss=0.09921, over 4254.00 frames. ], tot_loss[loss=0.2556, simple_loss=0.3057, pruned_loss=0.1027, over 954409.21 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:27:08,388 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.927e+02 2.166e+02 2.564e+02 5.587e+02, threshold=4.333e+02, percent-clipped=2.0 +2023-03-26 01:27:24,282 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=13737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:27:42,151 INFO [finetune.py:976] (0/7) Epoch 3, batch 2300, loss[loss=0.2437, simple_loss=0.3055, pruned_loss=0.09093, over 4887.00 frames. ], tot_loss[loss=0.2548, simple_loss=0.3053, pruned_loss=0.1021, over 955557.19 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:28:25,836 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=13798.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 01:28:36,681 INFO [finetune.py:976] (0/7) Epoch 3, batch 2350, loss[loss=0.2488, simple_loss=0.3016, pruned_loss=0.09798, over 4923.00 frames. ], tot_loss[loss=0.2528, simple_loss=0.3028, pruned_loss=0.1014, over 958121.06 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:28:50,254 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.589e+01 1.791e+02 2.182e+02 2.579e+02 6.380e+02, threshold=4.365e+02, percent-clipped=2.0 +2023-03-26 01:29:04,347 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6208, 1.5497, 1.4684, 1.7772, 2.0943, 1.7153, 1.2361, 1.3615], + device='cuda:0'), covar=tensor([0.2542, 0.2467, 0.2072, 0.1854, 0.2051, 0.1288, 0.3049, 0.2009], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0206, 0.0193, 0.0180, 0.0228, 0.0170, 0.0210, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:29:15,828 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 01:29:31,310 INFO [finetune.py:976] (0/7) Epoch 3, batch 2400, loss[loss=0.2339, simple_loss=0.2842, pruned_loss=0.09176, over 4761.00 frames. ], tot_loss[loss=0.2505, simple_loss=0.3002, pruned_loss=0.1004, over 958101.06 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:29:40,686 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4407, 1.3221, 1.4904, 2.3748, 1.7098, 2.0202, 0.7010, 1.9684], + device='cuda:0'), covar=tensor([0.1919, 0.1764, 0.1453, 0.1039, 0.1021, 0.1531, 0.2032, 0.0876], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0120, 0.0139, 0.0165, 0.0105, 0.0146, 0.0131, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 01:29:49,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3047, 2.9099, 3.0279, 3.2414, 3.0653, 2.8677, 3.3866, 1.1036], + device='cuda:0'), covar=tensor([0.1171, 0.0961, 0.1122, 0.1138, 0.1733, 0.1724, 0.1019, 0.4972], + device='cuda:0'), in_proj_covar=tensor([0.0366, 0.0242, 0.0273, 0.0293, 0.0337, 0.0284, 0.0309, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:30:02,996 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-03-26 01:30:15,690 INFO [finetune.py:976] (0/7) Epoch 3, batch 2450, loss[loss=0.2346, simple_loss=0.2917, pruned_loss=0.0888, over 4838.00 frames. ], tot_loss[loss=0.2472, simple_loss=0.2968, pruned_loss=0.09879, over 957626.63 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:30:29,103 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.953e+01 1.906e+02 2.150e+02 2.578e+02 4.181e+02, threshold=4.299e+02, percent-clipped=0.0 +2023-03-26 01:30:59,053 INFO [finetune.py:976] (0/7) Epoch 3, batch 2500, loss[loss=0.2854, simple_loss=0.3422, pruned_loss=0.1143, over 4824.00 frames. ], tot_loss[loss=0.2477, simple_loss=0.2974, pruned_loss=0.09897, over 955199.52 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:31:23,221 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1026, 1.9500, 1.6052, 1.4498, 2.1363, 2.4035, 2.0918, 1.8647], + device='cuda:0'), covar=tensor([0.0297, 0.0358, 0.0536, 0.0422, 0.0372, 0.0330, 0.0293, 0.0393], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0114, 0.0136, 0.0117, 0.0104, 0.0099, 0.0089, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.5155e-05, 9.0049e-05, 1.0938e-04, 9.2592e-05, 8.2573e-05, 7.3611e-05, + 6.8996e-05, 8.5595e-05], device='cuda:0') +2023-03-26 01:31:33,355 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5230, 1.2766, 1.2863, 1.2578, 1.6406, 1.6675, 1.4628, 1.2481], + device='cuda:0'), covar=tensor([0.0261, 0.0316, 0.0547, 0.0303, 0.0218, 0.0324, 0.0289, 0.0394], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0114, 0.0136, 0.0117, 0.0104, 0.0099, 0.0089, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.5119e-05, 8.9955e-05, 1.0932e-04, 9.2515e-05, 8.2488e-05, 7.3581e-05, + 6.8915e-05, 8.5522e-05], device='cuda:0') +2023-03-26 01:31:43,818 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-14000.pt +2023-03-26 01:31:53,427 INFO [finetune.py:976] (0/7) Epoch 3, batch 2550, loss[loss=0.2474, simple_loss=0.3054, pruned_loss=0.09471, over 4924.00 frames. ], tot_loss[loss=0.2508, simple_loss=0.3012, pruned_loss=0.1002, over 956268.50 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:32:02,446 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.924e+01 1.822e+02 2.075e+02 2.520e+02 4.375e+02, threshold=4.150e+02, percent-clipped=1.0 +2023-03-26 01:32:07,591 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4074, 2.0949, 1.8423, 0.8463, 2.0346, 1.8654, 1.6085, 1.8918], + device='cuda:0'), covar=tensor([0.0914, 0.1063, 0.1821, 0.2644, 0.1625, 0.2347, 0.2355, 0.1350], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0197, 0.0203, 0.0189, 0.0215, 0.0209, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:32:30,596 INFO [finetune.py:976] (0/7) Epoch 3, batch 2600, loss[loss=0.2385, simple_loss=0.2889, pruned_loss=0.09409, over 4791.00 frames. ], tot_loss[loss=0.2512, simple_loss=0.302, pruned_loss=0.1002, over 956488.73 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:32:41,345 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:32:43,720 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0806, 1.8650, 2.6538, 1.6491, 2.3014, 2.4506, 1.8361, 2.5069], + device='cuda:0'), covar=tensor([0.1763, 0.2260, 0.1525, 0.2877, 0.1169, 0.1874, 0.2743, 0.1215], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0203, 0.0204, 0.0195, 0.0180, 0.0225, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:32:45,614 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-26 01:33:04,857 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14093.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:33:16,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:33:18,153 INFO [finetune.py:976] (0/7) Epoch 3, batch 2650, loss[loss=0.2479, simple_loss=0.3089, pruned_loss=0.09339, over 4746.00 frames. ], tot_loss[loss=0.251, simple_loss=0.3026, pruned_loss=0.09972, over 957631.72 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:33:34,829 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.796e+02 2.195e+02 2.771e+02 4.502e+02, threshold=4.390e+02, percent-clipped=2.0 +2023-03-26 01:33:37,262 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14120.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:33:56,190 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14133.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:34:18,020 INFO [finetune.py:976] (0/7) Epoch 3, batch 2700, loss[loss=0.2396, simple_loss=0.282, pruned_loss=0.09857, over 4828.00 frames. ], tot_loss[loss=0.2495, simple_loss=0.3015, pruned_loss=0.09881, over 956594.52 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:34:28,739 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:34:50,944 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14181.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 01:35:13,611 INFO [finetune.py:976] (0/7) Epoch 3, batch 2750, loss[loss=0.2523, simple_loss=0.2897, pruned_loss=0.1074, over 4765.00 frames. ], tot_loss[loss=0.2469, simple_loss=0.2985, pruned_loss=0.09763, over 957096.36 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:35:20,817 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.613e+02 1.949e+02 2.418e+02 3.837e+02, threshold=3.898e+02, percent-clipped=0.0 +2023-03-26 01:35:50,233 INFO [finetune.py:976] (0/7) Epoch 3, batch 2800, loss[loss=0.2318, simple_loss=0.2782, pruned_loss=0.09266, over 4929.00 frames. ], tot_loss[loss=0.2442, simple_loss=0.2953, pruned_loss=0.09658, over 957774.62 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:36:35,462 INFO [finetune.py:976] (0/7) Epoch 3, batch 2850, loss[loss=0.2911, simple_loss=0.3362, pruned_loss=0.123, over 4793.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.2939, pruned_loss=0.09589, over 956898.50 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:36:48,618 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.210e+02 1.714e+02 2.069e+02 2.397e+02 3.427e+02, threshold=4.138e+02, percent-clipped=0.0 +2023-03-26 01:36:59,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6687, 0.8041, 1.4337, 1.3659, 1.2679, 1.2748, 1.2173, 1.3296], + device='cuda:0'), covar=tensor([0.8307, 1.4385, 1.1197, 1.2463, 1.3643, 0.9766, 1.5508, 1.0177], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0255, 0.0252, 0.0267, 0.0243, 0.0217, 0.0277, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:37:26,625 INFO [finetune.py:976] (0/7) Epoch 3, batch 2900, loss[loss=0.2288, simple_loss=0.2927, pruned_loss=0.08247, over 4829.00 frames. ], tot_loss[loss=0.2464, simple_loss=0.2973, pruned_loss=0.09772, over 953446.80 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:37:52,773 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 01:38:03,308 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8568, 1.6447, 2.3090, 1.5513, 2.1558, 2.1909, 1.6017, 2.3113], + device='cuda:0'), covar=tensor([0.1691, 0.2365, 0.1739, 0.2472, 0.1013, 0.1814, 0.3179, 0.1048], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0204, 0.0204, 0.0195, 0.0180, 0.0225, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:38:06,359 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:38:25,744 INFO [finetune.py:976] (0/7) Epoch 3, batch 2950, loss[loss=0.2074, simple_loss=0.2747, pruned_loss=0.06998, over 4795.00 frames. ], tot_loss[loss=0.2498, simple_loss=0.3014, pruned_loss=0.09907, over 953262.85 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:38:35,701 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14413.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:38:38,198 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.306e+02 1.859e+02 2.169e+02 2.721e+02 5.785e+02, threshold=4.339e+02, percent-clipped=3.0 +2023-03-26 01:38:50,256 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:39:01,607 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.13 vs. limit=5.0 +2023-03-26 01:39:02,088 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:39:20,293 INFO [finetune.py:976] (0/7) Epoch 3, batch 3000, loss[loss=0.2531, simple_loss=0.298, pruned_loss=0.1041, over 4742.00 frames. ], tot_loss[loss=0.2496, simple_loss=0.3015, pruned_loss=0.09888, over 953199.31 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:39:20,294 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 01:39:22,204 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8220, 1.5445, 1.3446, 1.3231, 1.4751, 1.5018, 1.4499, 2.2685], + device='cuda:0'), covar=tensor([1.3068, 1.2765, 1.0193, 1.2623, 0.9992, 0.6660, 1.2548, 0.4126], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0247, 0.0219, 0.0283, 0.0235, 0.0196, 0.0238, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:39:24,961 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6122, 1.4378, 1.9625, 2.9334, 2.0434, 2.2208, 1.0116, 2.2784], + device='cuda:0'), covar=tensor([0.1835, 0.1743, 0.1243, 0.0604, 0.0893, 0.1142, 0.1861, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0138, 0.0165, 0.0104, 0.0145, 0.0130, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 01:39:25,316 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5666, 1.6424, 1.9388, 1.3489, 1.6914, 1.8237, 1.6233, 2.0025], + device='cuda:0'), covar=tensor([0.1595, 0.2196, 0.1428, 0.1960, 0.1063, 0.1469, 0.2681, 0.0911], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0204, 0.0204, 0.0196, 0.0181, 0.0225, 0.0214, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:39:27,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7830, 1.6672, 1.6998, 1.7138, 1.1221, 3.0828, 1.3145, 1.8293], + device='cuda:0'), covar=tensor([0.3141, 0.2203, 0.1929, 0.2187, 0.1897, 0.0263, 0.2443, 0.1257], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0119, 0.0116, 0.0097, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 01:39:37,115 INFO [finetune.py:1010] (0/7) Epoch 3, validation: loss=0.1777, simple_loss=0.2485, pruned_loss=0.05342, over 2265189.00 frames. +2023-03-26 01:39:37,116 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6259MB +2023-03-26 01:39:42,137 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:39:45,718 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2577, 2.8113, 2.7082, 1.2323, 2.9439, 2.1015, 0.7144, 1.8396], + device='cuda:0'), covar=tensor([0.2554, 0.1865, 0.1765, 0.3507, 0.1388, 0.1138, 0.4230, 0.1620], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0167, 0.0165, 0.0129, 0.0155, 0.0120, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:39:53,762 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0443, 1.7882, 1.4082, 1.8713, 1.9520, 1.6585, 2.2890, 1.9344], + device='cuda:0'), covar=tensor([0.2141, 0.4281, 0.5110, 0.4525, 0.3469, 0.2280, 0.4737, 0.3064], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0195, 0.0239, 0.0254, 0.0220, 0.0184, 0.0209, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:39:56,810 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:40:03,359 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14476.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:40:15,392 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14487.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:40:36,153 INFO [finetune.py:976] (0/7) Epoch 3, batch 3050, loss[loss=0.2729, simple_loss=0.3244, pruned_loss=0.1107, over 4814.00 frames. ], tot_loss[loss=0.2509, simple_loss=0.3032, pruned_loss=0.09928, over 954615.98 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:40:43,252 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0355, 1.8518, 1.5513, 2.0440, 1.8570, 1.8086, 1.7934, 2.8546], + device='cuda:0'), covar=tensor([1.2229, 1.2429, 0.9736, 1.2400, 1.0366, 0.6287, 1.2930, 0.3658], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0247, 0.0219, 0.0283, 0.0235, 0.0196, 0.0238, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:40:46,679 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 01:40:52,883 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.290e+02 1.934e+02 2.277e+02 2.724e+02 4.940e+02, threshold=4.554e+02, percent-clipped=2.0 +2023-03-26 01:40:57,957 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 01:41:17,995 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14548.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:41:22,791 INFO [finetune.py:976] (0/7) Epoch 3, batch 3100, loss[loss=0.2045, simple_loss=0.2613, pruned_loss=0.07389, over 4801.00 frames. ], tot_loss[loss=0.2494, simple_loss=0.3015, pruned_loss=0.09867, over 954632.38 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:41:24,673 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0174, 1.8780, 1.6459, 1.9076, 2.1093, 1.7635, 2.3621, 1.9927], + device='cuda:0'), covar=tensor([0.2014, 0.3766, 0.4435, 0.3982, 0.3061, 0.2186, 0.4076, 0.2727], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0195, 0.0238, 0.0253, 0.0220, 0.0183, 0.0208, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:42:10,663 INFO [finetune.py:976] (0/7) Epoch 3, batch 3150, loss[loss=0.3022, simple_loss=0.3426, pruned_loss=0.1309, over 4912.00 frames. ], tot_loss[loss=0.2446, simple_loss=0.297, pruned_loss=0.09615, over 955106.25 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:42:18,343 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.775e+02 2.189e+02 2.683e+02 4.981e+02, threshold=4.378e+02, percent-clipped=2.0 +2023-03-26 01:42:25,619 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.02 vs. limit=5.0 +2023-03-26 01:43:00,562 INFO [finetune.py:976] (0/7) Epoch 3, batch 3200, loss[loss=0.296, simple_loss=0.3391, pruned_loss=0.1265, over 4865.00 frames. ], tot_loss[loss=0.2408, simple_loss=0.2929, pruned_loss=0.09429, over 956728.22 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:43:19,342 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5934, 1.6712, 1.7015, 1.1834, 1.6593, 2.0101, 1.9653, 1.5592], + device='cuda:0'), covar=tensor([0.0973, 0.0562, 0.0509, 0.0549, 0.0441, 0.0469, 0.0329, 0.0603], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0159, 0.0119, 0.0137, 0.0133, 0.0120, 0.0148, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.9313e-05, 1.1823e-04, 8.6670e-05, 1.0070e-04, 9.6367e-05, 8.9073e-05, + 1.1064e-04, 1.0709e-04], device='cuda:0') +2023-03-26 01:43:41,546 INFO [finetune.py:976] (0/7) Epoch 3, batch 3250, loss[loss=0.2635, simple_loss=0.3188, pruned_loss=0.1041, over 4823.00 frames. ], tot_loss[loss=0.2399, simple_loss=0.2914, pruned_loss=0.09419, over 955290.78 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:43:41,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7080, 1.5278, 1.2915, 1.2916, 1.4609, 1.4695, 1.4217, 2.2161], + device='cuda:0'), covar=tensor([1.1758, 1.0470, 0.9009, 1.1418, 0.8752, 0.6276, 1.0502, 0.3599], + device='cuda:0'), in_proj_covar=tensor([0.0269, 0.0245, 0.0217, 0.0281, 0.0233, 0.0194, 0.0236, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:43:54,504 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.718e+02 2.102e+02 2.544e+02 5.358e+02, threshold=4.204e+02, percent-clipped=1.0 +2023-03-26 01:44:08,042 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14728.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:15,442 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:31,817 INFO [finetune.py:976] (0/7) Epoch 3, batch 3300, loss[loss=0.2943, simple_loss=0.3438, pruned_loss=0.1224, over 4872.00 frames. ], tot_loss[loss=0.2441, simple_loss=0.296, pruned_loss=0.09609, over 953464.07 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:44:33,798 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:42,696 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14769.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:48,715 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:48,741 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14776.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:45:08,523 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:45:16,878 INFO [finetune.py:976] (0/7) Epoch 3, batch 3350, loss[loss=0.2645, simple_loss=0.318, pruned_loss=0.1055, over 4865.00 frames. ], tot_loss[loss=0.2446, simple_loss=0.2973, pruned_loss=0.09596, over 954392.82 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:45:17,524 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14807.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:45:29,676 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.682e+02 2.084e+02 2.593e+02 4.183e+02, threshold=4.169e+02, percent-clipped=0.0 +2023-03-26 01:45:39,451 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14824.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 01:45:39,483 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5701, 1.5462, 1.8983, 1.9696, 1.8055, 4.2390, 1.4903, 1.8450], + device='cuda:0'), covar=tensor([0.1082, 0.1745, 0.1350, 0.1075, 0.1543, 0.0184, 0.1435, 0.1762], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0079, 0.0080, 0.0093, 0.0084, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:45:52,839 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14843.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:46:01,259 INFO [finetune.py:976] (0/7) Epoch 3, batch 3400, loss[loss=0.2422, simple_loss=0.2969, pruned_loss=0.09376, over 4733.00 frames. ], tot_loss[loss=0.2488, simple_loss=0.301, pruned_loss=0.09825, over 955331.59 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:46:09,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8610, 1.5403, 1.3974, 1.3459, 1.9090, 2.0209, 1.7923, 1.3958], + device='cuda:0'), covar=tensor([0.0216, 0.0404, 0.0486, 0.0393, 0.0219, 0.0319, 0.0237, 0.0417], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0112, 0.0134, 0.0115, 0.0103, 0.0097, 0.0088, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.4254e-05, 8.8925e-05, 1.0767e-04, 9.1112e-05, 8.1935e-05, 7.2160e-05, + 6.7647e-05, 8.3703e-05], device='cuda:0') +2023-03-26 01:46:15,310 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.26 vs. limit=5.0 +2023-03-26 01:46:19,994 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3755, 1.3947, 1.2871, 1.5980, 1.5589, 3.0520, 1.3205, 1.5711], + device='cuda:0'), covar=tensor([0.1100, 0.1762, 0.1401, 0.1087, 0.1628, 0.0267, 0.1501, 0.1713], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 01:46:41,116 INFO [finetune.py:976] (0/7) Epoch 3, batch 3450, loss[loss=0.277, simple_loss=0.317, pruned_loss=0.1185, over 4902.00 frames. ], tot_loss[loss=0.249, simple_loss=0.3013, pruned_loss=0.09833, over 955312.36 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:46:53,149 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.352e+02 1.935e+02 2.237e+02 2.692e+02 3.962e+02, threshold=4.475e+02, percent-clipped=0.0 +2023-03-26 01:47:17,592 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 01:47:27,611 INFO [finetune.py:976] (0/7) Epoch 3, batch 3500, loss[loss=0.2201, simple_loss=0.2777, pruned_loss=0.08124, over 4769.00 frames. ], tot_loss[loss=0.2454, simple_loss=0.2978, pruned_loss=0.09646, over 955750.60 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:47:58,135 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 01:48:19,805 INFO [finetune.py:976] (0/7) Epoch 3, batch 3550, loss[loss=0.2035, simple_loss=0.242, pruned_loss=0.08255, over 4218.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.2942, pruned_loss=0.09566, over 954887.75 frames. ], batch size: 17, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:48:26,974 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.780e+02 2.199e+02 2.800e+02 5.904e+02, threshold=4.398e+02, percent-clipped=2.0 +2023-03-26 01:49:03,540 INFO [finetune.py:976] (0/7) Epoch 3, batch 3600, loss[loss=0.2572, simple_loss=0.2996, pruned_loss=0.1074, over 4821.00 frames. ], tot_loss[loss=0.2404, simple_loss=0.2915, pruned_loss=0.09463, over 955226.08 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:49:06,153 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9909, 1.6343, 1.5260, 1.7316, 2.0026, 1.6335, 2.0798, 1.8432], + device='cuda:0'), covar=tensor([0.2116, 0.4242, 0.5052, 0.4038, 0.3421, 0.2416, 0.3793, 0.3084], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0198, 0.0241, 0.0256, 0.0223, 0.0186, 0.0211, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:49:12,115 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:49:42,972 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:49:47,964 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9990, 2.1440, 1.9251, 1.4310, 2.3033, 2.2741, 2.1811, 1.7831], + device='cuda:0'), covar=tensor([0.0781, 0.0619, 0.0890, 0.1044, 0.0395, 0.0813, 0.0742, 0.1081], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0132, 0.0145, 0.0130, 0.0110, 0.0143, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:49:49,067 INFO [finetune.py:976] (0/7) Epoch 3, batch 3650, loss[loss=0.2643, simple_loss=0.3225, pruned_loss=0.1031, over 4734.00 frames. ], tot_loss[loss=0.2426, simple_loss=0.2943, pruned_loss=0.0955, over 955530.35 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:49:50,472 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8757, 1.7056, 1.4171, 1.4942, 1.8808, 1.5816, 2.3771, 1.7520], + device='cuda:0'), covar=tensor([0.1766, 0.3537, 0.4330, 0.4088, 0.2969, 0.1890, 0.2841, 0.2747], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0198, 0.0240, 0.0256, 0.0223, 0.0187, 0.0211, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:49:56,315 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.927e+02 2.238e+02 2.686e+02 4.916e+02, threshold=4.476e+02, percent-clipped=1.0 +2023-03-26 01:49:56,384 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:50:28,401 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15143.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:50:41,688 INFO [finetune.py:976] (0/7) Epoch 3, batch 3700, loss[loss=0.2491, simple_loss=0.2957, pruned_loss=0.1012, over 4897.00 frames. ], tot_loss[loss=0.2454, simple_loss=0.2973, pruned_loss=0.09673, over 954447.85 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:51:16,177 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:51:16,370 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 01:51:19,608 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:51:29,140 INFO [finetune.py:976] (0/7) Epoch 3, batch 3750, loss[loss=0.286, simple_loss=0.3235, pruned_loss=0.1242, over 4821.00 frames. ], tot_loss[loss=0.2465, simple_loss=0.2987, pruned_loss=0.09712, over 953980.29 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:51:40,515 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.801e+02 2.153e+02 2.622e+02 6.720e+02, threshold=4.305e+02, percent-clipped=1.0 +2023-03-26 01:52:33,794 INFO [finetune.py:976] (0/7) Epoch 3, batch 3800, loss[loss=0.2796, simple_loss=0.3251, pruned_loss=0.117, over 4837.00 frames. ], tot_loss[loss=0.2481, simple_loss=0.3008, pruned_loss=0.09773, over 955091.97 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:52:33,920 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15256.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:53:22,426 INFO [finetune.py:976] (0/7) Epoch 3, batch 3850, loss[loss=0.2434, simple_loss=0.2959, pruned_loss=0.09548, over 4774.00 frames. ], tot_loss[loss=0.2465, simple_loss=0.299, pruned_loss=0.09699, over 951009.73 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:53:39,316 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.276e+02 1.876e+02 2.253e+02 2.579e+02 5.032e+02, threshold=4.505e+02, percent-clipped=1.0 +2023-03-26 01:54:26,485 INFO [finetune.py:976] (0/7) Epoch 3, batch 3900, loss[loss=0.2313, simple_loss=0.2809, pruned_loss=0.09089, over 4758.00 frames. ], tot_loss[loss=0.2433, simple_loss=0.2957, pruned_loss=0.09544, over 952834.52 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:54:33,125 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8191, 1.5930, 1.4594, 1.5059, 1.8540, 1.5524, 2.0370, 1.7637], + device='cuda:0'), covar=tensor([0.2113, 0.3763, 0.4326, 0.3750, 0.3098, 0.2278, 0.3908, 0.2554], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0239, 0.0255, 0.0221, 0.0186, 0.0209, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:54:34,372 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2940, 2.0596, 1.6092, 2.2083, 2.3487, 1.8032, 2.8561, 2.2059], + device='cuda:0'), covar=tensor([0.2108, 0.5080, 0.5050, 0.4898, 0.3318, 0.2326, 0.3961, 0.3014], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0239, 0.0255, 0.0221, 0.0185, 0.0209, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:54:48,579 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9783, 2.1845, 2.0679, 1.4499, 2.2279, 2.1844, 2.1969, 1.8495], + device='cuda:0'), covar=tensor([0.0645, 0.0561, 0.0738, 0.1029, 0.0464, 0.0729, 0.0676, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0132, 0.0144, 0.0130, 0.0110, 0.0143, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 01:54:49,187 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8186, 1.6285, 1.6112, 1.7533, 1.0114, 3.5802, 1.3667, 1.9283], + device='cuda:0'), covar=tensor([0.3327, 0.2588, 0.2044, 0.2256, 0.2135, 0.0187, 0.2729, 0.1413], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 01:54:52,413 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 01:55:10,107 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:55:14,087 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 01:55:14,720 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 01:55:16,819 INFO [finetune.py:976] (0/7) Epoch 3, batch 3950, loss[loss=0.2099, simple_loss=0.2634, pruned_loss=0.07824, over 4905.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.2913, pruned_loss=0.09302, over 954189.35 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:55:25,254 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.678e+02 2.164e+02 2.472e+02 4.231e+02, threshold=4.328e+02, percent-clipped=0.0 +2023-03-26 01:55:52,136 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:55:59,827 INFO [finetune.py:976] (0/7) Epoch 3, batch 4000, loss[loss=0.3257, simple_loss=0.3625, pruned_loss=0.1445, over 4815.00 frames. ], tot_loss[loss=0.2414, simple_loss=0.2929, pruned_loss=0.09491, over 954785.20 frames. ], batch size: 41, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:57:05,460 INFO [finetune.py:976] (0/7) Epoch 3, batch 4050, loss[loss=0.2547, simple_loss=0.3142, pruned_loss=0.09758, over 4836.00 frames. ], tot_loss[loss=0.2426, simple_loss=0.2944, pruned_loss=0.09533, over 954724.88 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:57:20,271 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.798e+02 2.110e+02 2.647e+02 5.396e+02, threshold=4.219e+02, percent-clipped=2.0 +2023-03-26 01:57:29,828 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9823, 1.7963, 1.5404, 1.8550, 1.6983, 1.7169, 1.6491, 2.5637], + device='cuda:0'), covar=tensor([1.0825, 1.1801, 0.8397, 1.1803, 1.0245, 0.5914, 1.2626, 0.3305], + device='cuda:0'), in_proj_covar=tensor([0.0273, 0.0249, 0.0218, 0.0284, 0.0235, 0.0196, 0.0239, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 01:57:58,329 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15551.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:57:58,430 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 01:58:05,295 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 01:58:06,365 INFO [finetune.py:976] (0/7) Epoch 3, batch 4100, loss[loss=0.2518, simple_loss=0.3071, pruned_loss=0.0982, over 4859.00 frames. ], tot_loss[loss=0.2446, simple_loss=0.2972, pruned_loss=0.09602, over 955195.15 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:59:02,792 INFO [finetune.py:976] (0/7) Epoch 3, batch 4150, loss[loss=0.2765, simple_loss=0.3264, pruned_loss=0.1132, over 4924.00 frames. ], tot_loss[loss=0.2461, simple_loss=0.2986, pruned_loss=0.09684, over 955756.13 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:59:10,589 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.791e+02 2.157e+02 2.467e+02 4.537e+02, threshold=4.313e+02, percent-clipped=1.0 +2023-03-26 01:59:11,309 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5303, 3.5929, 3.4176, 1.7816, 3.7374, 2.9523, 0.9622, 2.5195], + device='cuda:0'), covar=tensor([0.2862, 0.1631, 0.1584, 0.3120, 0.0910, 0.0843, 0.4112, 0.1484], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0168, 0.0164, 0.0128, 0.0154, 0.0121, 0.0146, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 01:59:51,469 INFO [finetune.py:976] (0/7) Epoch 3, batch 4200, loss[loss=0.268, simple_loss=0.3143, pruned_loss=0.1109, over 4834.00 frames. ], tot_loss[loss=0.2462, simple_loss=0.2993, pruned_loss=0.09652, over 954291.75 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:00:53,262 INFO [finetune.py:976] (0/7) Epoch 3, batch 4250, loss[loss=0.2226, simple_loss=0.2805, pruned_loss=0.0824, over 4911.00 frames. ], tot_loss[loss=0.2438, simple_loss=0.2967, pruned_loss=0.09548, over 954288.59 frames. ], batch size: 46, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:01:00,001 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.731e+02 2.073e+02 2.469e+02 5.386e+02, threshold=4.147e+02, percent-clipped=2.0 +2023-03-26 02:01:14,006 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 02:01:43,526 INFO [finetune.py:976] (0/7) Epoch 3, batch 4300, loss[loss=0.2272, simple_loss=0.2789, pruned_loss=0.0878, over 4832.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.2931, pruned_loss=0.09421, over 954659.67 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:02:02,407 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-26 02:02:43,903 INFO [finetune.py:976] (0/7) Epoch 3, batch 4350, loss[loss=0.2851, simple_loss=0.3363, pruned_loss=0.1169, over 4798.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.29, pruned_loss=0.09278, over 955297.78 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:03:01,710 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.712e+02 2.007e+02 2.460e+02 4.679e+02, threshold=4.015e+02, percent-clipped=1.0 +2023-03-26 02:03:33,440 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:03:36,424 INFO [finetune.py:976] (0/7) Epoch 3, batch 4400, loss[loss=0.3453, simple_loss=0.3791, pruned_loss=0.1557, over 4787.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.292, pruned_loss=0.09448, over 954027.41 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:04:05,477 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15883.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:04:06,137 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7660, 1.1008, 1.4635, 1.4117, 1.2611, 1.2911, 1.3753, 1.4132], + device='cuda:0'), covar=tensor([0.8720, 1.4105, 1.1132, 1.2650, 1.3299, 0.9433, 1.6418, 1.0123], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0253, 0.0252, 0.0265, 0.0242, 0.0216, 0.0277, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 02:04:20,878 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:04:30,835 INFO [finetune.py:976] (0/7) Epoch 3, batch 4450, loss[loss=0.2242, simple_loss=0.2808, pruned_loss=0.08382, over 4819.00 frames. ], tot_loss[loss=0.2433, simple_loss=0.296, pruned_loss=0.09532, over 954989.29 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:04:48,084 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.872e+02 2.224e+02 2.526e+02 5.583e+02, threshold=4.448e+02, percent-clipped=1.0 +2023-03-26 02:05:16,013 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:05:23,247 INFO [finetune.py:976] (0/7) Epoch 3, batch 4500, loss[loss=0.2872, simple_loss=0.3354, pruned_loss=0.1195, over 4900.00 frames. ], tot_loss[loss=0.2445, simple_loss=0.2974, pruned_loss=0.09577, over 953151.29 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:05:44,029 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3526, 1.4813, 1.5461, 1.7006, 1.6309, 3.2158, 1.3848, 1.5565], + device='cuda:0'), covar=tensor([0.1077, 0.1652, 0.1166, 0.1036, 0.1473, 0.0231, 0.1380, 0.1695], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0080, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:06:15,550 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-16000.pt +2023-03-26 02:06:17,329 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16001.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:06:17,340 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1781, 1.8747, 2.2764, 0.9877, 2.2687, 2.6855, 1.9439, 2.1770], + device='cuda:0'), covar=tensor([0.1080, 0.1047, 0.0554, 0.0927, 0.0778, 0.0515, 0.0682, 0.0697], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0160, 0.0119, 0.0140, 0.0135, 0.0122, 0.0149, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.9956e-05, 1.1930e-04, 8.6946e-05, 1.0256e-04, 9.7726e-05, 9.0125e-05, + 1.1088e-04, 1.0832e-04], device='cuda:0') +2023-03-26 02:06:17,417 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 02:06:17,467 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 02:06:24,310 INFO [finetune.py:976] (0/7) Epoch 3, batch 4550, loss[loss=0.197, simple_loss=0.2567, pruned_loss=0.06862, over 4744.00 frames. ], tot_loss[loss=0.2448, simple_loss=0.2977, pruned_loss=0.0959, over 953037.42 frames. ], batch size: 23, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:06:36,868 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.237e+02 1.760e+02 2.085e+02 2.487e+02 3.865e+02, threshold=4.170e+02, percent-clipped=0.0 +2023-03-26 02:06:44,084 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4418, 1.5523, 1.7110, 1.9350, 1.6394, 3.4210, 1.2510, 1.6829], + device='cuda:0'), covar=tensor([0.1018, 0.1630, 0.1227, 0.0948, 0.1534, 0.0220, 0.1533, 0.1675], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0080, 0.0077, 0.0079, 0.0091, 0.0083, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:06:46,085 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 02:07:05,903 INFO [finetune.py:976] (0/7) Epoch 3, batch 4600, loss[loss=0.253, simple_loss=0.2951, pruned_loss=0.1055, over 4871.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.2964, pruned_loss=0.09465, over 954578.36 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:07:11,395 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16062.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:07:35,220 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16085.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:07:59,849 INFO [finetune.py:976] (0/7) Epoch 3, batch 4650, loss[loss=0.205, simple_loss=0.2604, pruned_loss=0.07479, over 4935.00 frames. ], tot_loss[loss=0.2394, simple_loss=0.2923, pruned_loss=0.09324, over 953724.66 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:08:07,260 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.817e+02 2.201e+02 2.598e+02 3.850e+02, threshold=4.403e+02, percent-clipped=0.0 +2023-03-26 02:08:29,947 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9463, 1.9737, 1.9683, 1.3119, 2.0702, 2.1306, 2.1302, 1.7455], + device='cuda:0'), covar=tensor([0.0612, 0.0609, 0.0703, 0.0998, 0.0536, 0.0641, 0.0575, 0.1062], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0134, 0.0146, 0.0131, 0.0111, 0.0144, 0.0148, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:08:30,561 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5278, 1.6212, 1.7562, 1.9724, 1.6857, 3.4983, 1.3550, 1.7130], + device='cuda:0'), covar=tensor([0.1051, 0.1688, 0.1147, 0.0995, 0.1641, 0.0242, 0.1494, 0.1719], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:08:39,866 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16146.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:08:42,108 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16149.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:08:52,382 INFO [finetune.py:976] (0/7) Epoch 3, batch 4700, loss[loss=0.2396, simple_loss=0.2838, pruned_loss=0.09775, over 4839.00 frames. ], tot_loss[loss=0.2363, simple_loss=0.2891, pruned_loss=0.09177, over 953667.91 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:09:03,341 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6418, 1.6353, 2.1730, 2.0417, 1.9068, 4.4035, 1.4235, 1.9089], + device='cuda:0'), covar=tensor([0.1077, 0.1790, 0.1145, 0.1116, 0.1640, 0.0229, 0.1538, 0.1722], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:09:30,875 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8201, 1.4858, 2.0646, 3.3955, 2.5399, 2.4039, 0.9323, 2.6401], + device='cuda:0'), covar=tensor([0.1771, 0.1652, 0.1511, 0.0628, 0.0763, 0.1600, 0.2027, 0.0668], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0120, 0.0139, 0.0166, 0.0104, 0.0144, 0.0130, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 02:09:41,914 INFO [finetune.py:976] (0/7) Epoch 3, batch 4750, loss[loss=0.2215, simple_loss=0.276, pruned_loss=0.08356, over 4802.00 frames. ], tot_loss[loss=0.235, simple_loss=0.2875, pruned_loss=0.09118, over 955023.73 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:09:44,972 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16210.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:09:49,753 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.299e+02 1.759e+02 2.065e+02 2.416e+02 4.123e+02, threshold=4.129e+02, percent-clipped=0.0 +2023-03-26 02:09:52,424 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9945, 1.7866, 1.4458, 1.8979, 1.8048, 1.6609, 1.6688, 2.8250], + device='cuda:0'), covar=tensor([1.1248, 1.2781, 0.8640, 1.2589, 1.0245, 0.6516, 1.1923, 0.3178], + device='cuda:0'), in_proj_covar=tensor([0.0273, 0.0249, 0.0218, 0.0283, 0.0234, 0.0195, 0.0238, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 02:10:03,634 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16239.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:10:07,390 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-26 02:10:16,138 INFO [finetune.py:976] (0/7) Epoch 3, batch 4800, loss[loss=0.2068, simple_loss=0.2619, pruned_loss=0.07586, over 4806.00 frames. ], tot_loss[loss=0.2372, simple_loss=0.2903, pruned_loss=0.09204, over 953662.22 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:10:16,262 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0830, 1.7814, 2.4263, 1.5977, 2.2926, 2.3085, 1.8535, 2.6485], + device='cuda:0'), covar=tensor([0.1466, 0.2212, 0.1451, 0.2074, 0.0937, 0.1396, 0.2544, 0.0825], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0207, 0.0204, 0.0197, 0.0183, 0.0227, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:11:10,320 INFO [finetune.py:976] (0/7) Epoch 3, batch 4850, loss[loss=0.2298, simple_loss=0.2754, pruned_loss=0.09209, over 4803.00 frames. ], tot_loss[loss=0.2404, simple_loss=0.2936, pruned_loss=0.09362, over 953332.14 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:11:12,280 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-03-26 02:11:18,711 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.278e+02 1.868e+02 2.289e+02 2.628e+02 4.977e+02, threshold=4.577e+02, percent-clipped=4.0 +2023-03-26 02:11:45,418 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 02:11:53,001 INFO [finetune.py:976] (0/7) Epoch 3, batch 4900, loss[loss=0.275, simple_loss=0.3176, pruned_loss=0.1162, over 4892.00 frames. ], tot_loss[loss=0.2418, simple_loss=0.2951, pruned_loss=0.09424, over 952016.12 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:11:53,738 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16357.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:12:04,220 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16371.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:12:04,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6768, 1.4999, 1.4172, 1.3579, 1.7943, 1.4711, 1.6784, 1.6127], + device='cuda:0'), covar=tensor([0.1790, 0.3124, 0.3928, 0.3190, 0.2616, 0.1993, 0.2865, 0.2486], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0195, 0.0237, 0.0253, 0.0220, 0.0185, 0.0209, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:12:34,568 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4786, 1.3818, 1.4258, 1.4042, 0.9432, 2.2560, 0.7938, 1.4056], + device='cuda:0'), covar=tensor([0.3497, 0.2424, 0.2163, 0.2499, 0.2031, 0.0366, 0.2739, 0.1432], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0111, 0.0116, 0.0119, 0.0116, 0.0096, 0.0101, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 02:12:35,109 INFO [finetune.py:976] (0/7) Epoch 3, batch 4950, loss[loss=0.2234, simple_loss=0.2839, pruned_loss=0.08147, over 4805.00 frames. ], tot_loss[loss=0.2454, simple_loss=0.298, pruned_loss=0.0964, over 952085.25 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:12:43,744 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.793e+02 2.170e+02 2.564e+02 4.726e+02, threshold=4.340e+02, percent-clipped=1.0 +2023-03-26 02:12:53,396 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:12:56,873 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2485, 3.7001, 3.9141, 4.0949, 4.0232, 3.8110, 4.3345, 1.4419], + device='cuda:0'), covar=tensor([0.0665, 0.0680, 0.0707, 0.0814, 0.0987, 0.1261, 0.0626, 0.4684], + device='cuda:0'), in_proj_covar=tensor([0.0365, 0.0246, 0.0278, 0.0297, 0.0344, 0.0289, 0.0314, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:12:59,278 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:13:11,535 INFO [finetune.py:976] (0/7) Epoch 3, batch 5000, loss[loss=0.2685, simple_loss=0.3032, pruned_loss=0.1169, over 4744.00 frames. ], tot_loss[loss=0.2425, simple_loss=0.2953, pruned_loss=0.09484, over 952850.00 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:13:20,387 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-03-26 02:14:08,361 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:14:08,894 INFO [finetune.py:976] (0/7) Epoch 3, batch 5050, loss[loss=0.2502, simple_loss=0.2957, pruned_loss=0.1024, over 4827.00 frames. ], tot_loss[loss=0.2402, simple_loss=0.2928, pruned_loss=0.09378, over 954389.10 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:14:27,727 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.680e+02 2.024e+02 2.446e+02 4.498e+02, threshold=4.048e+02, percent-clipped=1.0 +2023-03-26 02:14:39,966 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 02:14:49,226 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16539.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:15:09,827 INFO [finetune.py:976] (0/7) Epoch 3, batch 5100, loss[loss=0.2786, simple_loss=0.3116, pruned_loss=0.1228, over 4820.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.2886, pruned_loss=0.09143, over 954961.59 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:15:24,779 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6538, 1.4937, 1.4834, 1.6061, 1.1698, 3.6044, 1.3822, 1.9579], + device='cuda:0'), covar=tensor([0.3413, 0.2306, 0.2127, 0.2306, 0.1909, 0.0159, 0.2785, 0.1466], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0119, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 02:15:36,552 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16587.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:15:43,937 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 02:15:52,950 INFO [finetune.py:976] (0/7) Epoch 3, batch 5150, loss[loss=0.2634, simple_loss=0.3206, pruned_loss=0.1031, over 4813.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.2903, pruned_loss=0.09267, over 954325.66 frames. ], batch size: 40, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:16:04,173 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 02:16:12,054 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.762e+02 2.113e+02 2.590e+02 4.768e+02, threshold=4.226e+02, percent-clipped=2.0 +2023-03-26 02:16:47,969 INFO [finetune.py:976] (0/7) Epoch 3, batch 5200, loss[loss=0.2316, simple_loss=0.294, pruned_loss=0.08457, over 4763.00 frames. ], tot_loss[loss=0.2404, simple_loss=0.2938, pruned_loss=0.09351, over 955231.80 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:16:48,666 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:17:40,739 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16705.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:17:41,306 INFO [finetune.py:976] (0/7) Epoch 3, batch 5250, loss[loss=0.2207, simple_loss=0.2831, pruned_loss=0.07911, over 4816.00 frames. ], tot_loss[loss=0.2415, simple_loss=0.2957, pruned_loss=0.09362, over 953268.17 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:17:53,304 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.211e+02 1.833e+02 2.103e+02 2.647e+02 4.683e+02, threshold=4.205e+02, percent-clipped=1.0 +2023-03-26 02:18:00,066 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:18:12,242 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:18:21,889 INFO [finetune.py:976] (0/7) Epoch 3, batch 5300, loss[loss=0.2025, simple_loss=0.263, pruned_loss=0.07097, over 4113.00 frames. ], tot_loss[loss=0.2434, simple_loss=0.2975, pruned_loss=0.09461, over 952077.78 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:18:49,194 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16789.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:18:50,711 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 02:19:07,617 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16805.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:19:08,225 INFO [finetune.py:976] (0/7) Epoch 3, batch 5350, loss[loss=0.2222, simple_loss=0.28, pruned_loss=0.08218, over 4758.00 frames. ], tot_loss[loss=0.2432, simple_loss=0.2974, pruned_loss=0.09445, over 951134.01 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:19:21,972 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.880e+02 2.227e+02 2.511e+02 3.677e+02, threshold=4.454e+02, percent-clipped=0.0 +2023-03-26 02:19:46,291 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16853.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:19:48,084 INFO [finetune.py:976] (0/7) Epoch 3, batch 5400, loss[loss=0.2616, simple_loss=0.3031, pruned_loss=0.1101, over 4792.00 frames. ], tot_loss[loss=0.2412, simple_loss=0.2948, pruned_loss=0.09378, over 952370.02 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:19:49,445 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6445, 1.5421, 1.5034, 1.7926, 2.0875, 1.7266, 1.1737, 1.3857], + device='cuda:0'), covar=tensor([0.2635, 0.2490, 0.2241, 0.1881, 0.2121, 0.1431, 0.3407, 0.2122], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0209, 0.0197, 0.0182, 0.0233, 0.0172, 0.0213, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:19:55,537 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16868.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:20:31,377 INFO [finetune.py:976] (0/7) Epoch 3, batch 5450, loss[loss=0.181, simple_loss=0.2543, pruned_loss=0.05389, over 4824.00 frames. ], tot_loss[loss=0.2372, simple_loss=0.2905, pruned_loss=0.09201, over 952626.40 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:20:31,476 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:20:38,653 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.665e+02 2.000e+02 2.450e+02 4.433e+02, threshold=4.000e+02, percent-clipped=0.0 +2023-03-26 02:20:43,565 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0329, 1.8904, 1.6320, 2.0438, 2.1292, 1.8153, 2.4978, 2.1409], + device='cuda:0'), covar=tensor([0.1877, 0.3828, 0.4135, 0.3782, 0.2918, 0.1944, 0.3597, 0.2488], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0240, 0.0255, 0.0223, 0.0187, 0.0210, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:20:46,033 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16929.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:21:14,236 INFO [finetune.py:976] (0/7) Epoch 3, batch 5500, loss[loss=0.2128, simple_loss=0.2701, pruned_loss=0.07782, over 4750.00 frames. ], tot_loss[loss=0.2352, simple_loss=0.2877, pruned_loss=0.09133, over 953234.68 frames. ], batch size: 27, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:21:16,790 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3531, 2.2568, 2.8932, 1.7808, 2.4834, 2.7519, 2.1120, 2.8512], + device='cuda:0'), covar=tensor([0.1736, 0.1977, 0.1590, 0.2648, 0.1141, 0.1808, 0.2581, 0.1363], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0209, 0.0206, 0.0200, 0.0185, 0.0229, 0.0219, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:21:26,174 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16967.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:22:07,777 INFO [finetune.py:976] (0/7) Epoch 3, batch 5550, loss[loss=0.2185, simple_loss=0.2715, pruned_loss=0.0827, over 4791.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.2884, pruned_loss=0.09149, over 954226.49 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:22:15,686 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.712e+02 2.015e+02 2.380e+02 4.122e+02, threshold=4.030e+02, percent-clipped=1.0 +2023-03-26 02:22:21,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:22:41,436 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1015, 3.5577, 3.7270, 3.9784, 3.8382, 3.6464, 4.1603, 1.3027], + device='cuda:0'), covar=tensor([0.0799, 0.0861, 0.0840, 0.0954, 0.1279, 0.1500, 0.0737, 0.5364], + device='cuda:0'), in_proj_covar=tensor([0.0363, 0.0245, 0.0277, 0.0295, 0.0342, 0.0289, 0.0311, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:22:53,359 INFO [finetune.py:976] (0/7) Epoch 3, batch 5600, loss[loss=0.2286, simple_loss=0.2875, pruned_loss=0.08489, over 4837.00 frames. ], tot_loss[loss=0.2403, simple_loss=0.2934, pruned_loss=0.09362, over 955104.95 frames. ], batch size: 30, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:23:10,678 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17075.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:23:21,399 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 02:23:38,817 INFO [finetune.py:976] (0/7) Epoch 3, batch 5650, loss[loss=0.2947, simple_loss=0.3505, pruned_loss=0.1195, over 4743.00 frames. ], tot_loss[loss=0.2411, simple_loss=0.295, pruned_loss=0.09359, over 953364.59 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:23:45,803 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.755e+02 2.152e+02 2.681e+02 4.789e+02, threshold=4.305e+02, percent-clipped=1.0 +2023-03-26 02:24:09,619 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 02:24:15,373 INFO [finetune.py:976] (0/7) Epoch 3, batch 5700, loss[loss=0.2679, simple_loss=0.288, pruned_loss=0.124, over 4177.00 frames. ], tot_loss[loss=0.2397, simple_loss=0.2917, pruned_loss=0.09386, over 933771.66 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:24:41,800 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-3.pt +2023-03-26 02:24:56,870 INFO [finetune.py:976] (0/7) Epoch 4, batch 0, loss[loss=0.3484, simple_loss=0.3756, pruned_loss=0.1606, over 4837.00 frames. ], tot_loss[loss=0.3484, simple_loss=0.3756, pruned_loss=0.1606, over 4837.00 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:24:56,872 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 02:25:04,198 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2064, 1.4092, 1.4845, 0.8207, 1.2804, 1.5477, 1.6479, 1.4174], + device='cuda:0'), covar=tensor([0.1051, 0.0644, 0.0569, 0.0557, 0.0494, 0.0703, 0.0364, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0157, 0.0117, 0.0136, 0.0132, 0.0121, 0.0146, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.8181e-05, 1.1719e-04, 8.5409e-05, 1.0017e-04, 9.5355e-05, 9.0035e-05, + 1.0877e-04, 1.0736e-04], device='cuda:0') +2023-03-26 02:25:18,215 INFO [finetune.py:1010] (0/7) Epoch 4, validation: loss=0.1768, simple_loss=0.2473, pruned_loss=0.0532, over 2265189.00 frames. +2023-03-26 02:25:18,215 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 02:25:22,722 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17189.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:25:55,799 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.713e+02 2.128e+02 2.708e+02 4.853e+02, threshold=4.257e+02, percent-clipped=3.0 +2023-03-26 02:25:59,546 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17224.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 02:26:01,330 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6649, 0.5603, 1.5374, 1.3822, 1.3598, 1.3307, 1.2004, 1.4449], + device='cuda:0'), covar=tensor([0.6496, 1.0714, 0.9376, 0.9607, 1.0181, 0.7343, 1.2054, 0.8317], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0254, 0.0256, 0.0266, 0.0245, 0.0219, 0.0280, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:26:05,891 INFO [finetune.py:976] (0/7) Epoch 4, batch 50, loss[loss=0.2268, simple_loss=0.2808, pruned_loss=0.08642, over 4862.00 frames. ], tot_loss[loss=0.2418, simple_loss=0.2955, pruned_loss=0.09403, over 216437.10 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:26:29,255 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:26:34,025 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7366, 1.0061, 0.7655, 1.6250, 2.1094, 1.3547, 1.2674, 1.6833], + device='cuda:0'), covar=tensor([0.1427, 0.2212, 0.2236, 0.1216, 0.1868, 0.2212, 0.1569, 0.1799], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0100, 0.0118, 0.0095, 0.0125, 0.0098, 0.0102, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 02:26:36,462 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17262.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:26:55,371 INFO [finetune.py:976] (0/7) Epoch 4, batch 100, loss[loss=0.2227, simple_loss=0.2791, pruned_loss=0.0831, over 4890.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.2894, pruned_loss=0.09199, over 381602.47 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:27:26,862 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.697e+02 1.982e+02 2.273e+02 3.827e+02, threshold=3.964e+02, percent-clipped=0.0 +2023-03-26 02:27:36,644 INFO [finetune.py:976] (0/7) Epoch 4, batch 150, loss[loss=0.2235, simple_loss=0.2678, pruned_loss=0.08957, over 4833.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.2837, pruned_loss=0.08956, over 509577.93 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:27:45,320 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3489, 1.5325, 1.7966, 1.7325, 1.6118, 3.3310, 1.2507, 1.6568], + device='cuda:0'), covar=tensor([0.0981, 0.1506, 0.1152, 0.1021, 0.1414, 0.0237, 0.1401, 0.1589], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0079, 0.0092, 0.0082, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:28:01,924 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:28:16,143 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 02:28:25,903 INFO [finetune.py:976] (0/7) Epoch 4, batch 200, loss[loss=0.2447, simple_loss=0.2834, pruned_loss=0.103, over 4902.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.2813, pruned_loss=0.08892, over 608557.49 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:28:26,032 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17383.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:28:55,774 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.314e+02 1.771e+02 2.098e+02 2.514e+02 4.657e+02, threshold=4.195e+02, percent-clipped=1.0 +2023-03-26 02:28:59,995 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7047, 1.5646, 1.4735, 1.6828, 2.2333, 1.8303, 1.1885, 1.3906], + device='cuda:0'), covar=tensor([0.2571, 0.2525, 0.2187, 0.2013, 0.2077, 0.1308, 0.3348, 0.2027], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0209, 0.0198, 0.0184, 0.0233, 0.0174, 0.0215, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:29:01,214 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:29:09,065 INFO [finetune.py:976] (0/7) Epoch 4, batch 250, loss[loss=0.2869, simple_loss=0.3376, pruned_loss=0.1181, over 4815.00 frames. ], tot_loss[loss=0.2344, simple_loss=0.2873, pruned_loss=0.09073, over 686744.96 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:29:10,385 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9296, 2.1677, 1.8947, 1.3498, 2.2932, 2.2094, 2.1853, 1.8958], + device='cuda:0'), covar=tensor([0.0712, 0.0579, 0.0857, 0.1013, 0.0478, 0.0768, 0.0731, 0.1041], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0133, 0.0144, 0.0129, 0.0111, 0.0142, 0.0147, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:29:17,851 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17444.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:29:49,005 INFO [finetune.py:976] (0/7) Epoch 4, batch 300, loss[loss=0.2676, simple_loss=0.323, pruned_loss=0.1061, over 4817.00 frames. ], tot_loss[loss=0.2388, simple_loss=0.2921, pruned_loss=0.09272, over 748285.24 frames. ], batch size: 40, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:30:17,235 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1981, 1.3176, 1.0242, 1.2711, 1.4613, 2.4998, 1.1974, 1.4558], + device='cuda:0'), covar=tensor([0.1062, 0.1935, 0.1255, 0.1079, 0.1690, 0.0344, 0.1652, 0.1828], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:30:35,039 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.302e+02 1.964e+02 2.271e+02 2.699e+02 6.272e+02, threshold=4.542e+02, percent-clipped=2.0 +2023-03-26 02:30:39,299 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17524.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:30:44,122 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4879, 1.5486, 1.1777, 1.3801, 1.8622, 1.7684, 1.5891, 1.2989], + device='cuda:0'), covar=tensor([0.0343, 0.0350, 0.0712, 0.0371, 0.0220, 0.0423, 0.0343, 0.0456], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0114, 0.0137, 0.0117, 0.0105, 0.0099, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.6045e-05, 8.9945e-05, 1.1061e-04, 9.2694e-05, 8.3039e-05, 7.4165e-05, + 6.9962e-05, 8.4910e-05], device='cuda:0') +2023-03-26 02:30:44,608 INFO [finetune.py:976] (0/7) Epoch 4, batch 350, loss[loss=0.2478, simple_loss=0.3123, pruned_loss=0.09166, over 4921.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.2947, pruned_loss=0.09319, over 794681.49 frames. ], batch size: 42, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:30:53,138 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:31:10,201 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17562.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:31:10,236 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7328, 1.3367, 0.9363, 0.2594, 1.2115, 1.4788, 1.2617, 1.4719], + device='cuda:0'), covar=tensor([0.0851, 0.1151, 0.1662, 0.2537, 0.1724, 0.2575, 0.2867, 0.0992], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0200, 0.0204, 0.0190, 0.0217, 0.0211, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:31:16,309 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17572.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:31:23,349 INFO [finetune.py:976] (0/7) Epoch 4, batch 400, loss[loss=0.2397, simple_loss=0.2912, pruned_loss=0.09412, over 4760.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.2951, pruned_loss=0.09314, over 829888.78 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:31:27,062 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-03-26 02:31:34,667 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 02:31:46,281 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:31:51,079 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.790e+02 1.987e+02 2.567e+02 5.687e+02, threshold=3.975e+02, percent-clipped=1.0 +2023-03-26 02:32:10,368 INFO [finetune.py:976] (0/7) Epoch 4, batch 450, loss[loss=0.2247, simple_loss=0.2824, pruned_loss=0.08352, over 4902.00 frames. ], tot_loss[loss=0.2402, simple_loss=0.2944, pruned_loss=0.09299, over 858000.77 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:33:00,232 INFO [finetune.py:976] (0/7) Epoch 4, batch 500, loss[loss=0.2694, simple_loss=0.311, pruned_loss=0.1139, over 4871.00 frames. ], tot_loss[loss=0.2377, simple_loss=0.2913, pruned_loss=0.09201, over 879749.04 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:33:10,451 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0227, 1.9214, 1.7627, 2.1174, 2.5445, 2.0806, 1.8375, 1.5255], + device='cuda:0'), covar=tensor([0.2298, 0.2191, 0.1943, 0.1788, 0.2067, 0.1217, 0.2624, 0.1873], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0207, 0.0197, 0.0182, 0.0232, 0.0172, 0.0213, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:33:11,010 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17700.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:24,346 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.264e+02 1.808e+02 2.091e+02 2.485e+02 4.480e+02, threshold=4.181e+02, percent-clipped=1.0 +2023-03-26 02:33:24,426 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:31,878 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:34,059 INFO [finetune.py:976] (0/7) Epoch 4, batch 550, loss[loss=0.2433, simple_loss=0.2939, pruned_loss=0.09629, over 4901.00 frames. ], tot_loss[loss=0.2362, simple_loss=0.2893, pruned_loss=0.09153, over 896149.66 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:33:37,769 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17739.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 02:33:40,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7633, 1.0404, 1.5534, 1.5434, 1.4532, 1.4172, 1.4239, 1.4846], + device='cuda:0'), covar=tensor([0.7346, 1.1450, 0.8997, 0.9954, 1.0407, 0.7686, 1.2775, 0.7904], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0253, 0.0256, 0.0265, 0.0243, 0.0218, 0.0279, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:34:03,942 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17761.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:34:17,632 INFO [finetune.py:976] (0/7) Epoch 4, batch 600, loss[loss=0.2088, simple_loss=0.2604, pruned_loss=0.07863, over 4762.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.2887, pruned_loss=0.09075, over 910176.47 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:34:22,560 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17791.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:34:41,399 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.729e+02 2.101e+02 2.480e+02 7.519e+02, threshold=4.202e+02, percent-clipped=3.0 +2023-03-26 02:34:50,414 INFO [finetune.py:976] (0/7) Epoch 4, batch 650, loss[loss=0.2247, simple_loss=0.2942, pruned_loss=0.07763, over 4803.00 frames. ], tot_loss[loss=0.2398, simple_loss=0.2934, pruned_loss=0.09313, over 919891.76 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:34:59,951 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:35:31,912 INFO [finetune.py:976] (0/7) Epoch 4, batch 700, loss[loss=0.2906, simple_loss=0.3142, pruned_loss=0.1335, over 3954.00 frames. ], tot_loss[loss=0.2409, simple_loss=0.2951, pruned_loss=0.09338, over 927903.05 frames. ], batch size: 17, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:35:46,359 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17893.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:36:18,031 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.804e+02 2.027e+02 2.461e+02 4.855e+02, threshold=4.055e+02, percent-clipped=2.0 +2023-03-26 02:36:34,757 INFO [finetune.py:976] (0/7) Epoch 4, batch 750, loss[loss=0.2661, simple_loss=0.319, pruned_loss=0.1066, over 4885.00 frames. ], tot_loss[loss=0.2434, simple_loss=0.2975, pruned_loss=0.09464, over 936108.02 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:36:49,408 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5268, 2.1438, 1.7077, 0.9244, 1.9390, 2.1292, 1.8629, 1.9387], + device='cuda:0'), covar=tensor([0.0860, 0.0956, 0.1653, 0.2369, 0.1603, 0.2329, 0.2297, 0.1002], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0204, 0.0190, 0.0218, 0.0211, 0.0219, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:37:08,564 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:37:30,394 INFO [finetune.py:976] (0/7) Epoch 4, batch 800, loss[loss=0.2553, simple_loss=0.3129, pruned_loss=0.09884, over 4889.00 frames. ], tot_loss[loss=0.2395, simple_loss=0.2941, pruned_loss=0.09244, over 939953.20 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:37:46,960 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-18000.pt +2023-03-26 02:38:05,563 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.784e+02 2.191e+02 2.808e+02 5.190e+02, threshold=4.382e+02, percent-clipped=3.0 +2023-03-26 02:38:05,685 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:08,553 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18021.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:09,865 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-03-26 02:38:20,834 INFO [finetune.py:976] (0/7) Epoch 4, batch 850, loss[loss=0.2633, simple_loss=0.3049, pruned_loss=0.1108, over 4822.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.2926, pruned_loss=0.09238, over 942451.18 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:38:26,592 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18039.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:38:26,708 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 02:38:39,163 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:43,425 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5360, 1.6663, 1.7679, 1.1577, 1.6058, 1.9208, 1.9488, 1.5417], + device='cuda:0'), covar=tensor([0.0903, 0.0553, 0.0402, 0.0512, 0.0404, 0.0541, 0.0273, 0.0496], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0159, 0.0117, 0.0138, 0.0134, 0.0123, 0.0147, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.8537e-05, 1.1809e-04, 8.5377e-05, 1.0129e-04, 9.7212e-05, 9.1131e-05, + 1.0991e-04, 1.0767e-04], device='cuda:0') +2023-03-26 02:38:45,698 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:57,811 INFO [finetune.py:976] (0/7) Epoch 4, batch 900, loss[loss=0.1841, simple_loss=0.2482, pruned_loss=0.06003, over 4840.00 frames. ], tot_loss[loss=0.235, simple_loss=0.2888, pruned_loss=0.09056, over 945681.29 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:38:59,689 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18086.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:39:00,271 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18087.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:39:19,669 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.720e+02 1.921e+02 2.312e+02 4.297e+02, threshold=3.842e+02, percent-clipped=0.0 +2023-03-26 02:39:35,995 INFO [finetune.py:976] (0/7) Epoch 4, batch 950, loss[loss=0.2442, simple_loss=0.2916, pruned_loss=0.09838, over 4823.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.2883, pruned_loss=0.09125, over 947400.59 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:40:07,904 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6269, 1.7110, 1.8733, 1.0948, 1.7875, 1.9834, 2.0288, 1.5724], + device='cuda:0'), covar=tensor([0.1041, 0.0693, 0.0377, 0.0629, 0.0356, 0.0601, 0.0300, 0.0700], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0158, 0.0117, 0.0138, 0.0133, 0.0122, 0.0147, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.8408e-05, 1.1778e-04, 8.5075e-05, 1.0144e-04, 9.6654e-05, 9.0824e-05, + 1.0967e-04, 1.0721e-04], device='cuda:0') +2023-03-26 02:40:22,944 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18174.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:40:23,588 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18175.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:40:29,236 INFO [finetune.py:976] (0/7) Epoch 4, batch 1000, loss[loss=0.268, simple_loss=0.3124, pruned_loss=0.1118, over 4904.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2911, pruned_loss=0.09175, over 951081.98 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:41:07,311 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.690e+02 2.099e+02 2.471e+02 3.966e+02, threshold=4.198e+02, percent-clipped=1.0 +2023-03-26 02:41:28,636 INFO [finetune.py:976] (0/7) Epoch 4, batch 1050, loss[loss=0.2713, simple_loss=0.3191, pruned_loss=0.1117, over 4800.00 frames. ], tot_loss[loss=0.2392, simple_loss=0.2938, pruned_loss=0.09229, over 951958.65 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:41:29,966 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:41:30,595 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:41:31,860 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9169, 1.4867, 1.6924, 1.6819, 1.5011, 1.4794, 1.6121, 1.6703], + device='cuda:0'), covar=tensor([0.9728, 1.3666, 1.0111, 1.3037, 1.3338, 1.0058, 1.6632, 0.9148], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0253, 0.0255, 0.0263, 0.0242, 0.0217, 0.0277, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 02:41:50,423 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2884, 1.3126, 1.3714, 0.7486, 1.4465, 1.4068, 1.3547, 1.2626], + device='cuda:0'), covar=tensor([0.0640, 0.0702, 0.0689, 0.0986, 0.0735, 0.0712, 0.0643, 0.1156], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0128, 0.0110, 0.0141, 0.0146, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:42:02,449 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6438, 1.6550, 1.6480, 1.0433, 1.6874, 1.9333, 1.9040, 1.5771], + device='cuda:0'), covar=tensor([0.1100, 0.0651, 0.0498, 0.0658, 0.0495, 0.0562, 0.0411, 0.0637], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0158, 0.0117, 0.0138, 0.0133, 0.0122, 0.0147, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.8208e-05, 1.1771e-04, 8.5318e-05, 1.0122e-04, 9.6518e-05, 9.0740e-05, + 1.0969e-04, 1.0730e-04], device='cuda:0') +2023-03-26 02:42:11,261 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2850, 3.7130, 3.8517, 4.1372, 3.9878, 3.7252, 4.3305, 1.4551], + device='cuda:0'), covar=tensor([0.0636, 0.0749, 0.0812, 0.0818, 0.1103, 0.1416, 0.0627, 0.4774], + device='cuda:0'), in_proj_covar=tensor([0.0362, 0.0244, 0.0279, 0.0295, 0.0342, 0.0287, 0.0309, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:42:18,623 INFO [finetune.py:976] (0/7) Epoch 4, batch 1100, loss[loss=0.2724, simple_loss=0.3228, pruned_loss=0.111, over 4890.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.2954, pruned_loss=0.09294, over 951846.06 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:42:49,104 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18316.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:42:50,749 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.851e+02 2.220e+02 2.754e+02 4.687e+02, threshold=4.440e+02, percent-clipped=1.0 +2023-03-26 02:43:08,651 INFO [finetune.py:976] (0/7) Epoch 4, batch 1150, loss[loss=0.2186, simple_loss=0.2805, pruned_loss=0.07838, over 4767.00 frames. ], tot_loss[loss=0.2398, simple_loss=0.2947, pruned_loss=0.09246, over 953211.81 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:43:25,363 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18356.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:03,369 INFO [finetune.py:976] (0/7) Epoch 4, batch 1200, loss[loss=0.204, simple_loss=0.2605, pruned_loss=0.07376, over 4817.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2921, pruned_loss=0.09131, over 953166.15 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:44:05,845 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:28,293 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18404.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:28,547 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 02:44:35,872 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 02:44:47,472 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.754e+02 2.082e+02 2.492e+02 3.668e+02, threshold=4.164e+02, percent-clipped=0.0 +2023-03-26 02:45:07,911 INFO [finetune.py:976] (0/7) Epoch 4, batch 1250, loss[loss=0.1807, simple_loss=0.2349, pruned_loss=0.06327, over 4732.00 frames. ], tot_loss[loss=0.2345, simple_loss=0.2892, pruned_loss=0.08996, over 954147.46 frames. ], batch size: 23, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:45:09,084 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18434.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:45:59,526 INFO [finetune.py:976] (0/7) Epoch 4, batch 1300, loss[loss=0.1981, simple_loss=0.2467, pruned_loss=0.07474, over 4787.00 frames. ], tot_loss[loss=0.2311, simple_loss=0.2856, pruned_loss=0.08835, over 954781.47 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:46:10,655 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18490.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:22,659 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18508.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:46:22,758 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-03-26 02:46:29,709 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.658e+02 2.008e+02 2.632e+02 4.281e+02, threshold=4.017e+02, percent-clipped=1.0 +2023-03-26 02:46:36,931 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:37,485 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18531.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:41,947 INFO [finetune.py:976] (0/7) Epoch 4, batch 1350, loss[loss=0.2382, simple_loss=0.3041, pruned_loss=0.08613, over 4821.00 frames. ], tot_loss[loss=0.2331, simple_loss=0.287, pruned_loss=0.08956, over 955259.25 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:46:55,193 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:47:11,773 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18569.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:47:25,813 INFO [finetune.py:976] (0/7) Epoch 4, batch 1400, loss[loss=0.1937, simple_loss=0.2569, pruned_loss=0.06528, over 4791.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.2916, pruned_loss=0.09098, over 955995.85 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:47:26,499 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4390, 1.4521, 1.4840, 0.8285, 1.6100, 1.6342, 1.4768, 1.4305], + device='cuda:0'), covar=tensor([0.0696, 0.0788, 0.0738, 0.1072, 0.0743, 0.0740, 0.0756, 0.1169], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0133, 0.0144, 0.0129, 0.0110, 0.0142, 0.0147, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:47:31,082 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 02:47:53,797 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:47:55,463 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.863e+02 2.207e+02 2.712e+02 5.337e+02, threshold=4.415e+02, percent-clipped=2.0 +2023-03-26 02:48:10,032 INFO [finetune.py:976] (0/7) Epoch 4, batch 1450, loss[loss=0.2326, simple_loss=0.2871, pruned_loss=0.08903, over 4926.00 frames. ], tot_loss[loss=0.2382, simple_loss=0.2935, pruned_loss=0.09142, over 956226.12 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:48:43,321 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18664.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:48:49,419 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:48:56,991 INFO [finetune.py:976] (0/7) Epoch 4, batch 1500, loss[loss=0.2976, simple_loss=0.3315, pruned_loss=0.1318, over 4860.00 frames. ], tot_loss[loss=0.2408, simple_loss=0.2954, pruned_loss=0.09308, over 954880.48 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:49:36,301 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.377e+02 1.810e+02 2.122e+02 2.509e+02 6.153e+02, threshold=4.245e+02, percent-clipped=1.0 +2023-03-26 02:49:37,009 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6599, 1.5141, 1.5643, 1.5986, 1.1026, 3.5719, 1.3796, 1.9462], + device='cuda:0'), covar=tensor([0.3690, 0.2643, 0.2235, 0.2472, 0.2147, 0.0163, 0.2793, 0.1455], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 02:49:54,648 INFO [finetune.py:976] (0/7) Epoch 4, batch 1550, loss[loss=0.293, simple_loss=0.3373, pruned_loss=0.1244, over 4886.00 frames. ], tot_loss[loss=0.2401, simple_loss=0.2949, pruned_loss=0.09267, over 953333.26 frames. ], batch size: 32, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:49:55,956 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:50:21,040 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0149, 1.9113, 1.5366, 1.9358, 1.9519, 1.6810, 2.4282, 2.0180], + device='cuda:0'), covar=tensor([0.1922, 0.3887, 0.4472, 0.4318, 0.3260, 0.2278, 0.4062, 0.2692], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0193, 0.0237, 0.0253, 0.0222, 0.0185, 0.0209, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:50:34,584 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 02:50:39,061 INFO [finetune.py:976] (0/7) Epoch 4, batch 1600, loss[loss=0.1466, simple_loss=0.2212, pruned_loss=0.03598, over 4782.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.292, pruned_loss=0.09183, over 953482.40 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:51:19,525 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.645e+02 2.011e+02 2.399e+02 5.772e+02, threshold=4.021e+02, percent-clipped=1.0 +2023-03-26 02:51:23,307 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5740, 1.4916, 1.9556, 1.9248, 1.8066, 4.1708, 1.3397, 1.7217], + device='cuda:0'), covar=tensor([0.0970, 0.1739, 0.1192, 0.0978, 0.1569, 0.0141, 0.1535, 0.1745], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0079, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:51:30,345 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18830.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:51:30,937 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:51:32,084 INFO [finetune.py:976] (0/7) Epoch 4, batch 1650, loss[loss=0.2258, simple_loss=0.2826, pruned_loss=0.08449, over 4841.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2873, pruned_loss=0.0893, over 952959.12 frames. ], batch size: 30, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:51:32,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3975, 2.1860, 2.8415, 1.7248, 2.5937, 2.5557, 2.1712, 2.8658], + device='cuda:0'), covar=tensor([0.1756, 0.1977, 0.1439, 0.2561, 0.0974, 0.1895, 0.2324, 0.1081], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0206, 0.0203, 0.0197, 0.0182, 0.0226, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:51:38,067 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 02:51:48,108 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18846.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:05,189 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18864.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:52:23,038 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:23,625 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18879.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:25,972 INFO [finetune.py:976] (0/7) Epoch 4, batch 1700, loss[loss=0.2727, simple_loss=0.3255, pruned_loss=0.1099, over 4905.00 frames. ], tot_loss[loss=0.2316, simple_loss=0.2861, pruned_loss=0.08851, over 955725.21 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:52:26,156 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.30 vs. limit=5.0 +2023-03-26 02:52:26,180 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-26 02:53:00,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.767e+02 2.149e+02 2.599e+02 5.673e+02, threshold=4.299e+02, percent-clipped=2.0 +2023-03-26 02:53:06,481 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-03-26 02:53:07,525 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18930.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:53:09,236 INFO [finetune.py:976] (0/7) Epoch 4, batch 1750, loss[loss=0.3079, simple_loss=0.3569, pruned_loss=0.1295, over 4867.00 frames. ], tot_loss[loss=0.2355, simple_loss=0.2901, pruned_loss=0.09042, over 955507.52 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:53:22,087 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.0268, 4.3234, 4.5427, 4.8795, 4.7279, 4.5015, 5.1123, 1.5142], + device='cuda:0'), covar=tensor([0.0572, 0.0685, 0.0629, 0.0728, 0.1107, 0.1278, 0.0493, 0.5222], + device='cuda:0'), in_proj_covar=tensor([0.0361, 0.0244, 0.0277, 0.0293, 0.0339, 0.0287, 0.0310, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:53:52,364 INFO [finetune.py:976] (0/7) Epoch 4, batch 1800, loss[loss=0.2811, simple_loss=0.3339, pruned_loss=0.1141, over 4808.00 frames. ], tot_loss[loss=0.2395, simple_loss=0.2945, pruned_loss=0.09226, over 958019.15 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:53:57,493 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:54:32,259 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.317e+02 1.869e+02 2.115e+02 2.590e+02 5.981e+02, threshold=4.230e+02, percent-clipped=1.0 +2023-03-26 02:54:40,545 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0375, 1.7463, 2.4329, 1.5535, 2.2367, 2.2820, 1.8073, 2.5483], + device='cuda:0'), covar=tensor([0.1551, 0.2134, 0.1421, 0.1972, 0.0862, 0.1672, 0.2510, 0.0865], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0207, 0.0204, 0.0198, 0.0182, 0.0227, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:54:50,216 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:54:52,012 INFO [finetune.py:976] (0/7) Epoch 4, batch 1850, loss[loss=0.2289, simple_loss=0.286, pruned_loss=0.08583, over 4816.00 frames. ], tot_loss[loss=0.2393, simple_loss=0.2945, pruned_loss=0.09204, over 957690.80 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:55:19,739 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0009, 1.4532, 1.0672, 1.7891, 2.3425, 1.4505, 1.6751, 1.9153], + device='cuda:0'), covar=tensor([0.1417, 0.1967, 0.2163, 0.1170, 0.1830, 0.2016, 0.1374, 0.1908], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0100, 0.0118, 0.0095, 0.0126, 0.0098, 0.0101, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 02:55:39,892 INFO [finetune.py:976] (0/7) Epoch 4, batch 1900, loss[loss=0.2142, simple_loss=0.2843, pruned_loss=0.07206, over 4894.00 frames. ], tot_loss[loss=0.2398, simple_loss=0.2952, pruned_loss=0.09222, over 957529.89 frames. ], batch size: 43, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:55:56,800 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3358, 1.4140, 1.3662, 1.6243, 1.6636, 3.0144, 1.3906, 1.5577], + device='cuda:0'), covar=tensor([0.1106, 0.1754, 0.1126, 0.1013, 0.1525, 0.0286, 0.1459, 0.1706], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 02:56:12,754 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 1.752e+02 2.082e+02 2.658e+02 3.786e+02, threshold=4.164e+02, percent-clipped=0.0 +2023-03-26 02:56:25,558 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19129.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:56:33,399 INFO [finetune.py:976] (0/7) Epoch 4, batch 1950, loss[loss=0.2925, simple_loss=0.3229, pruned_loss=0.131, over 4348.00 frames. ], tot_loss[loss=0.2369, simple_loss=0.2922, pruned_loss=0.0908, over 956238.95 frames. ], batch size: 66, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:56:41,420 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19146.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:56:53,386 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19164.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:57:09,562 INFO [finetune.py:976] (0/7) Epoch 4, batch 2000, loss[loss=0.243, simple_loss=0.2919, pruned_loss=0.09707, over 4806.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.289, pruned_loss=0.0898, over 958063.32 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:57:14,700 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19190.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:57:17,079 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:57:22,294 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 02:57:34,646 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19212.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:57:39,346 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.251e+02 1.681e+02 2.009e+02 2.396e+02 5.395e+02, threshold=4.017e+02, percent-clipped=3.0 +2023-03-26 02:57:42,682 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 02:57:49,438 INFO [finetune.py:976] (0/7) Epoch 4, batch 2050, loss[loss=0.2336, simple_loss=0.2763, pruned_loss=0.0955, over 4919.00 frames. ], tot_loss[loss=0.2299, simple_loss=0.2846, pruned_loss=0.0876, over 959322.86 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:58:07,579 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9740, 1.5564, 2.2350, 3.5476, 2.6036, 2.6329, 0.8673, 2.8392], + device='cuda:0'), covar=tensor([0.1732, 0.1724, 0.1377, 0.0550, 0.0742, 0.1803, 0.2084, 0.0562], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0166, 0.0105, 0.0144, 0.0129, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 02:58:19,734 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7864, 1.5731, 1.5458, 1.3852, 1.8911, 1.5417, 1.7727, 1.7470], + device='cuda:0'), covar=tensor([0.1920, 0.3197, 0.4170, 0.3222, 0.2925, 0.2069, 0.3072, 0.2462], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0195, 0.0238, 0.0255, 0.0223, 0.0186, 0.0210, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:58:31,806 INFO [finetune.py:976] (0/7) Epoch 4, batch 2100, loss[loss=0.2167, simple_loss=0.2935, pruned_loss=0.06995, over 4907.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2851, pruned_loss=0.08825, over 958861.41 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:58:34,245 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19286.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:58:46,647 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9692, 1.4625, 1.7742, 1.6508, 1.5668, 1.5927, 1.6296, 1.7332], + device='cuda:0'), covar=tensor([0.7007, 0.9667, 0.8343, 0.9611, 0.9712, 0.7536, 1.2584, 0.7159], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0253, 0.0256, 0.0264, 0.0243, 0.0218, 0.0278, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:58:56,630 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5239, 2.4007, 2.3508, 2.6401, 2.2570, 4.9672, 2.2789, 2.9015], + device='cuda:0'), covar=tensor([0.2775, 0.2111, 0.1650, 0.1778, 0.1366, 0.0067, 0.1969, 0.1069], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0112, 0.0117, 0.0120, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 02:59:09,861 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.737e+02 2.057e+02 2.371e+02 3.601e+02, threshold=4.115e+02, percent-clipped=0.0 +2023-03-26 02:59:20,164 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 02:59:25,759 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:59:27,977 INFO [finetune.py:976] (0/7) Epoch 4, batch 2150, loss[loss=0.2836, simple_loss=0.3342, pruned_loss=0.1165, over 4832.00 frames. ], tot_loss[loss=0.2359, simple_loss=0.2902, pruned_loss=0.09076, over 955423.40 frames. ], batch size: 49, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:59:29,867 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8984, 1.3287, 1.6830, 1.6207, 1.5064, 1.5309, 1.5658, 1.5594], + device='cuda:0'), covar=tensor([0.7819, 1.0375, 0.8703, 0.9891, 1.0562, 0.8050, 1.2276, 0.8144], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0253, 0.0257, 0.0264, 0.0243, 0.0218, 0.0278, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 02:59:30,393 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6871, 3.5739, 3.4927, 1.7775, 3.7405, 2.7585, 0.8867, 2.6154], + device='cuda:0'), covar=tensor([0.2952, 0.1638, 0.1499, 0.3298, 0.0922, 0.0942, 0.4428, 0.1369], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0169, 0.0161, 0.0128, 0.0154, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 03:00:10,912 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19378.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:00:13,945 INFO [finetune.py:976] (0/7) Epoch 4, batch 2200, loss[loss=0.2522, simple_loss=0.2991, pruned_loss=0.1027, over 4883.00 frames. ], tot_loss[loss=0.237, simple_loss=0.292, pruned_loss=0.09099, over 955409.63 frames. ], batch size: 32, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:00:32,875 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 03:00:33,398 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:01:05,356 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.822e+02 2.092e+02 2.549e+02 4.918e+02, threshold=4.184e+02, percent-clipped=2.0 +2023-03-26 03:01:14,912 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4169, 2.9049, 2.7809, 1.2869, 2.9855, 2.1290, 0.6174, 1.8424], + device='cuda:0'), covar=tensor([0.2237, 0.1978, 0.1932, 0.3327, 0.1250, 0.1237, 0.4188, 0.1656], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0168, 0.0161, 0.0128, 0.0154, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 03:01:19,364 INFO [finetune.py:976] (0/7) Epoch 4, batch 2250, loss[loss=0.2753, simple_loss=0.3323, pruned_loss=0.1092, over 4727.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.2928, pruned_loss=0.09141, over 953620.29 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:01:45,255 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 03:01:46,318 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:01:47,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0175, 1.6995, 2.3282, 3.7747, 2.7862, 2.6336, 0.8597, 2.9692], + device='cuda:0'), covar=tensor([0.1791, 0.1606, 0.1408, 0.0557, 0.0788, 0.1652, 0.2134, 0.0573], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0120, 0.0138, 0.0167, 0.0105, 0.0144, 0.0130, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 03:02:09,425 INFO [finetune.py:976] (0/7) Epoch 4, batch 2300, loss[loss=0.2876, simple_loss=0.3385, pruned_loss=0.1184, over 4842.00 frames. ], tot_loss[loss=0.2364, simple_loss=0.2922, pruned_loss=0.09033, over 954905.53 frames. ], batch size: 49, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:02:12,903 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19485.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 03:02:21,458 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4273, 1.4361, 1.5392, 1.7221, 1.6025, 3.2624, 1.4782, 1.6102], + device='cuda:0'), covar=tensor([0.1143, 0.1820, 0.1281, 0.1125, 0.1716, 0.0253, 0.1477, 0.1855], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:02:37,772 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.717e+02 2.025e+02 2.639e+02 4.089e+02, threshold=4.050e+02, percent-clipped=0.0 +2023-03-26 03:02:53,438 INFO [finetune.py:976] (0/7) Epoch 4, batch 2350, loss[loss=0.1564, simple_loss=0.213, pruned_loss=0.04991, over 4685.00 frames. ], tot_loss[loss=0.2331, simple_loss=0.2888, pruned_loss=0.08866, over 954884.13 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:03:37,789 INFO [finetune.py:976] (0/7) Epoch 4, batch 2400, loss[loss=0.2209, simple_loss=0.2671, pruned_loss=0.08736, over 4875.00 frames. ], tot_loss[loss=0.2299, simple_loss=0.2856, pruned_loss=0.08706, over 956048.74 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:03:40,239 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:00,384 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6822, 1.5344, 1.4662, 1.7012, 2.0552, 1.7535, 1.0106, 1.4243], + device='cuda:0'), covar=tensor([0.2542, 0.2405, 0.2180, 0.1957, 0.1904, 0.1249, 0.3233, 0.2000], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0208, 0.0197, 0.0183, 0.0232, 0.0173, 0.0213, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:04:10,039 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19618.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:10,537 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.651e+02 1.936e+02 2.390e+02 3.810e+02, threshold=3.872e+02, percent-clipped=0.0 +2023-03-26 03:04:14,405 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2691, 2.1237, 1.7295, 2.4407, 2.2705, 1.9079, 2.8366, 2.2553], + device='cuda:0'), covar=tensor([0.1837, 0.4162, 0.4573, 0.4177, 0.3278, 0.2191, 0.4657, 0.2845], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0194, 0.0237, 0.0254, 0.0223, 0.0186, 0.0210, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:04:19,623 INFO [finetune.py:976] (0/7) Epoch 4, batch 2450, loss[loss=0.2713, simple_loss=0.3142, pruned_loss=0.1142, over 4811.00 frames. ], tot_loss[loss=0.2264, simple_loss=0.2819, pruned_loss=0.08548, over 955366.05 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:04:20,303 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19634.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:22,675 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8838, 1.2771, 1.6757, 1.6014, 1.4692, 1.4936, 1.5408, 1.5217], + device='cuda:0'), covar=tensor([0.6573, 1.0139, 0.8250, 0.9721, 1.0436, 0.7552, 1.1742, 0.7791], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0253, 0.0258, 0.0265, 0.0243, 0.0219, 0.0279, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:04:25,746 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1823, 1.6028, 1.7030, 1.8213, 1.5786, 1.6892, 1.7606, 1.7227], + device='cuda:0'), covar=tensor([1.0321, 1.5225, 1.2512, 1.4089, 1.7779, 1.1197, 1.9276, 1.1377], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0253, 0.0258, 0.0264, 0.0243, 0.0219, 0.0279, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:04:59,953 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:05:02,277 INFO [finetune.py:976] (0/7) Epoch 4, batch 2500, loss[loss=0.2158, simple_loss=0.2908, pruned_loss=0.07043, over 4830.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2848, pruned_loss=0.08738, over 953720.97 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:05:02,987 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9878, 1.8874, 1.8585, 1.9869, 1.4470, 3.5864, 1.6732, 2.2583], + device='cuda:0'), covar=tensor([0.2966, 0.2365, 0.1849, 0.2087, 0.1827, 0.0199, 0.2432, 0.1183], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0100, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 03:05:30,656 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.296e+02 1.721e+02 2.075e+02 2.470e+02 4.533e+02, threshold=4.150e+02, percent-clipped=4.0 +2023-03-26 03:05:40,523 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1308, 1.3735, 0.8834, 2.0461, 2.3121, 1.5556, 1.7088, 2.0678], + device='cuda:0'), covar=tensor([0.1478, 0.2311, 0.2364, 0.1198, 0.2041, 0.2259, 0.1459, 0.2024], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0100, 0.0118, 0.0095, 0.0126, 0.0098, 0.0101, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 03:05:45,256 INFO [finetune.py:976] (0/7) Epoch 4, batch 2550, loss[loss=0.2158, simple_loss=0.2878, pruned_loss=0.07194, over 4788.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.2878, pruned_loss=0.08749, over 954145.61 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:05:58,469 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19750.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:06:31,492 INFO [finetune.py:976] (0/7) Epoch 4, batch 2600, loss[loss=0.2655, simple_loss=0.2977, pruned_loss=0.1166, over 4746.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2894, pruned_loss=0.08824, over 954738.34 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:06:33,302 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19785.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 03:06:34,094 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-03-26 03:07:15,443 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4285, 2.2249, 1.9834, 1.5601, 2.4273, 2.7913, 2.4514, 2.1082], + device='cuda:0'), covar=tensor([0.0236, 0.0413, 0.0511, 0.0418, 0.0284, 0.0488, 0.0410, 0.0453], + device='cuda:0'), in_proj_covar=tensor([0.0085, 0.0114, 0.0138, 0.0118, 0.0104, 0.0099, 0.0092, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.6606e-05, 9.0303e-05, 1.1115e-04, 9.3248e-05, 8.2552e-05, 7.3940e-05, + 7.0877e-05, 8.5557e-05], device='cuda:0') +2023-03-26 03:07:15,921 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.335e+02 1.804e+02 2.222e+02 2.871e+02 4.406e+02, threshold=4.445e+02, percent-clipped=2.0 +2023-03-26 03:07:33,334 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6356, 1.4610, 1.9716, 1.7724, 1.6648, 4.0889, 1.2664, 1.6421], + device='cuda:0'), covar=tensor([0.1336, 0.2235, 0.1451, 0.1347, 0.1928, 0.0228, 0.2133, 0.2372], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0081, 0.0078, 0.0080, 0.0093, 0.0084, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:07:34,472 INFO [finetune.py:976] (0/7) Epoch 4, batch 2650, loss[loss=0.2259, simple_loss=0.2591, pruned_loss=0.09632, over 4091.00 frames. ], tot_loss[loss=0.2348, simple_loss=0.2908, pruned_loss=0.08942, over 952819.72 frames. ], batch size: 17, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:07:34,543 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19833.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 03:07:44,350 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5429, 1.3542, 1.7581, 3.0140, 2.0754, 2.2187, 0.9472, 2.4290], + device='cuda:0'), covar=tensor([0.2194, 0.2105, 0.1730, 0.0970, 0.1070, 0.1605, 0.2270, 0.0917], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0105, 0.0143, 0.0130, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 03:08:34,281 INFO [finetune.py:976] (0/7) Epoch 4, batch 2700, loss[loss=0.1806, simple_loss=0.2486, pruned_loss=0.05625, over 4862.00 frames. ], tot_loss[loss=0.2321, simple_loss=0.2884, pruned_loss=0.08792, over 953854.43 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:09:16,013 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.716e+02 2.005e+02 2.489e+02 3.950e+02, threshold=4.009e+02, percent-clipped=0.0 +2023-03-26 03:09:22,973 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7409, 1.6083, 1.9938, 1.8402, 1.8558, 4.3010, 1.5664, 1.8407], + device='cuda:0'), covar=tensor([0.1070, 0.1784, 0.1220, 0.1159, 0.1571, 0.0223, 0.1575, 0.1792], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:09:26,957 INFO [finetune.py:976] (0/7) Epoch 4, batch 2750, loss[loss=0.2365, simple_loss=0.2921, pruned_loss=0.09041, over 4830.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2852, pruned_loss=0.08718, over 952344.35 frames. ], batch size: 30, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:09:40,295 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-03-26 03:09:52,998 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4350, 1.3951, 1.6250, 1.6664, 1.5654, 3.2833, 1.2781, 1.5524], + device='cuda:0'), covar=tensor([0.1135, 0.1885, 0.1339, 0.1196, 0.1626, 0.0270, 0.1744, 0.1946], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:09:54,788 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19974.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:10:00,205 INFO [finetune.py:976] (0/7) Epoch 4, batch 2800, loss[loss=0.2036, simple_loss=0.2508, pruned_loss=0.0782, over 4940.00 frames. ], tot_loss[loss=0.2266, simple_loss=0.2815, pruned_loss=0.08581, over 950720.66 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:10:11,550 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-20000.pt +2023-03-26 03:10:24,998 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.781e+02 2.091e+02 2.518e+02 3.954e+02, threshold=4.183e+02, percent-clipped=0.0 +2023-03-26 03:10:26,215 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:10:34,544 INFO [finetune.py:976] (0/7) Epoch 4, batch 2850, loss[loss=0.2389, simple_loss=0.2845, pruned_loss=0.09664, over 4912.00 frames. ], tot_loss[loss=0.2264, simple_loss=0.2811, pruned_loss=0.08589, over 952226.44 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:10:45,846 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:21,568 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:22,670 INFO [finetune.py:976] (0/7) Epoch 4, batch 2900, loss[loss=0.2344, simple_loss=0.2827, pruned_loss=0.093, over 4172.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2843, pruned_loss=0.08726, over 948508.86 frames. ], batch size: 18, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:11:31,803 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:42,682 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:56,201 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20112.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:12:03,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7415, 1.5279, 1.3106, 1.4094, 1.4747, 1.4080, 1.4353, 2.1935], + device='cuda:0'), covar=tensor([0.8196, 0.7717, 0.6453, 0.7835, 0.6841, 0.4347, 0.7785, 0.2787], + device='cuda:0'), in_proj_covar=tensor([0.0278, 0.0253, 0.0220, 0.0285, 0.0237, 0.0198, 0.0241, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 03:12:05,858 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.342e+02 1.862e+02 2.201e+02 2.748e+02 4.534e+02, threshold=4.402e+02, percent-clipped=1.0 +2023-03-26 03:12:16,533 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 03:12:27,411 INFO [finetune.py:976] (0/7) Epoch 4, batch 2950, loss[loss=0.1919, simple_loss=0.2521, pruned_loss=0.06584, over 4709.00 frames. ], tot_loss[loss=0.2323, simple_loss=0.2875, pruned_loss=0.0886, over 948571.43 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:12:48,260 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:13:10,797 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20173.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:13:18,826 INFO [finetune.py:976] (0/7) Epoch 4, batch 3000, loss[loss=0.223, simple_loss=0.2843, pruned_loss=0.08079, over 4829.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.2897, pruned_loss=0.08945, over 948468.63 frames. ], batch size: 47, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:13:18,827 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 03:13:35,068 INFO [finetune.py:1010] (0/7) Epoch 4, validation: loss=0.169, simple_loss=0.2409, pruned_loss=0.04857, over 2265189.00 frames. +2023-03-26 03:13:35,069 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 03:13:58,692 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:13:59,284 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8365, 1.1772, 0.9696, 1.7310, 2.0337, 1.3644, 1.4831, 1.6709], + device='cuda:0'), covar=tensor([0.1357, 0.2194, 0.2165, 0.1205, 0.2114, 0.2321, 0.1435, 0.2060], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0100, 0.0119, 0.0095, 0.0127, 0.0098, 0.0102, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 03:14:18,154 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.917e+02 2.134e+02 2.804e+02 4.274e+02, threshold=4.268e+02, percent-clipped=0.0 +2023-03-26 03:14:27,242 INFO [finetune.py:976] (0/7) Epoch 4, batch 3050, loss[loss=0.2227, simple_loss=0.2856, pruned_loss=0.07995, over 4907.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.2906, pruned_loss=0.08961, over 947771.75 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:14:36,506 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.9229, 4.2976, 4.4994, 4.7280, 4.6535, 4.4233, 5.0463, 1.5636], + device='cuda:0'), covar=tensor([0.0674, 0.0762, 0.0706, 0.0789, 0.1138, 0.1376, 0.0553, 0.5056], + device='cuda:0'), in_proj_covar=tensor([0.0358, 0.0243, 0.0277, 0.0292, 0.0338, 0.0285, 0.0308, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:15:05,422 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 03:15:06,981 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:15:16,632 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20274.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:15:24,345 INFO [finetune.py:976] (0/7) Epoch 4, batch 3100, loss[loss=0.2067, simple_loss=0.2705, pruned_loss=0.07146, over 4817.00 frames. ], tot_loss[loss=0.2305, simple_loss=0.2865, pruned_loss=0.08726, over 949593.23 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:15:27,089 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-03-26 03:16:01,239 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.549e+02 1.969e+02 2.570e+02 5.632e+02, threshold=3.937e+02, percent-clipped=1.0 +2023-03-26 03:16:03,109 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:16:14,157 INFO [finetune.py:976] (0/7) Epoch 4, batch 3150, loss[loss=0.1754, simple_loss=0.2369, pruned_loss=0.05698, over 4758.00 frames. ], tot_loss[loss=0.2288, simple_loss=0.2837, pruned_loss=0.0869, over 950982.17 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:16:56,932 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:01,160 INFO [finetune.py:976] (0/7) Epoch 4, batch 3200, loss[loss=0.2042, simple_loss=0.2656, pruned_loss=0.07137, over 4804.00 frames. ], tot_loss[loss=0.226, simple_loss=0.2807, pruned_loss=0.08562, over 952123.83 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:17:04,845 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:40,191 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:40,661 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.625e+02 1.959e+02 2.342e+02 5.079e+02, threshold=3.919e+02, percent-clipped=3.0 +2023-03-26 03:17:58,484 INFO [finetune.py:976] (0/7) Epoch 4, batch 3250, loss[loss=0.2304, simple_loss=0.29, pruned_loss=0.08542, over 4925.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2811, pruned_loss=0.08569, over 952186.02 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:18:06,354 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20445.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:08,865 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20449.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:21,888 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20468.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:29,601 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:31,929 INFO [finetune.py:976] (0/7) Epoch 4, batch 3300, loss[loss=0.2252, simple_loss=0.279, pruned_loss=0.08567, over 4782.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2847, pruned_loss=0.08706, over 951342.85 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:19:09,458 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.750e+02 2.039e+02 2.534e+02 4.074e+02, threshold=4.078e+02, percent-clipped=2.0 +2023-03-26 03:19:18,597 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.08 vs. limit=5.0 +2023-03-26 03:19:29,493 INFO [finetune.py:976] (0/7) Epoch 4, batch 3350, loss[loss=0.2274, simple_loss=0.2766, pruned_loss=0.08906, over 4900.00 frames. ], tot_loss[loss=0.2312, simple_loss=0.287, pruned_loss=0.08776, over 951379.66 frames. ], batch size: 32, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:19:59,020 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20560.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:20:17,266 INFO [finetune.py:976] (0/7) Epoch 4, batch 3400, loss[loss=0.2395, simple_loss=0.292, pruned_loss=0.09344, over 4922.00 frames. ], tot_loss[loss=0.2322, simple_loss=0.2883, pruned_loss=0.08802, over 951324.97 frames. ], batch size: 42, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:20:20,414 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:20:46,762 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.735e+02 2.048e+02 2.538e+02 3.974e+02, threshold=4.096e+02, percent-clipped=0.0 +2023-03-26 03:21:05,097 INFO [finetune.py:976] (0/7) Epoch 4, batch 3450, loss[loss=0.2199, simple_loss=0.2803, pruned_loss=0.07969, over 4890.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2861, pruned_loss=0.08634, over 952099.68 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:21:16,898 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3102, 2.9310, 3.0562, 3.2527, 3.0991, 2.9182, 3.3555, 0.9367], + device='cuda:0'), covar=tensor([0.1068, 0.0893, 0.1002, 0.1001, 0.1579, 0.1640, 0.1040, 0.5045], + device='cuda:0'), in_proj_covar=tensor([0.0356, 0.0242, 0.0276, 0.0291, 0.0336, 0.0283, 0.0305, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:21:22,508 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:21:45,701 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20676.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:21:51,858 INFO [finetune.py:976] (0/7) Epoch 4, batch 3500, loss[loss=0.17, simple_loss=0.2377, pruned_loss=0.05111, over 4750.00 frames. ], tot_loss[loss=0.2289, simple_loss=0.2848, pruned_loss=0.08655, over 950931.55 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:21:54,274 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6308, 1.4711, 1.4110, 1.4922, 1.1809, 3.3448, 1.4018, 1.8581], + device='cuda:0'), covar=tensor([0.4136, 0.3038, 0.2466, 0.2949, 0.1923, 0.0276, 0.2733, 0.1412], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0112, 0.0117, 0.0120, 0.0117, 0.0097, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 03:22:31,509 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.236e+02 1.688e+02 2.022e+02 2.523e+02 5.341e+02, threshold=4.043e+02, percent-clipped=2.0 +2023-03-26 03:22:35,121 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20724.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:22:37,697 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.01 vs. limit=5.0 +2023-03-26 03:22:43,660 INFO [finetune.py:976] (0/7) Epoch 4, batch 3550, loss[loss=0.237, simple_loss=0.285, pruned_loss=0.09452, over 4744.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.2829, pruned_loss=0.08633, over 952675.30 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:22:56,117 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:22:56,768 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:24,106 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20768.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:33,432 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:39,155 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-03-26 03:23:41,428 INFO [finetune.py:976] (0/7) Epoch 4, batch 3600, loss[loss=0.2482, simple_loss=0.2962, pruned_loss=0.1001, over 4856.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2811, pruned_loss=0.08565, over 954180.32 frames. ], batch size: 44, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:23:52,902 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:24:24,357 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:24:31,499 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.828e+01 1.831e+02 2.152e+02 2.506e+02 5.159e+02, threshold=4.304e+02, percent-clipped=1.0 +2023-03-26 03:24:50,041 INFO [finetune.py:976] (0/7) Epoch 4, batch 3650, loss[loss=0.2946, simple_loss=0.3388, pruned_loss=0.1252, over 4743.00 frames. ], tot_loss[loss=0.2292, simple_loss=0.2841, pruned_loss=0.08713, over 952822.97 frames. ], batch size: 59, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:25:24,783 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20860.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:25:52,755 INFO [finetune.py:976] (0/7) Epoch 4, batch 3700, loss[loss=0.2347, simple_loss=0.2887, pruned_loss=0.0903, over 4774.00 frames. ], tot_loss[loss=0.2332, simple_loss=0.2885, pruned_loss=0.08895, over 955312.39 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:26:17,147 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20908.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:26:24,267 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.710e+02 2.022e+02 2.626e+02 5.956e+02, threshold=4.044e+02, percent-clipped=2.0 +2023-03-26 03:26:27,141 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7271, 1.5970, 1.2278, 1.3886, 1.5391, 1.4365, 1.4793, 2.2414], + device='cuda:0'), covar=tensor([0.8253, 0.8476, 0.6428, 0.8655, 0.7033, 0.4605, 0.8056, 0.2976], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0254, 0.0220, 0.0284, 0.0237, 0.0197, 0.0241, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 03:26:34,612 INFO [finetune.py:976] (0/7) Epoch 4, batch 3750, loss[loss=0.1905, simple_loss=0.2327, pruned_loss=0.07414, over 3997.00 frames. ], tot_loss[loss=0.2338, simple_loss=0.2897, pruned_loss=0.08898, over 953104.14 frames. ], batch size: 17, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:26:46,551 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:27:33,087 INFO [finetune.py:976] (0/7) Epoch 4, batch 3800, loss[loss=0.2588, simple_loss=0.3086, pruned_loss=0.1045, over 4157.00 frames. ], tot_loss[loss=0.2325, simple_loss=0.2889, pruned_loss=0.08806, over 952208.40 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:28:14,227 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.718e+02 1.983e+02 2.372e+02 3.493e+02, threshold=3.966e+02, percent-clipped=0.0 +2023-03-26 03:28:29,875 INFO [finetune.py:976] (0/7) Epoch 4, batch 3850, loss[loss=0.2453, simple_loss=0.3054, pruned_loss=0.09266, over 4867.00 frames. ], tot_loss[loss=0.2305, simple_loss=0.2869, pruned_loss=0.08705, over 952831.87 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:28:37,722 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:28:45,318 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:08,434 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:19,707 INFO [finetune.py:976] (0/7) Epoch 4, batch 3900, loss[loss=0.2358, simple_loss=0.2896, pruned_loss=0.09099, over 4934.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2854, pruned_loss=0.08713, over 954483.03 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:29:26,803 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:43,725 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:46,752 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:47,847 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.709e+02 1.984e+02 2.370e+02 6.134e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-03-26 03:29:49,142 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:59,343 INFO [finetune.py:976] (0/7) Epoch 4, batch 3950, loss[loss=0.234, simple_loss=0.2811, pruned_loss=0.09351, over 4821.00 frames. ], tot_loss[loss=0.2261, simple_loss=0.2815, pruned_loss=0.08531, over 953202.95 frames. ], batch size: 25, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:30:28,102 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 03:30:47,484 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:30:50,288 INFO [finetune.py:976] (0/7) Epoch 4, batch 4000, loss[loss=0.218, simple_loss=0.2761, pruned_loss=0.07998, over 4820.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2811, pruned_loss=0.08509, over 954169.21 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:30:56,466 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 03:31:13,532 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 03:31:14,605 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9629, 1.9660, 1.5162, 2.1215, 2.0212, 1.6602, 2.5979, 2.0407], + device='cuda:0'), covar=tensor([0.2132, 0.3996, 0.4445, 0.4153, 0.3349, 0.2226, 0.3893, 0.2625], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0196, 0.0238, 0.0257, 0.0227, 0.0188, 0.0212, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:31:19,051 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.740e+02 2.148e+02 2.427e+02 4.357e+02, threshold=4.296e+02, percent-clipped=1.0 +2023-03-26 03:31:27,483 INFO [finetune.py:976] (0/7) Epoch 4, batch 4050, loss[loss=0.2746, simple_loss=0.3242, pruned_loss=0.1125, over 4862.00 frames. ], tot_loss[loss=0.2295, simple_loss=0.2855, pruned_loss=0.08673, over 954690.05 frames. ], batch size: 31, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:31:35,214 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:32:12,119 INFO [finetune.py:976] (0/7) Epoch 4, batch 4100, loss[loss=0.2139, simple_loss=0.2765, pruned_loss=0.07568, over 4761.00 frames. ], tot_loss[loss=0.2323, simple_loss=0.289, pruned_loss=0.08775, over 955301.22 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:32:18,162 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21292.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:32:44,633 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:32:53,584 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.825e+02 2.074e+02 2.571e+02 5.101e+02, threshold=4.147e+02, percent-clipped=2.0 +2023-03-26 03:33:04,187 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21328.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:33:07,022 INFO [finetune.py:976] (0/7) Epoch 4, batch 4150, loss[loss=0.2046, simple_loss=0.2774, pruned_loss=0.06584, over 4844.00 frames. ], tot_loss[loss=0.2337, simple_loss=0.2904, pruned_loss=0.08852, over 955253.87 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:33:28,036 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 03:33:35,601 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 03:33:45,547 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21373.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:33:52,602 INFO [finetune.py:976] (0/7) Epoch 4, batch 4200, loss[loss=0.2673, simple_loss=0.318, pruned_loss=0.1083, over 4887.00 frames. ], tot_loss[loss=0.2317, simple_loss=0.2891, pruned_loss=0.08712, over 955164.60 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:34:02,010 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:34:21,941 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:34:30,332 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5594, 1.5561, 1.7946, 1.8766, 1.6627, 3.6013, 1.4275, 1.7676], + device='cuda:0'), covar=tensor([0.0986, 0.1662, 0.1121, 0.0935, 0.1537, 0.0263, 0.1373, 0.1631], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:34:31,865 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 03:34:39,320 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.725e+02 2.024e+02 2.533e+02 3.913e+02, threshold=4.049e+02, percent-clipped=0.0 +2023-03-26 03:34:50,755 INFO [finetune.py:976] (0/7) Epoch 4, batch 4250, loss[loss=0.1856, simple_loss=0.2505, pruned_loss=0.06039, over 4803.00 frames. ], tot_loss[loss=0.2302, simple_loss=0.2868, pruned_loss=0.08676, over 953148.26 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:34:57,770 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 03:35:01,251 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6808, 1.2315, 1.0393, 1.5672, 2.0876, 1.2815, 1.5157, 1.7385], + device='cuda:0'), covar=tensor([0.1313, 0.1897, 0.1961, 0.1112, 0.1762, 0.2165, 0.1237, 0.1630], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0100, 0.0119, 0.0095, 0.0126, 0.0098, 0.0102, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 03:35:09,367 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7160, 1.4610, 1.4371, 1.4303, 1.7914, 1.8374, 1.6632, 1.3796], + device='cuda:0'), covar=tensor([0.0211, 0.0285, 0.0452, 0.0286, 0.0190, 0.0317, 0.0236, 0.0355], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0115, 0.0139, 0.0119, 0.0106, 0.0100, 0.0092, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.7459e-05, 9.0521e-05, 1.1212e-04, 9.3749e-05, 8.3984e-05, 7.4468e-05, + 7.0375e-05, 8.5935e-05], device='cuda:0') +2023-03-26 03:35:18,392 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:35:24,281 INFO [finetune.py:976] (0/7) Epoch 4, batch 4300, loss[loss=0.1808, simple_loss=0.2389, pruned_loss=0.06137, over 4822.00 frames. ], tot_loss[loss=0.226, simple_loss=0.2825, pruned_loss=0.08478, over 953123.33 frames. ], batch size: 41, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:35:51,539 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2056, 2.1003, 2.0550, 2.3704, 2.9782, 2.2837, 2.1417, 1.7509], + device='cuda:0'), covar=tensor([0.2309, 0.2212, 0.1854, 0.1760, 0.1823, 0.1054, 0.2533, 0.1873], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0207, 0.0197, 0.0183, 0.0234, 0.0173, 0.0213, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:35:53,224 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.671e+02 2.065e+02 2.560e+02 4.445e+02, threshold=4.130e+02, percent-clipped=1.0 +2023-03-26 03:36:01,104 INFO [finetune.py:976] (0/7) Epoch 4, batch 4350, loss[loss=0.2262, simple_loss=0.286, pruned_loss=0.08321, over 4834.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.2793, pruned_loss=0.08361, over 953446.52 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:36:11,372 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0752, 1.6153, 1.8265, 1.8466, 1.6586, 1.6700, 1.7558, 1.7013], + device='cuda:0'), covar=tensor([0.7768, 1.1408, 0.8488, 1.0302, 1.1136, 0.8381, 1.3657, 0.7901], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0251, 0.0255, 0.0261, 0.0241, 0.0218, 0.0277, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-26 03:36:40,601 INFO [finetune.py:976] (0/7) Epoch 4, batch 4400, loss[loss=0.2073, simple_loss=0.2683, pruned_loss=0.07317, over 4783.00 frames. ], tot_loss[loss=0.2265, simple_loss=0.2819, pruned_loss=0.08558, over 951568.88 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:36:48,986 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:05,929 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:15,011 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.721e+02 2.036e+02 2.627e+02 4.967e+02, threshold=4.072e+02, percent-clipped=2.0 +2023-03-26 03:37:22,898 INFO [finetune.py:976] (0/7) Epoch 4, batch 4450, loss[loss=0.2049, simple_loss=0.2681, pruned_loss=0.07088, over 4870.00 frames. ], tot_loss[loss=0.2295, simple_loss=0.2855, pruned_loss=0.08677, over 951203.28 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:37:38,545 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:39,220 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-03-26 03:37:50,994 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:57,906 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 03:38:01,165 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:12,007 INFO [finetune.py:976] (0/7) Epoch 4, batch 4500, loss[loss=0.2288, simple_loss=0.2926, pruned_loss=0.08251, over 4835.00 frames. ], tot_loss[loss=0.2299, simple_loss=0.2864, pruned_loss=0.08671, over 950043.47 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:38:12,184 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 03:38:12,672 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:37,419 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 03:38:41,393 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:49,076 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.721e+02 2.032e+02 2.543e+02 5.339e+02, threshold=4.063e+02, percent-clipped=3.0 +2023-03-26 03:38:58,599 INFO [finetune.py:976] (0/7) Epoch 4, batch 4550, loss[loss=0.2766, simple_loss=0.321, pruned_loss=0.1161, over 4707.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.2876, pruned_loss=0.08649, over 952113.40 frames. ], batch size: 54, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:38:59,955 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1142, 3.5328, 3.7011, 3.9627, 3.8055, 3.6745, 4.1797, 1.3023], + device='cuda:0'), covar=tensor([0.0795, 0.0810, 0.0791, 0.0892, 0.1416, 0.1397, 0.0734, 0.5020], + device='cuda:0'), in_proj_covar=tensor([0.0360, 0.0245, 0.0279, 0.0295, 0.0340, 0.0287, 0.0309, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:38:59,966 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5219, 1.5038, 1.8783, 3.0754, 2.1047, 2.1728, 0.9805, 2.4175], + device='cuda:0'), covar=tensor([0.1768, 0.1467, 0.1277, 0.0557, 0.0824, 0.1315, 0.1853, 0.0628], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0166, 0.0104, 0.0143, 0.0130, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 03:39:14,116 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:18,463 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:30,494 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:41,972 INFO [finetune.py:976] (0/7) Epoch 4, batch 4600, loss[loss=0.2069, simple_loss=0.2603, pruned_loss=0.07678, over 4886.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2874, pruned_loss=0.08614, over 955528.35 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:40:25,680 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.789e+02 2.167e+02 2.687e+02 4.147e+02, threshold=4.334e+02, percent-clipped=1.0 +2023-03-26 03:40:32,333 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21822.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:40:33,573 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:40:44,876 INFO [finetune.py:976] (0/7) Epoch 4, batch 4650, loss[loss=0.19, simple_loss=0.2494, pruned_loss=0.06527, over 4765.00 frames. ], tot_loss[loss=0.228, simple_loss=0.2848, pruned_loss=0.08563, over 957140.06 frames. ], batch size: 54, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:41:06,184 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:41:28,342 INFO [finetune.py:976] (0/7) Epoch 4, batch 4700, loss[loss=0.2234, simple_loss=0.2716, pruned_loss=0.08765, over 4807.00 frames. ], tot_loss[loss=0.2266, simple_loss=0.2829, pruned_loss=0.0851, over 959210.26 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:41:41,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2170, 2.2638, 2.5786, 2.3995, 2.4334, 4.7058, 2.2722, 2.3928], + device='cuda:0'), covar=tensor([0.0977, 0.1464, 0.1017, 0.1001, 0.1405, 0.0260, 0.1224, 0.1495], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:42:02,067 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:08,439 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.749e+02 2.081e+02 2.546e+02 7.973e+02, threshold=4.162e+02, percent-clipped=1.0 +2023-03-26 03:42:09,773 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8325, 3.7042, 3.5894, 1.8806, 3.7950, 2.8173, 0.7178, 2.5745], + device='cuda:0'), covar=tensor([0.2749, 0.2052, 0.1440, 0.3153, 0.0986, 0.0971, 0.4692, 0.1579], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0172, 0.0164, 0.0129, 0.0156, 0.0123, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 03:42:16,914 INFO [finetune.py:976] (0/7) Epoch 4, batch 4750, loss[loss=0.2226, simple_loss=0.2819, pruned_loss=0.08169, over 4761.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2814, pruned_loss=0.08501, over 958627.22 frames. ], batch size: 28, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:42:29,490 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:46,712 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:53,028 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:53,716 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:02,052 INFO [finetune.py:976] (0/7) Epoch 4, batch 4800, loss[loss=0.2018, simple_loss=0.2621, pruned_loss=0.0707, over 4784.00 frames. ], tot_loss[loss=0.2263, simple_loss=0.2829, pruned_loss=0.08485, over 958834.86 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:43:02,776 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21984.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:04,626 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 03:43:14,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8719, 1.7511, 1.3323, 1.5164, 1.5914, 1.5619, 1.5278, 2.4490], + device='cuda:0'), covar=tensor([0.8238, 0.7911, 0.6512, 0.8467, 0.6995, 0.4616, 0.7607, 0.2822], + device='cuda:0'), in_proj_covar=tensor([0.0280, 0.0255, 0.0221, 0.0285, 0.0238, 0.0199, 0.0243, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:43:22,670 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-22000.pt +2023-03-26 03:43:34,457 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:37,302 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.769e+02 1.987e+02 2.631e+02 5.032e+02, threshold=3.974e+02, percent-clipped=2.0 +2023-03-26 03:43:44,123 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22031.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:44,662 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22032.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:45,187 INFO [finetune.py:976] (0/7) Epoch 4, batch 4850, loss[loss=0.254, simple_loss=0.3112, pruned_loss=0.09834, over 4931.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.2869, pruned_loss=0.08614, over 959407.16 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:44:19,577 INFO [finetune.py:976] (0/7) Epoch 4, batch 4900, loss[loss=0.2392, simple_loss=0.288, pruned_loss=0.09521, over 4750.00 frames. ], tot_loss[loss=0.2291, simple_loss=0.2866, pruned_loss=0.08583, over 957594.48 frames. ], batch size: 27, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:44:56,526 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8396, 1.5292, 1.3511, 1.3343, 1.4325, 1.4879, 1.4097, 2.2748], + device='cuda:0'), covar=tensor([0.7368, 0.7508, 0.5477, 0.7139, 0.6578, 0.4003, 0.7204, 0.2564], + device='cuda:0'), in_proj_covar=tensor([0.0278, 0.0254, 0.0219, 0.0283, 0.0237, 0.0198, 0.0241, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:44:57,636 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8084, 1.5881, 1.3530, 1.3601, 1.4750, 1.5067, 1.4609, 2.3144], + device='cuda:0'), covar=tensor([0.6985, 0.7320, 0.5502, 0.6883, 0.6229, 0.3969, 0.7039, 0.2428], + device='cuda:0'), in_proj_covar=tensor([0.0278, 0.0254, 0.0219, 0.0283, 0.0237, 0.0198, 0.0241, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:45:00,544 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:45:01,046 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.764e+02 2.232e+02 2.515e+02 4.523e+02, threshold=4.464e+02, percent-clipped=3.0 +2023-03-26 03:45:01,140 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0890, 3.5550, 3.6900, 3.9661, 3.8293, 3.6538, 4.1874, 1.3036], + device='cuda:0'), covar=tensor([0.0766, 0.0763, 0.0818, 0.0921, 0.1314, 0.1347, 0.0714, 0.5113], + device='cuda:0'), in_proj_covar=tensor([0.0362, 0.0246, 0.0279, 0.0295, 0.0340, 0.0287, 0.0309, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:45:18,640 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:45:20,457 INFO [finetune.py:976] (0/7) Epoch 4, batch 4950, loss[loss=0.2177, simple_loss=0.2883, pruned_loss=0.07358, over 4912.00 frames. ], tot_loss[loss=0.2313, simple_loss=0.289, pruned_loss=0.08685, over 958054.88 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:46:12,760 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:19,513 INFO [finetune.py:976] (0/7) Epoch 4, batch 5000, loss[loss=0.1756, simple_loss=0.2392, pruned_loss=0.056, over 4940.00 frames. ], tot_loss[loss=0.2282, simple_loss=0.2859, pruned_loss=0.08525, over 955846.31 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:46:24,502 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:35,047 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:40,529 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:43,480 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.598e+02 2.028e+02 2.482e+02 4.524e+02, threshold=4.056e+02, percent-clipped=1.0 +2023-03-26 03:46:50,677 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0780, 1.4276, 1.1745, 2.0691, 2.3924, 1.9544, 1.8465, 2.1091], + device='cuda:0'), covar=tensor([0.1172, 0.1849, 0.1960, 0.0957, 0.1661, 0.1859, 0.1145, 0.1477], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0117, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 03:46:59,735 INFO [finetune.py:976] (0/7) Epoch 4, batch 5050, loss[loss=0.2241, simple_loss=0.2752, pruned_loss=0.0865, over 4831.00 frames. ], tot_loss[loss=0.2264, simple_loss=0.2834, pruned_loss=0.08469, over 954206.34 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:47:02,254 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:02,266 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:16,044 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:28,181 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22269.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:32,932 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:40,352 INFO [finetune.py:976] (0/7) Epoch 4, batch 5100, loss[loss=0.2204, simple_loss=0.2802, pruned_loss=0.08035, over 4771.00 frames. ], tot_loss[loss=0.2232, simple_loss=0.2798, pruned_loss=0.08329, over 955158.26 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:47:48,143 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1892, 2.6995, 2.5175, 1.3417, 2.7089, 2.3258, 2.1489, 2.3934], + device='cuda:0'), covar=tensor([0.0789, 0.1112, 0.1892, 0.2641, 0.1895, 0.1979, 0.2148, 0.1283], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0202, 0.0204, 0.0192, 0.0219, 0.0210, 0.0221, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:47:49,912 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22298.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:50,453 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:54,136 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 03:48:03,280 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22317.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:48:05,001 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.872e+01 1.620e+02 1.853e+02 2.165e+02 3.345e+02, threshold=3.706e+02, percent-clipped=0.0 +2023-03-26 03:48:08,723 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:48:13,376 INFO [finetune.py:976] (0/7) Epoch 4, batch 5150, loss[loss=0.2237, simple_loss=0.287, pruned_loss=0.08024, over 4846.00 frames. ], tot_loss[loss=0.223, simple_loss=0.2788, pruned_loss=0.08354, over 953188.05 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:48:14,250 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 03:48:49,766 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8725, 1.7891, 1.4316, 1.9557, 1.9708, 1.5309, 2.2641, 1.8658], + device='cuda:0'), covar=tensor([0.1669, 0.3089, 0.3726, 0.3276, 0.2671, 0.1947, 0.3701, 0.2247], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0196, 0.0239, 0.0256, 0.0227, 0.0189, 0.0212, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:48:51,907 INFO [finetune.py:976] (0/7) Epoch 4, batch 5200, loss[loss=0.2443, simple_loss=0.3054, pruned_loss=0.09163, over 4860.00 frames. ], tot_loss[loss=0.2255, simple_loss=0.2816, pruned_loss=0.08466, over 954193.56 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:48:53,234 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:48:58,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8820, 1.8119, 1.9138, 1.1198, 2.1130, 1.9517, 1.9206, 1.6403], + device='cuda:0'), covar=tensor([0.0632, 0.0735, 0.0695, 0.1049, 0.0520, 0.0718, 0.0658, 0.1139], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0133, 0.0144, 0.0128, 0.0110, 0.0143, 0.0147, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:49:07,638 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5575, 1.5453, 1.5561, 0.8991, 1.5194, 1.8103, 1.8422, 1.4062], + device='cuda:0'), covar=tensor([0.1053, 0.0665, 0.0472, 0.0639, 0.0439, 0.0444, 0.0255, 0.0627], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0158, 0.0118, 0.0136, 0.0132, 0.0121, 0.0147, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.7510e-05, 1.1711e-04, 8.5997e-05, 9.9232e-05, 9.5373e-05, 8.9978e-05, + 1.0973e-04, 1.0737e-04], device='cuda:0') +2023-03-26 03:49:26,382 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22419.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:49:26,881 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.279e+02 1.811e+02 2.155e+02 2.737e+02 4.498e+02, threshold=4.310e+02, percent-clipped=4.0 +2023-03-26 03:49:40,615 INFO [finetune.py:976] (0/7) Epoch 4, batch 5250, loss[loss=0.2893, simple_loss=0.3297, pruned_loss=0.1244, over 4833.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2845, pruned_loss=0.0852, over 954668.69 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:49:54,898 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22446.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:15,923 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 03:50:18,144 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:27,827 INFO [finetune.py:976] (0/7) Epoch 4, batch 5300, loss[loss=0.2207, simple_loss=0.2809, pruned_loss=0.08022, over 4851.00 frames. ], tot_loss[loss=0.2301, simple_loss=0.2874, pruned_loss=0.08639, over 954759.11 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:50:30,398 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:49,040 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:51:10,354 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.644e+02 2.088e+02 2.525e+02 4.526e+02, threshold=4.176e+02, percent-clipped=1.0 +2023-03-26 03:51:29,037 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22532.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:51:29,579 INFO [finetune.py:976] (0/7) Epoch 4, batch 5350, loss[loss=0.2393, simple_loss=0.2942, pruned_loss=0.09215, over 4895.00 frames. ], tot_loss[loss=0.2292, simple_loss=0.287, pruned_loss=0.0857, over 956317.33 frames. ], batch size: 37, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:51:43,431 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22554.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:04,982 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:20,638 INFO [finetune.py:976] (0/7) Epoch 4, batch 5400, loss[loss=0.2167, simple_loss=0.2813, pruned_loss=0.07607, over 4778.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.2853, pruned_loss=0.08515, over 957349.40 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:52:27,238 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:33,935 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6205, 0.5432, 1.4757, 1.3820, 1.3523, 1.3317, 1.2800, 1.3798], + device='cuda:0'), covar=tensor([0.5208, 0.8323, 0.7031, 0.7085, 0.7808, 0.5906, 0.8731, 0.6176], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0251, 0.0256, 0.0260, 0.0240, 0.0218, 0.0276, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:52:41,185 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6140, 1.6790, 1.9563, 2.0119, 1.8093, 3.7228, 1.5536, 1.8188], + device='cuda:0'), covar=tensor([0.1060, 0.1761, 0.1111, 0.1059, 0.1615, 0.0249, 0.1527, 0.1729], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 03:52:45,272 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.598e+02 1.961e+02 2.261e+02 4.832e+02, threshold=3.922e+02, percent-clipped=2.0 +2023-03-26 03:52:50,436 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22626.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:54,124 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.13 vs. limit=2.0 +2023-03-26 03:52:54,548 INFO [finetune.py:976] (0/7) Epoch 4, batch 5450, loss[loss=0.2403, simple_loss=0.2832, pruned_loss=0.09874, over 4807.00 frames. ], tot_loss[loss=0.226, simple_loss=0.2828, pruned_loss=0.08458, over 956314.79 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:52:58,339 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:53:30,710 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:53:37,509 INFO [finetune.py:976] (0/7) Epoch 4, batch 5500, loss[loss=0.2152, simple_loss=0.2721, pruned_loss=0.07911, over 4788.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.2804, pruned_loss=0.084, over 954649.27 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:53:48,417 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22700.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 03:54:00,991 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.350e+02 1.787e+02 2.088e+02 2.580e+02 6.017e+02, threshold=4.176e+02, percent-clipped=5.0 +2023-03-26 03:54:16,191 INFO [finetune.py:976] (0/7) Epoch 4, batch 5550, loss[loss=0.2868, simple_loss=0.3075, pruned_loss=0.1331, over 4251.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.283, pruned_loss=0.08559, over 955815.82 frames. ], batch size: 18, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:54:23,406 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:54:30,758 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 03:54:55,508 INFO [finetune.py:976] (0/7) Epoch 4, batch 5600, loss[loss=0.2857, simple_loss=0.3246, pruned_loss=0.1234, over 4879.00 frames. ], tot_loss[loss=0.2319, simple_loss=0.2881, pruned_loss=0.08779, over 956766.08 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:54:57,326 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22786.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:01,954 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22794.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:02,606 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 03:55:06,233 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9713, 1.7875, 1.8137, 1.9779, 2.4839, 1.9266, 1.7976, 1.5484], + device='cuda:0'), covar=tensor([0.2206, 0.2299, 0.1831, 0.1785, 0.2134, 0.1242, 0.2541, 0.1848], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0209, 0.0197, 0.0184, 0.0234, 0.0174, 0.0214, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:55:28,583 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.750e+02 2.147e+02 2.459e+02 4.993e+02, threshold=4.295e+02, percent-clipped=1.0 +2023-03-26 03:55:31,580 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22825.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:40,676 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:41,210 INFO [finetune.py:976] (0/7) Epoch 4, batch 5650, loss[loss=0.2268, simple_loss=0.2938, pruned_loss=0.07993, over 4920.00 frames. ], tot_loss[loss=0.2329, simple_loss=0.2902, pruned_loss=0.08776, over 955850.98 frames. ], batch size: 36, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:55:41,955 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:06,194 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22855.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:08,559 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3783, 2.0397, 2.5383, 1.7957, 2.3945, 2.6126, 2.1081, 2.6913], + device='cuda:0'), covar=tensor([0.1221, 0.1802, 0.1385, 0.1836, 0.0920, 0.1188, 0.2105, 0.0757], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0204, 0.0201, 0.0195, 0.0181, 0.0222, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:56:21,412 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:31,128 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22880.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:32,781 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 03:56:33,110 INFO [finetune.py:976] (0/7) Epoch 4, batch 5700, loss[loss=0.208, simple_loss=0.2439, pruned_loss=0.08605, over 4434.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2861, pruned_loss=0.087, over 939220.82 frames. ], batch size: 19, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:56:34,950 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:41,776 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22893.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:57:03,752 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-4.pt +2023-03-26 03:57:20,681 INFO [finetune.py:976] (0/7) Epoch 5, batch 0, loss[loss=0.2157, simple_loss=0.2827, pruned_loss=0.07429, over 4861.00 frames. ], tot_loss[loss=0.2157, simple_loss=0.2827, pruned_loss=0.07429, over 4861.00 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:57:20,682 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 03:57:32,556 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2312, 1.1647, 1.2041, 1.2184, 1.5226, 1.3850, 1.2872, 1.1457], + device='cuda:0'), covar=tensor([0.0384, 0.0333, 0.0508, 0.0297, 0.0238, 0.0484, 0.0360, 0.0380], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0115, 0.0138, 0.0120, 0.0105, 0.0101, 0.0092, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.7570e-05, 9.0440e-05, 1.1130e-04, 9.4555e-05, 8.3660e-05, 7.5062e-05, + 7.0309e-05, 8.6112e-05], device='cuda:0') +2023-03-26 03:57:37,527 INFO [finetune.py:1010] (0/7) Epoch 5, validation: loss=0.1701, simple_loss=0.2413, pruned_loss=0.0494, over 2265189.00 frames. +2023-03-26 03:57:37,528 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 03:57:48,822 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22919.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:57:49,348 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.844e+01 1.616e+02 1.840e+02 2.309e+02 3.969e+02, threshold=3.680e+02, percent-clipped=0.0 +2023-03-26 03:57:59,135 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 03:58:02,581 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:58:12,459 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8675, 1.7293, 1.6466, 1.8853, 2.4369, 1.9133, 1.5617, 1.4544], + device='cuda:0'), covar=tensor([0.2228, 0.2262, 0.1964, 0.1746, 0.1955, 0.1231, 0.2848, 0.1932], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0210, 0.0198, 0.0185, 0.0236, 0.0175, 0.0215, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 03:58:13,504 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6978, 1.2740, 0.8941, 1.5196, 2.1173, 1.0472, 1.4607, 1.6382], + device='cuda:0'), covar=tensor([0.1627, 0.2134, 0.2162, 0.1269, 0.1951, 0.2140, 0.1435, 0.1976], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0093, 0.0124, 0.0097, 0.0100, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 03:58:29,015 INFO [finetune.py:976] (0/7) Epoch 5, batch 50, loss[loss=0.2623, simple_loss=0.3166, pruned_loss=0.104, over 4836.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.2882, pruned_loss=0.08564, over 217060.31 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:59:05,718 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22995.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 03:59:16,886 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 03:59:17,857 INFO [finetune.py:976] (0/7) Epoch 5, batch 100, loss[loss=0.2294, simple_loss=0.2892, pruned_loss=0.08477, over 4898.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2819, pruned_loss=0.08466, over 378863.31 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:59:23,738 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.764e+02 2.029e+02 2.456e+02 6.922e+02, threshold=4.057e+02, percent-clipped=5.0 +2023-03-26 03:59:36,982 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:59:51,326 INFO [finetune.py:976] (0/7) Epoch 5, batch 150, loss[loss=0.1985, simple_loss=0.2606, pruned_loss=0.06822, over 4919.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2763, pruned_loss=0.08264, over 507612.06 frames. ], batch size: 37, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:00:08,773 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:00:13,242 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6516, 1.1749, 0.9987, 1.5774, 2.0588, 1.0726, 1.3887, 1.6822], + device='cuda:0'), covar=tensor([0.1534, 0.2083, 0.1951, 0.1171, 0.1923, 0.2091, 0.1496, 0.1780], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0117, 0.0093, 0.0124, 0.0097, 0.0101, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 04:00:22,917 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:00:30,574 INFO [finetune.py:976] (0/7) Epoch 5, batch 200, loss[loss=0.2177, simple_loss=0.2692, pruned_loss=0.08314, over 4850.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2771, pruned_loss=0.08297, over 608510.19 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:00:42,594 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.662e+02 1.994e+02 2.595e+02 4.858e+02, threshold=3.989e+02, percent-clipped=3.0 +2023-03-26 04:01:01,383 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:01:09,514 INFO [finetune.py:976] (0/7) Epoch 5, batch 250, loss[loss=0.2026, simple_loss=0.2551, pruned_loss=0.07504, over 4779.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2818, pruned_loss=0.08527, over 685038.04 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:01:18,064 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:01:27,430 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4027, 1.4914, 1.6409, 1.8332, 1.4991, 3.4093, 1.3800, 1.5523], + device='cuda:0'), covar=tensor([0.1014, 0.1758, 0.1263, 0.1076, 0.1749, 0.0271, 0.1481, 0.1721], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0082, 0.0077, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 04:01:31,041 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:01:57,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2294, 1.2270, 1.5652, 1.0798, 1.2560, 1.4193, 1.2269, 1.5793], + device='cuda:0'), covar=tensor([0.1326, 0.2214, 0.1229, 0.1503, 0.0988, 0.1300, 0.2851, 0.0843], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0205, 0.0202, 0.0197, 0.0183, 0.0224, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:02:00,761 INFO [finetune.py:976] (0/7) Epoch 5, batch 300, loss[loss=0.2509, simple_loss=0.313, pruned_loss=0.09439, over 4906.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2864, pruned_loss=0.08675, over 747970.52 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:02:11,928 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:02:20,054 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.235e+02 1.736e+02 2.142e+02 2.597e+02 5.294e+02, threshold=4.284e+02, percent-clipped=3.0 +2023-03-26 04:02:31,212 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:02:31,879 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8614, 1.7523, 1.4785, 1.5610, 1.6242, 1.6653, 1.6184, 2.4170], + device='cuda:0'), covar=tensor([0.7751, 0.7669, 0.6015, 0.7537, 0.6815, 0.4141, 0.7186, 0.2585], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0254, 0.0219, 0.0284, 0.0236, 0.0198, 0.0241, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:02:52,126 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 04:03:03,008 INFO [finetune.py:976] (0/7) Epoch 5, batch 350, loss[loss=0.1685, simple_loss=0.2332, pruned_loss=0.05191, over 4817.00 frames. ], tot_loss[loss=0.233, simple_loss=0.2889, pruned_loss=0.08856, over 791888.53 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:03:20,885 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23278.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:31,491 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 04:03:34,244 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:41,711 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23295.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:51,622 INFO [finetune.py:976] (0/7) Epoch 5, batch 400, loss[loss=0.2128, simple_loss=0.2666, pruned_loss=0.07951, over 4918.00 frames. ], tot_loss[loss=0.2309, simple_loss=0.2879, pruned_loss=0.08694, over 829870.56 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:03:58,185 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.903e+01 1.731e+02 2.116e+02 2.565e+02 5.981e+02, threshold=4.232e+02, percent-clipped=1.0 +2023-03-26 04:04:13,596 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:04:24,904 INFO [finetune.py:976] (0/7) Epoch 5, batch 450, loss[loss=0.207, simple_loss=0.2808, pruned_loss=0.06656, over 4775.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2852, pruned_loss=0.08567, over 855272.11 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:05:10,552 INFO [finetune.py:976] (0/7) Epoch 5, batch 500, loss[loss=0.2268, simple_loss=0.2884, pruned_loss=0.08262, over 4788.00 frames. ], tot_loss[loss=0.2262, simple_loss=0.2828, pruned_loss=0.08483, over 876760.05 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:05:16,626 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.775e+02 2.029e+02 2.615e+02 5.539e+02, threshold=4.057e+02, percent-clipped=1.0 +2023-03-26 04:05:33,331 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0100, 1.7133, 2.3491, 1.6442, 2.0732, 2.3932, 1.8099, 2.5065], + device='cuda:0'), covar=tensor([0.1423, 0.2135, 0.1449, 0.1766, 0.0956, 0.1165, 0.2591, 0.0850], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0205, 0.0202, 0.0196, 0.0183, 0.0223, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:05:42,934 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23450.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:05:52,285 INFO [finetune.py:976] (0/7) Epoch 5, batch 550, loss[loss=0.2154, simple_loss=0.2774, pruned_loss=0.07669, over 4893.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.2797, pruned_loss=0.08324, over 895607.49 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:05:54,204 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:15,707 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:23,760 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23488.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:30,288 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23498.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:38,164 INFO [finetune.py:976] (0/7) Epoch 5, batch 600, loss[loss=0.2117, simple_loss=0.2818, pruned_loss=0.07078, over 4832.00 frames. ], tot_loss[loss=0.2227, simple_loss=0.279, pruned_loss=0.08318, over 907955.88 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:06:44,763 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.759e+02 2.043e+02 2.434e+02 4.744e+02, threshold=4.086e+02, percent-clipped=2.0 +2023-03-26 04:06:51,175 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23529.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:07:18,714 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23549.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 04:07:25,882 INFO [finetune.py:976] (0/7) Epoch 5, batch 650, loss[loss=0.2591, simple_loss=0.3279, pruned_loss=0.09519, over 4817.00 frames. ], tot_loss[loss=0.2261, simple_loss=0.2831, pruned_loss=0.08451, over 920005.82 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:07:33,748 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23573.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:07:42,964 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:08:01,899 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7840, 2.4303, 2.1924, 1.0732, 2.3123, 2.0963, 1.9569, 2.1419], + device='cuda:0'), covar=tensor([0.0959, 0.0996, 0.1605, 0.2488, 0.1668, 0.2266, 0.2246, 0.1165], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0202, 0.0190, 0.0216, 0.0209, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:08:14,421 INFO [finetune.py:976] (0/7) Epoch 5, batch 700, loss[loss=0.2174, simple_loss=0.2707, pruned_loss=0.08198, over 4275.00 frames. ], tot_loss[loss=0.2304, simple_loss=0.2871, pruned_loss=0.08684, over 927580.02 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:08:30,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.701e+02 2.127e+02 2.576e+02 5.648e+02, threshold=4.253e+02, percent-clipped=2.0 +2023-03-26 04:08:54,262 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-26 04:09:25,209 INFO [finetune.py:976] (0/7) Epoch 5, batch 750, loss[loss=0.205, simple_loss=0.2849, pruned_loss=0.06253, over 4887.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2893, pruned_loss=0.08804, over 934946.48 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:09:46,254 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 04:10:02,096 INFO [finetune.py:976] (0/7) Epoch 5, batch 800, loss[loss=0.2675, simple_loss=0.3103, pruned_loss=0.1124, over 4897.00 frames. ], tot_loss[loss=0.2309, simple_loss=0.2879, pruned_loss=0.08692, over 938995.84 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:10:08,708 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.635e+02 1.954e+02 2.429e+02 4.773e+02, threshold=3.908e+02, percent-clipped=1.0 +2023-03-26 04:10:14,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4824, 1.3226, 1.2917, 1.3472, 1.7364, 1.7035, 1.5175, 1.2763], + device='cuda:0'), covar=tensor([0.0300, 0.0276, 0.0495, 0.0272, 0.0195, 0.0321, 0.0268, 0.0336], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0114, 0.0139, 0.0120, 0.0105, 0.0102, 0.0091, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.8048e-05, 8.9916e-05, 1.1181e-04, 9.4600e-05, 8.3294e-05, 7.5891e-05, + 7.0006e-05, 8.5844e-05], device='cuda:0') +2023-03-26 04:10:56,824 INFO [finetune.py:976] (0/7) Epoch 5, batch 850, loss[loss=0.2568, simple_loss=0.3037, pruned_loss=0.1049, over 4892.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2854, pruned_loss=0.08583, over 943251.35 frames. ], batch size: 32, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:11:04,267 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23764.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:11:54,660 INFO [finetune.py:976] (0/7) Epoch 5, batch 900, loss[loss=0.2202, simple_loss=0.2792, pruned_loss=0.0806, over 4915.00 frames. ], tot_loss[loss=0.2251, simple_loss=0.2821, pruned_loss=0.08406, over 946681.30 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:11:55,331 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23812.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:12:00,783 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.642e+02 1.957e+02 2.389e+02 4.840e+02, threshold=3.913e+02, percent-clipped=2.0 +2023-03-26 04:12:10,526 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5706, 1.3293, 1.1741, 1.2270, 1.7628, 1.8221, 1.6282, 1.3278], + device='cuda:0'), covar=tensor([0.0270, 0.0385, 0.0717, 0.0415, 0.0204, 0.0424, 0.0283, 0.0385], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0114, 0.0138, 0.0119, 0.0105, 0.0101, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.7800e-05, 8.9402e-05, 1.1132e-04, 9.4017e-05, 8.2807e-05, 7.5348e-05, + 6.9809e-05, 8.5522e-05], device='cuda:0') +2023-03-26 04:12:21,815 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23844.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:12:37,545 INFO [finetune.py:976] (0/7) Epoch 5, batch 950, loss[loss=0.217, simple_loss=0.2611, pruned_loss=0.08644, over 4386.00 frames. ], tot_loss[loss=0.2244, simple_loss=0.2807, pruned_loss=0.084, over 948672.47 frames. ], batch size: 19, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:12:44,899 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23873.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:12:47,365 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9924, 2.5203, 2.3864, 1.3230, 2.3794, 2.0940, 1.8462, 2.2453], + device='cuda:0'), covar=tensor([0.1058, 0.0950, 0.2028, 0.2443, 0.2240, 0.2802, 0.2640, 0.1492], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0199, 0.0202, 0.0189, 0.0216, 0.0208, 0.0220, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:12:52,623 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:13:28,835 INFO [finetune.py:976] (0/7) Epoch 5, batch 1000, loss[loss=0.2329, simple_loss=0.2687, pruned_loss=0.09851, over 4258.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.2821, pruned_loss=0.08487, over 947804.34 frames. ], batch size: 65, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:13:38,584 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.702e+02 2.066e+02 2.385e+02 5.722e+02, threshold=4.131e+02, percent-clipped=3.0 +2023-03-26 04:13:38,648 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23921.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:13:43,136 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 04:13:47,709 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:13:56,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7042, 1.6723, 1.7952, 1.8493, 1.8223, 3.3318, 1.4515, 1.7709], + device='cuda:0'), covar=tensor([0.0838, 0.1489, 0.0873, 0.0899, 0.1389, 0.0268, 0.1281, 0.1348], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0082, 0.0077, 0.0080, 0.0093, 0.0084, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 04:14:04,233 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 04:14:06,412 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6201, 1.5586, 1.6702, 0.8352, 1.6950, 1.8722, 1.8466, 1.5005], + device='cuda:0'), covar=tensor([0.0814, 0.0593, 0.0424, 0.0575, 0.0378, 0.0465, 0.0285, 0.0602], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0157, 0.0117, 0.0135, 0.0132, 0.0121, 0.0147, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.7062e-05, 1.1645e-04, 8.5260e-05, 9.9024e-05, 9.5047e-05, 8.9855e-05, + 1.0922e-04, 1.0676e-04], device='cuda:0') +2023-03-26 04:14:14,856 INFO [finetune.py:976] (0/7) Epoch 5, batch 1050, loss[loss=0.1992, simple_loss=0.2692, pruned_loss=0.06457, over 4895.00 frames. ], tot_loss[loss=0.2275, simple_loss=0.2841, pruned_loss=0.08539, over 948754.57 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:14:36,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7575, 1.4705, 2.1208, 3.3111, 2.3585, 2.4050, 1.0067, 2.5870], + device='cuda:0'), covar=tensor([0.1733, 0.1617, 0.1368, 0.0618, 0.0765, 0.1526, 0.1945, 0.0625], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0119, 0.0136, 0.0167, 0.0103, 0.0142, 0.0129, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 04:14:39,474 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5932, 1.5861, 1.8281, 1.8746, 1.7074, 3.6094, 1.3881, 1.7057], + device='cuda:0'), covar=tensor([0.0955, 0.1740, 0.1134, 0.1005, 0.1534, 0.0237, 0.1485, 0.1695], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 04:15:02,104 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-24000.pt +2023-03-26 04:15:19,193 INFO [finetune.py:976] (0/7) Epoch 5, batch 1100, loss[loss=0.2209, simple_loss=0.2868, pruned_loss=0.07749, over 4902.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2854, pruned_loss=0.08562, over 949850.16 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:15:28,406 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.825e+02 2.109e+02 2.589e+02 5.024e+02, threshold=4.219e+02, percent-clipped=4.0 +2023-03-26 04:15:38,249 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:15:48,778 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0108, 1.6487, 2.4081, 3.8448, 2.7058, 2.6080, 0.8425, 3.0502], + device='cuda:0'), covar=tensor([0.1824, 0.1645, 0.1437, 0.0538, 0.0796, 0.1778, 0.2181, 0.0581], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0118, 0.0136, 0.0167, 0.0102, 0.0142, 0.0128, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 04:15:54,483 INFO [finetune.py:976] (0/7) Epoch 5, batch 1150, loss[loss=0.2132, simple_loss=0.2769, pruned_loss=0.07476, over 4905.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.2865, pruned_loss=0.08634, over 951664.62 frames. ], batch size: 36, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:15:59,270 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5838, 2.2807, 1.9555, 0.9106, 2.1930, 1.9236, 1.7930, 2.0650], + device='cuda:0'), covar=tensor([0.0820, 0.1034, 0.1691, 0.2304, 0.1455, 0.2341, 0.2200, 0.1060], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0202, 0.0190, 0.0215, 0.0208, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:16:18,937 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:16:28,016 INFO [finetune.py:976] (0/7) Epoch 5, batch 1200, loss[loss=0.2796, simple_loss=0.3214, pruned_loss=0.1189, over 4849.00 frames. ], tot_loss[loss=0.2265, simple_loss=0.2838, pruned_loss=0.08465, over 950395.68 frames. ], batch size: 44, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:16:37,230 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.721e+02 2.129e+02 2.606e+02 7.150e+02, threshold=4.257e+02, percent-clipped=3.0 +2023-03-26 04:16:47,778 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7755, 1.1715, 0.8899, 1.6657, 2.1090, 1.2587, 1.6604, 1.7567], + device='cuda:0'), covar=tensor([0.1440, 0.2090, 0.2236, 0.1189, 0.1998, 0.2151, 0.1362, 0.1825], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0098, 0.0117, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 04:16:51,871 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24144.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:17:03,454 INFO [finetune.py:976] (0/7) Epoch 5, batch 1250, loss[loss=0.214, simple_loss=0.2616, pruned_loss=0.08324, over 4783.00 frames. ], tot_loss[loss=0.2232, simple_loss=0.2804, pruned_loss=0.08298, over 951749.17 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:17:28,512 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24192.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:17:42,424 INFO [finetune.py:976] (0/7) Epoch 5, batch 1300, loss[loss=0.1934, simple_loss=0.2633, pruned_loss=0.06174, over 4926.00 frames. ], tot_loss[loss=0.2189, simple_loss=0.2763, pruned_loss=0.08075, over 953977.47 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:17:56,617 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.607e+02 1.851e+02 2.364e+02 3.844e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 04:18:19,560 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9077, 1.7546, 2.2593, 1.5716, 2.1163, 2.1872, 1.7433, 2.2610], + device='cuda:0'), covar=tensor([0.1144, 0.1693, 0.0934, 0.1620, 0.0643, 0.0997, 0.2057, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0204, 0.0200, 0.0195, 0.0182, 0.0221, 0.0214, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:18:34,750 INFO [finetune.py:976] (0/7) Epoch 5, batch 1350, loss[loss=0.2165, simple_loss=0.2817, pruned_loss=0.07569, over 4874.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2765, pruned_loss=0.08131, over 954908.44 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:18:37,160 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0328, 2.2089, 2.0842, 1.4690, 2.2628, 2.2516, 2.1677, 1.8762], + device='cuda:0'), covar=tensor([0.0755, 0.0558, 0.0808, 0.1006, 0.0476, 0.0722, 0.0676, 0.1029], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0134, 0.0145, 0.0128, 0.0112, 0.0144, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:19:12,805 INFO [finetune.py:976] (0/7) Epoch 5, batch 1400, loss[loss=0.2189, simple_loss=0.2724, pruned_loss=0.08273, over 4718.00 frames. ], tot_loss[loss=0.2236, simple_loss=0.2814, pruned_loss=0.08293, over 956748.90 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:19:21,620 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.714e+02 2.138e+02 2.571e+02 4.877e+02, threshold=4.276e+02, percent-clipped=6.0 +2023-03-26 04:19:44,457 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 04:19:56,902 INFO [finetune.py:976] (0/7) Epoch 5, batch 1450, loss[loss=0.2283, simple_loss=0.2905, pruned_loss=0.08308, over 4824.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2836, pruned_loss=0.08357, over 954393.85 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:19:59,938 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7399, 1.5739, 1.6316, 1.7670, 1.0580, 3.7544, 1.4370, 2.0961], + device='cuda:0'), covar=tensor([0.3381, 0.2441, 0.1989, 0.2204, 0.1974, 0.0149, 0.2598, 0.1296], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0113, 0.0116, 0.0120, 0.0116, 0.0097, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 04:20:20,191 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:20:22,708 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6711, 1.4223, 1.2736, 1.1235, 1.3789, 1.3733, 1.3297, 2.0351], + device='cuda:0'), covar=tensor([0.6835, 0.6393, 0.5312, 0.6643, 0.5665, 0.3539, 0.6400, 0.2593], + device='cuda:0'), in_proj_covar=tensor([0.0280, 0.0254, 0.0220, 0.0283, 0.0236, 0.0199, 0.0242, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:20:28,757 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4841, 1.9805, 2.7727, 1.6296, 2.5924, 2.9191, 2.0479, 2.7430], + device='cuda:0'), covar=tensor([0.1642, 0.2329, 0.1858, 0.2733, 0.1156, 0.1666, 0.2634, 0.1407], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0205, 0.0201, 0.0195, 0.0182, 0.0222, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:20:31,768 INFO [finetune.py:976] (0/7) Epoch 5, batch 1500, loss[loss=0.1889, simple_loss=0.247, pruned_loss=0.0654, over 4786.00 frames. ], tot_loss[loss=0.2281, simple_loss=0.2862, pruned_loss=0.08505, over 953026.35 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:20:38,324 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.776e+02 2.138e+02 2.564e+02 4.291e+02, threshold=4.276e+02, percent-clipped=1.0 +2023-03-26 04:21:13,451 INFO [finetune.py:976] (0/7) Epoch 5, batch 1550, loss[loss=0.2743, simple_loss=0.3319, pruned_loss=0.1084, over 4824.00 frames. ], tot_loss[loss=0.2268, simple_loss=0.2856, pruned_loss=0.08406, over 954461.65 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:21:33,959 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4792, 1.3093, 1.7866, 2.7031, 1.8824, 2.0309, 0.9177, 2.1816], + device='cuda:0'), covar=tensor([0.1910, 0.1829, 0.1433, 0.0884, 0.0908, 0.1952, 0.1977, 0.0794], + device='cuda:0'), in_proj_covar=tensor([0.0105, 0.0119, 0.0137, 0.0168, 0.0103, 0.0144, 0.0130, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 04:21:47,121 INFO [finetune.py:976] (0/7) Epoch 5, batch 1600, loss[loss=0.2612, simple_loss=0.3122, pruned_loss=0.1051, over 4713.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.2825, pruned_loss=0.08256, over 954903.12 frames. ], batch size: 59, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:21:58,803 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.770e+02 2.018e+02 2.552e+02 5.194e+02, threshold=4.037e+02, percent-clipped=4.0 +2023-03-26 04:22:21,136 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:22:33,861 INFO [finetune.py:976] (0/7) Epoch 5, batch 1650, loss[loss=0.1771, simple_loss=0.2469, pruned_loss=0.05365, over 4830.00 frames. ], tot_loss[loss=0.2214, simple_loss=0.2795, pruned_loss=0.08166, over 955124.73 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:22:43,195 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:23:17,416 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24608.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:23:22,440 INFO [finetune.py:976] (0/7) Epoch 5, batch 1700, loss[loss=0.2177, simple_loss=0.2815, pruned_loss=0.07698, over 4755.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2769, pruned_loss=0.08116, over 953179.76 frames. ], batch size: 54, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:23:22,622 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 04:23:31,258 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.673e+02 1.915e+02 2.251e+02 4.027e+02, threshold=3.830e+02, percent-clipped=0.0 +2023-03-26 04:23:42,117 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:24:06,425 INFO [finetune.py:976] (0/7) Epoch 5, batch 1750, loss[loss=0.2759, simple_loss=0.3271, pruned_loss=0.1123, over 4809.00 frames. ], tot_loss[loss=0.223, simple_loss=0.2802, pruned_loss=0.0829, over 954246.09 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:24:17,937 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3055, 2.0281, 1.7698, 1.4728, 2.3895, 2.5203, 2.3081, 2.0400], + device='cuda:0'), covar=tensor([0.0252, 0.0425, 0.0482, 0.0432, 0.0302, 0.0401, 0.0252, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0114, 0.0140, 0.0120, 0.0106, 0.0102, 0.0092, 0.0111], + device='cuda:0'), out_proj_covar=tensor([6.8568e-05, 8.9943e-05, 1.1246e-04, 9.4538e-05, 8.4296e-05, 7.5892e-05, + 7.0802e-05, 8.6392e-05], device='cuda:0') +2023-03-26 04:24:19,128 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6169, 1.4277, 1.8978, 2.7410, 1.9820, 2.0721, 1.1945, 2.2875], + device='cuda:0'), covar=tensor([0.1572, 0.1389, 0.1101, 0.0553, 0.0772, 0.1896, 0.1513, 0.0647], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0118, 0.0135, 0.0165, 0.0102, 0.0142, 0.0128, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 04:24:27,994 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24693.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:24:30,418 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-03-26 04:24:32,782 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6852, 1.5127, 1.4600, 1.6627, 2.0433, 1.6253, 1.3386, 1.3786], + device='cuda:0'), covar=tensor([0.2148, 0.2285, 0.1910, 0.1745, 0.1984, 0.1193, 0.2717, 0.1840], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0209, 0.0199, 0.0185, 0.0236, 0.0175, 0.0215, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:24:39,305 INFO [finetune.py:976] (0/7) Epoch 5, batch 1800, loss[loss=0.1755, simple_loss=0.2419, pruned_loss=0.0545, over 4774.00 frames. ], tot_loss[loss=0.2253, simple_loss=0.283, pruned_loss=0.08377, over 953452.41 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:24:45,828 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.800e+02 2.166e+02 2.491e+02 4.201e+02, threshold=4.331e+02, percent-clipped=2.0 +2023-03-26 04:24:59,906 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:25:12,945 INFO [finetune.py:976] (0/7) Epoch 5, batch 1850, loss[loss=0.2654, simple_loss=0.3191, pruned_loss=0.1059, over 4812.00 frames. ], tot_loss[loss=0.2284, simple_loss=0.2862, pruned_loss=0.08527, over 952658.29 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:25:15,491 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:25:29,394 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 04:25:30,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4929, 1.3094, 1.2681, 1.4729, 1.6966, 1.4161, 0.9820, 1.2913], + device='cuda:0'), covar=tensor([0.2017, 0.2273, 0.1872, 0.1576, 0.1570, 0.1208, 0.2731, 0.1788], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0210, 0.0199, 0.0185, 0.0237, 0.0175, 0.0215, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:25:35,072 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7395, 1.4782, 2.1827, 3.3443, 2.3706, 2.4101, 1.2815, 2.7482], + device='cuda:0'), covar=tensor([0.1817, 0.1513, 0.1305, 0.0493, 0.0760, 0.1310, 0.1700, 0.0586], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0118, 0.0136, 0.0165, 0.0102, 0.0142, 0.0128, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 04:25:44,637 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0704, 3.5062, 3.6643, 3.8586, 3.8192, 3.5885, 4.1280, 1.5039], + device='cuda:0'), covar=tensor([0.0729, 0.0886, 0.0823, 0.0997, 0.1173, 0.1349, 0.0695, 0.4927], + device='cuda:0'), in_proj_covar=tensor([0.0361, 0.0247, 0.0277, 0.0295, 0.0341, 0.0286, 0.0307, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:25:46,395 INFO [finetune.py:976] (0/7) Epoch 5, batch 1900, loss[loss=0.2124, simple_loss=0.2785, pruned_loss=0.07313, over 4848.00 frames. ], tot_loss[loss=0.2287, simple_loss=0.287, pruned_loss=0.08515, over 953454.09 frames. ], batch size: 44, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:25:52,457 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.802e+02 2.061e+02 2.489e+02 6.200e+02, threshold=4.122e+02, percent-clipped=1.0 +2023-03-26 04:25:57,986 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:26:20,825 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:26:27,332 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.68 vs. limit=2.0 +2023-03-26 04:26:29,634 INFO [finetune.py:976] (0/7) Epoch 5, batch 1950, loss[loss=0.1989, simple_loss=0.2671, pruned_loss=0.06539, over 4924.00 frames. ], tot_loss[loss=0.2255, simple_loss=0.2847, pruned_loss=0.0832, over 954845.61 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:26:45,924 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2375, 1.8442, 1.8415, 0.8379, 2.0098, 2.2038, 1.9479, 1.8662], + device='cuda:0'), covar=tensor([0.0688, 0.0631, 0.0507, 0.0612, 0.0409, 0.0382, 0.0428, 0.0455], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0158, 0.0119, 0.0136, 0.0133, 0.0122, 0.0147, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.7788e-05, 1.1698e-04, 8.6040e-05, 9.9394e-05, 9.5944e-05, 9.0555e-05, + 1.0897e-04, 1.0718e-04], device='cuda:0') +2023-03-26 04:26:57,666 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:01,811 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:02,449 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8591, 1.6952, 1.6315, 1.9917, 2.2097, 1.8964, 1.4042, 1.5296], + device='cuda:0'), covar=tensor([0.2445, 0.2585, 0.2163, 0.1855, 0.2036, 0.1284, 0.2968, 0.1998], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0211, 0.0201, 0.0186, 0.0238, 0.0175, 0.0217, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:27:02,942 INFO [finetune.py:976] (0/7) Epoch 5, batch 2000, loss[loss=0.1974, simple_loss=0.2625, pruned_loss=0.06614, over 4820.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.2814, pruned_loss=0.0826, over 955031.87 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:27:13,003 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.611e+02 2.012e+02 2.424e+02 3.709e+02, threshold=4.024e+02, percent-clipped=0.0 +2023-03-26 04:27:15,539 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:50,325 INFO [finetune.py:976] (0/7) Epoch 5, batch 2050, loss[loss=0.195, simple_loss=0.2597, pruned_loss=0.06509, over 4798.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2782, pruned_loss=0.08141, over 955443.58 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:27:57,769 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9929, 1.7663, 1.7564, 1.9873, 1.2446, 4.6197, 1.7871, 2.4964], + device='cuda:0'), covar=tensor([0.3535, 0.2551, 0.2025, 0.2318, 0.1913, 0.0117, 0.2543, 0.1240], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0118, 0.0098, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 04:28:08,219 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-26 04:28:12,193 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:28:23,339 INFO [finetune.py:976] (0/7) Epoch 5, batch 2100, loss[loss=0.2543, simple_loss=0.323, pruned_loss=0.09285, over 4753.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2774, pruned_loss=0.08086, over 953442.11 frames. ], batch size: 59, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:28:39,037 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.247e+01 1.663e+02 1.989e+02 2.457e+02 4.446e+02, threshold=3.978e+02, percent-clipped=1.0 +2023-03-26 04:28:42,879 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5256, 2.3212, 1.8929, 0.9054, 2.0189, 1.9346, 1.7454, 1.9856], + device='cuda:0'), covar=tensor([0.0912, 0.0909, 0.1706, 0.2347, 0.1550, 0.2656, 0.2450, 0.1065], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0203, 0.0190, 0.0216, 0.0209, 0.0221, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:28:52,418 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.61 vs. limit=2.0 +2023-03-26 04:29:08,272 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25056.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 04:29:11,160 INFO [finetune.py:976] (0/7) Epoch 5, batch 2150, loss[loss=0.3094, simple_loss=0.3583, pruned_loss=0.1302, over 4810.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2798, pruned_loss=0.08181, over 953360.26 frames. ], batch size: 41, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:29:45,153 INFO [finetune.py:976] (0/7) Epoch 5, batch 2200, loss[loss=0.2535, simple_loss=0.3147, pruned_loss=0.09612, over 4899.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2832, pruned_loss=0.08299, over 954356.40 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:29:52,261 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.691e+02 1.983e+02 2.301e+02 4.176e+02, threshold=3.967e+02, percent-clipped=1.0 +2023-03-26 04:29:52,334 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:30:10,046 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1847, 1.3256, 1.4266, 0.5683, 1.1968, 1.5868, 1.6012, 1.2823], + device='cuda:0'), covar=tensor([0.0912, 0.0492, 0.0423, 0.0607, 0.0456, 0.0471, 0.0290, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0159, 0.0119, 0.0137, 0.0134, 0.0123, 0.0148, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.8574e-05, 1.1781e-04, 8.6504e-05, 1.0019e-04, 9.6497e-05, 9.1271e-05, + 1.1013e-04, 1.0784e-04], device='cuda:0') +2023-03-26 04:30:18,606 INFO [finetune.py:976] (0/7) Epoch 5, batch 2250, loss[loss=0.2484, simple_loss=0.3101, pruned_loss=0.09338, over 4815.00 frames. ], tot_loss[loss=0.227, simple_loss=0.2858, pruned_loss=0.08414, over 955896.73 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:30:28,811 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-26 04:30:46,282 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25203.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:30:46,973 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:30:53,042 INFO [finetune.py:976] (0/7) Epoch 5, batch 2300, loss[loss=0.2581, simple_loss=0.3125, pruned_loss=0.1019, over 4797.00 frames. ], tot_loss[loss=0.2269, simple_loss=0.286, pruned_loss=0.0839, over 956976.64 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:31:05,174 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.840e+02 2.117e+02 2.638e+02 5.911e+02, threshold=4.234e+02, percent-clipped=5.0 +2023-03-26 04:31:14,025 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:31:35,989 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:31:42,891 INFO [finetune.py:976] (0/7) Epoch 5, batch 2350, loss[loss=0.2331, simple_loss=0.2876, pruned_loss=0.08935, over 4872.00 frames. ], tot_loss[loss=0.2251, simple_loss=0.2839, pruned_loss=0.0831, over 957229.89 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:31:51,257 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:32:16,767 INFO [finetune.py:976] (0/7) Epoch 5, batch 2400, loss[loss=0.2049, simple_loss=0.2595, pruned_loss=0.07509, over 4937.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2805, pruned_loss=0.0822, over 957206.77 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:32:23,860 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.631e+02 1.900e+02 2.318e+02 5.058e+02, threshold=3.799e+02, percent-clipped=1.0 +2023-03-26 04:32:58,152 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25351.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 04:33:04,187 INFO [finetune.py:976] (0/7) Epoch 5, batch 2450, loss[loss=0.1738, simple_loss=0.2309, pruned_loss=0.05837, over 4706.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2761, pruned_loss=0.08017, over 956785.81 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:34:02,098 INFO [finetune.py:976] (0/7) Epoch 5, batch 2500, loss[loss=0.2589, simple_loss=0.3237, pruned_loss=0.09707, over 4896.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2773, pruned_loss=0.08097, over 956681.39 frames. ], batch size: 43, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:34:18,834 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.793e+02 2.115e+02 2.620e+02 5.379e+02, threshold=4.229e+02, percent-clipped=6.0 +2023-03-26 04:34:18,928 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:34:26,161 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25428.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:34:27,581 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 04:34:47,668 INFO [finetune.py:976] (0/7) Epoch 5, batch 2550, loss[loss=0.1681, simple_loss=0.2365, pruned_loss=0.04984, over 4796.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.2813, pruned_loss=0.08221, over 955679.51 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:34:53,565 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25469.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:34:54,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4387, 1.3825, 1.2270, 1.2848, 1.7083, 1.6242, 1.4532, 1.2417], + device='cuda:0'), covar=tensor([0.0350, 0.0298, 0.0567, 0.0316, 0.0206, 0.0396, 0.0287, 0.0410], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0113, 0.0139, 0.0119, 0.0106, 0.0102, 0.0092, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.8483e-05, 8.9237e-05, 1.1197e-04, 9.3955e-05, 8.3889e-05, 7.5654e-05, + 7.0233e-05, 8.5914e-05], device='cuda:0') +2023-03-26 04:35:07,247 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25489.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:35:16,166 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:35:20,835 INFO [finetune.py:976] (0/7) Epoch 5, batch 2600, loss[loss=0.1838, simple_loss=0.2534, pruned_loss=0.05708, over 4782.00 frames. ], tot_loss[loss=0.2237, simple_loss=0.2823, pruned_loss=0.08261, over 953781.49 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:35:27,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9160, 2.2231, 2.0094, 1.3752, 2.3274, 2.1473, 1.9714, 1.8770], + device='cuda:0'), covar=tensor([0.0661, 0.0530, 0.0756, 0.0938, 0.0412, 0.0698, 0.0689, 0.0929], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0126, 0.0110, 0.0142, 0.0145, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:35:28,040 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.693e+02 2.088e+02 2.425e+02 4.415e+02, threshold=4.177e+02, percent-clipped=1.0 +2023-03-26 04:35:48,652 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:35:54,563 INFO [finetune.py:976] (0/7) Epoch 5, batch 2650, loss[loss=0.2085, simple_loss=0.2639, pruned_loss=0.07649, over 4820.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.2821, pruned_loss=0.08226, over 952145.41 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:36:33,948 INFO [finetune.py:976] (0/7) Epoch 5, batch 2700, loss[loss=0.1728, simple_loss=0.243, pruned_loss=0.05128, over 4820.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2808, pruned_loss=0.0813, over 952112.65 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:36:50,929 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.711e+02 2.002e+02 2.331e+02 3.948e+02, threshold=4.004e+02, percent-clipped=0.0 +2023-03-26 04:36:52,260 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5977, 1.3451, 1.4181, 1.3563, 1.8318, 1.7755, 1.5888, 1.3507], + device='cuda:0'), covar=tensor([0.0260, 0.0323, 0.0499, 0.0310, 0.0181, 0.0387, 0.0266, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0113, 0.0139, 0.0118, 0.0105, 0.0102, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.8219e-05, 8.8987e-05, 1.1158e-04, 9.3413e-05, 8.3374e-05, 7.5707e-05, + 6.9799e-05, 8.5496e-05], device='cuda:0') +2023-03-26 04:37:22,107 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8306, 1.5405, 1.3794, 1.4664, 1.5249, 1.4881, 1.4942, 2.2636], + device='cuda:0'), covar=tensor([0.6685, 0.8013, 0.5457, 0.6376, 0.5887, 0.3946, 0.6335, 0.2521], + device='cuda:0'), in_proj_covar=tensor([0.0281, 0.0255, 0.0220, 0.0284, 0.0237, 0.0200, 0.0243, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:37:26,216 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25651.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:37:32,345 INFO [finetune.py:976] (0/7) Epoch 5, batch 2750, loss[loss=0.1805, simple_loss=0.2316, pruned_loss=0.06468, over 4261.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2775, pruned_loss=0.08007, over 951978.05 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:37:58,588 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:38:07,649 INFO [finetune.py:976] (0/7) Epoch 5, batch 2800, loss[loss=0.2011, simple_loss=0.2592, pruned_loss=0.07145, over 4756.00 frames. ], tot_loss[loss=0.2157, simple_loss=0.2737, pruned_loss=0.07882, over 953190.72 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:38:17,734 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 04:38:23,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.611e+02 1.888e+02 2.301e+02 3.388e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 04:38:52,630 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8482, 1.6892, 1.6402, 1.8865, 2.3554, 1.8519, 1.6031, 1.4971], + device='cuda:0'), covar=tensor([0.2113, 0.2152, 0.1789, 0.1660, 0.1752, 0.1225, 0.2508, 0.1759], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0208, 0.0198, 0.0183, 0.0234, 0.0173, 0.0214, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:39:03,402 INFO [finetune.py:976] (0/7) Epoch 5, batch 2850, loss[loss=0.2694, simple_loss=0.3274, pruned_loss=0.1057, over 4927.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2723, pruned_loss=0.07823, over 955750.70 frames. ], batch size: 42, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:39:18,029 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25784.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:39:21,419 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5598, 2.4262, 1.9991, 0.9964, 2.1195, 1.9215, 1.6359, 2.0278], + device='cuda:0'), covar=tensor([0.0801, 0.0821, 0.1869, 0.2183, 0.1683, 0.2377, 0.2415, 0.1184], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0203, 0.0189, 0.0217, 0.0208, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:39:35,321 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-26 04:39:37,534 INFO [finetune.py:976] (0/7) Epoch 5, batch 2900, loss[loss=0.2057, simple_loss=0.2712, pruned_loss=0.07011, over 4862.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.275, pruned_loss=0.07922, over 956213.85 frames. ], batch size: 34, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:39:44,762 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.812e+02 2.065e+02 2.463e+02 5.082e+02, threshold=4.130e+02, percent-clipped=4.0 +2023-03-26 04:40:02,054 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2442, 2.1377, 2.3938, 1.0942, 2.5250, 2.7210, 2.2797, 2.1701], + device='cuda:0'), covar=tensor([0.0938, 0.0643, 0.0356, 0.0696, 0.0379, 0.0741, 0.0373, 0.0606], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0159, 0.0119, 0.0138, 0.0134, 0.0123, 0.0149, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.9038e-05, 1.1800e-04, 8.6508e-05, 1.0063e-04, 9.6669e-05, 9.1254e-05, + 1.1064e-04, 1.0837e-04], device='cuda:0') +2023-03-26 04:40:09,021 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25858.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:40:10,735 INFO [finetune.py:976] (0/7) Epoch 5, batch 2950, loss[loss=0.2307, simple_loss=0.2854, pruned_loss=0.08799, over 4925.00 frames. ], tot_loss[loss=0.2195, simple_loss=0.2784, pruned_loss=0.08033, over 955665.67 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:40:43,942 INFO [finetune.py:976] (0/7) Epoch 5, batch 3000, loss[loss=0.1787, simple_loss=0.237, pruned_loss=0.06021, over 4730.00 frames. ], tot_loss[loss=0.2221, simple_loss=0.2813, pruned_loss=0.08142, over 957398.56 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:40:43,943 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 04:40:54,558 INFO [finetune.py:1010] (0/7) Epoch 5, validation: loss=0.1652, simple_loss=0.2371, pruned_loss=0.04667, over 2265189.00 frames. +2023-03-26 04:40:54,558 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 04:41:00,078 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25919.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:41:01,809 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.739e+02 2.096e+02 2.435e+02 4.160e+02, threshold=4.193e+02, percent-clipped=2.0 +2023-03-26 04:41:27,995 INFO [finetune.py:976] (0/7) Epoch 5, batch 3050, loss[loss=0.2389, simple_loss=0.2964, pruned_loss=0.09075, over 4706.00 frames. ], tot_loss[loss=0.2237, simple_loss=0.2832, pruned_loss=0.08207, over 956894.89 frames. ], batch size: 59, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:41:41,884 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 04:41:53,279 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-26000.pt +2023-03-26 04:42:08,229 INFO [finetune.py:976] (0/7) Epoch 5, batch 3100, loss[loss=0.2292, simple_loss=0.2875, pruned_loss=0.08543, over 4776.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2813, pruned_loss=0.08111, over 956558.03 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:42:25,415 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.603e+02 1.879e+02 2.413e+02 4.411e+02, threshold=3.758e+02, percent-clipped=2.0 +2023-03-26 04:43:10,311 INFO [finetune.py:976] (0/7) Epoch 5, batch 3150, loss[loss=0.2104, simple_loss=0.2717, pruned_loss=0.0745, over 4928.00 frames. ], tot_loss[loss=0.2203, simple_loss=0.2789, pruned_loss=0.08088, over 957035.66 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:43:30,920 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26084.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:43:42,331 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:43:49,562 INFO [finetune.py:976] (0/7) Epoch 5, batch 3200, loss[loss=0.2316, simple_loss=0.284, pruned_loss=0.08957, over 4899.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2752, pruned_loss=0.07949, over 957074.16 frames. ], batch size: 32, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:43:58,347 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.599e+02 1.999e+02 2.453e+02 4.323e+02, threshold=3.997e+02, percent-clipped=1.0 +2023-03-26 04:44:04,855 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26132.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:44:37,844 INFO [finetune.py:976] (0/7) Epoch 5, batch 3250, loss[loss=0.1772, simple_loss=0.2398, pruned_loss=0.05724, over 4796.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.2758, pruned_loss=0.07955, over 955069.27 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:44:43,909 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:44:47,971 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:45:40,772 INFO [finetune.py:976] (0/7) Epoch 5, batch 3300, loss[loss=0.2689, simple_loss=0.3363, pruned_loss=0.1007, over 4862.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2793, pruned_loss=0.08082, over 953939.30 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:45:47,653 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26214.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:46:00,062 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.721e+02 2.149e+02 2.490e+02 4.939e+02, threshold=4.298e+02, percent-clipped=1.0 +2023-03-26 04:46:06,004 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:46:11,930 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6260, 0.5817, 1.5666, 1.3463, 1.3032, 1.2328, 1.2495, 1.4759], + device='cuda:0'), covar=tensor([0.4938, 0.7493, 0.5731, 0.6332, 0.7019, 0.5394, 0.7692, 0.5360], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0248, 0.0254, 0.0258, 0.0241, 0.0218, 0.0274, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:46:14,894 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7735, 4.4613, 4.3385, 2.5634, 4.5226, 3.2637, 0.8395, 2.8993], + device='cuda:0'), covar=tensor([0.2853, 0.1376, 0.1300, 0.2970, 0.0841, 0.0975, 0.5192, 0.1426], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0174, 0.0166, 0.0130, 0.0157, 0.0124, 0.0148, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 04:46:19,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:46:29,542 INFO [finetune.py:976] (0/7) Epoch 5, batch 3350, loss[loss=0.2401, simple_loss=0.2992, pruned_loss=0.0905, over 4878.00 frames. ], tot_loss[loss=0.222, simple_loss=0.2815, pruned_loss=0.08124, over 955474.66 frames. ], batch size: 34, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:46:59,659 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2034, 3.6140, 3.7930, 4.0553, 3.9425, 3.6982, 4.3129, 1.4005], + device='cuda:0'), covar=tensor([0.0853, 0.0848, 0.0730, 0.0978, 0.1386, 0.1506, 0.0683, 0.5122], + device='cuda:0'), in_proj_covar=tensor([0.0358, 0.0245, 0.0275, 0.0293, 0.0337, 0.0284, 0.0305, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:47:12,172 INFO [finetune.py:976] (0/7) Epoch 5, batch 3400, loss[loss=0.2308, simple_loss=0.2894, pruned_loss=0.08608, over 4879.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2837, pruned_loss=0.08277, over 953360.86 frames. ], batch size: 43, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:47:12,304 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:47:19,911 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.624e+02 1.904e+02 2.361e+02 4.543e+02, threshold=3.807e+02, percent-clipped=1.0 +2023-03-26 04:47:31,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9006, 1.8270, 1.5007, 1.9228, 1.7907, 1.6771, 1.7084, 2.5453], + device='cuda:0'), covar=tensor([0.7508, 0.8803, 0.6189, 0.8227, 0.7307, 0.4343, 0.8369, 0.2665], + device='cuda:0'), in_proj_covar=tensor([0.0281, 0.0256, 0.0220, 0.0284, 0.0238, 0.0200, 0.0243, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:47:53,149 INFO [finetune.py:976] (0/7) Epoch 5, batch 3450, loss[loss=0.265, simple_loss=0.3226, pruned_loss=0.1037, over 4895.00 frames. ], tot_loss[loss=0.2235, simple_loss=0.2829, pruned_loss=0.08207, over 952808.26 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:48:37,958 INFO [finetune.py:976] (0/7) Epoch 5, batch 3500, loss[loss=0.1793, simple_loss=0.2479, pruned_loss=0.05538, over 4776.00 frames. ], tot_loss[loss=0.2206, simple_loss=0.2795, pruned_loss=0.08087, over 951712.83 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:48:43,880 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26414.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:48:51,309 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.586e+02 1.962e+02 2.229e+02 4.326e+02, threshold=3.925e+02, percent-clipped=1.0 +2023-03-26 04:49:23,753 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:49:25,514 INFO [finetune.py:976] (0/7) Epoch 5, batch 3550, loss[loss=0.1776, simple_loss=0.2433, pruned_loss=0.05592, over 4791.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2765, pruned_loss=0.08032, over 951011.89 frames. ], batch size: 29, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:49:32,242 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2643, 2.1510, 1.6517, 2.3775, 2.2523, 1.9349, 2.6297, 2.1831], + device='cuda:0'), covar=tensor([0.1673, 0.3198, 0.4023, 0.3301, 0.2897, 0.1853, 0.4098, 0.2262], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0193, 0.0235, 0.0253, 0.0229, 0.0188, 0.0210, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:49:34,668 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:50:11,928 INFO [finetune.py:976] (0/7) Epoch 5, batch 3600, loss[loss=0.2095, simple_loss=0.268, pruned_loss=0.0755, over 4935.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2737, pruned_loss=0.07928, over 950671.17 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:50:13,888 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26514.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:50:19,777 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.588e+02 1.920e+02 2.290e+02 3.397e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 04:50:20,520 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:50:55,027 INFO [finetune.py:976] (0/7) Epoch 5, batch 3650, loss[loss=0.1742, simple_loss=0.2402, pruned_loss=0.05412, over 4821.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2769, pruned_loss=0.08082, over 952800.43 frames. ], batch size: 30, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:50:55,692 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26562.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:50:57,712 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 04:51:10,034 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-26 04:51:14,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6139, 0.6081, 1.4431, 1.3436, 1.2594, 1.2431, 1.1864, 1.3509], + device='cuda:0'), covar=tensor([0.5494, 0.7654, 0.6426, 0.6572, 0.7923, 0.5886, 0.8611, 0.5966], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0248, 0.0255, 0.0258, 0.0241, 0.0218, 0.0274, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:51:31,358 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:51:34,853 INFO [finetune.py:976] (0/7) Epoch 5, batch 3700, loss[loss=0.1804, simple_loss=0.2552, pruned_loss=0.0528, over 4902.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.281, pruned_loss=0.08242, over 953593.33 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:51:42,594 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.866e+02 2.222e+02 2.784e+02 4.852e+02, threshold=4.444e+02, percent-clipped=6.0 +2023-03-26 04:51:59,432 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4128, 1.3463, 1.5982, 2.4586, 1.6774, 2.1277, 0.9540, 1.9499], + device='cuda:0'), covar=tensor([0.2007, 0.1592, 0.1273, 0.0815, 0.0984, 0.1329, 0.1639, 0.0860], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0166, 0.0102, 0.0142, 0.0129, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 04:52:02,454 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8220, 1.5363, 1.3289, 1.1206, 1.4791, 1.4890, 1.4152, 2.1485], + device='cuda:0'), covar=tensor([0.6474, 0.6224, 0.5123, 0.6607, 0.5637, 0.3524, 0.5902, 0.2397], + device='cuda:0'), in_proj_covar=tensor([0.0281, 0.0256, 0.0220, 0.0283, 0.0238, 0.0200, 0.0243, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:52:07,673 INFO [finetune.py:976] (0/7) Epoch 5, batch 3750, loss[loss=0.2488, simple_loss=0.3087, pruned_loss=0.09441, over 4854.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.2819, pruned_loss=0.08238, over 952523.40 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:52:37,002 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7970, 1.7874, 1.5348, 1.9038, 2.2550, 1.8053, 1.4277, 1.3966], + device='cuda:0'), covar=tensor([0.2504, 0.2368, 0.2200, 0.1891, 0.2100, 0.1334, 0.2901, 0.2147], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0208, 0.0198, 0.0183, 0.0234, 0.0173, 0.0213, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:53:00,962 INFO [finetune.py:976] (0/7) Epoch 5, batch 3800, loss[loss=0.2059, simple_loss=0.2671, pruned_loss=0.07238, over 4801.00 frames. ], tot_loss[loss=0.2252, simple_loss=0.2843, pruned_loss=0.08309, over 953615.30 frames. ], batch size: 45, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:53:08,702 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.710e+02 2.076e+02 2.649e+02 5.488e+02, threshold=4.152e+02, percent-clipped=2.0 +2023-03-26 04:53:30,035 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 04:53:37,457 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1472, 1.7431, 1.2749, 0.4786, 1.5281, 1.7533, 1.3871, 1.5868], + device='cuda:0'), covar=tensor([0.0791, 0.0998, 0.1578, 0.2147, 0.1552, 0.2609, 0.2608, 0.1083], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0202, 0.0204, 0.0191, 0.0218, 0.0211, 0.0223, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:53:39,326 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8127, 1.6421, 1.4103, 1.4206, 1.5660, 1.5265, 1.5651, 2.2857], + device='cuda:0'), covar=tensor([0.7270, 0.7132, 0.5493, 0.7288, 0.6349, 0.3858, 0.6673, 0.2738], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0257, 0.0221, 0.0284, 0.0239, 0.0201, 0.0244, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:53:57,420 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26758.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:53:59,695 INFO [finetune.py:976] (0/7) Epoch 5, batch 3850, loss[loss=0.1927, simple_loss=0.2514, pruned_loss=0.06697, over 4931.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.2824, pruned_loss=0.08211, over 953265.04 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:54:11,272 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:54:34,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0277, 1.4389, 1.0545, 2.0393, 2.3979, 1.7883, 1.7187, 1.8659], + device='cuda:0'), covar=tensor([0.1449, 0.2101, 0.2047, 0.1188, 0.1787, 0.1909, 0.1419, 0.1874], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0093, 0.0123, 0.0096, 0.0101, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 04:54:43,281 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-03-26 04:55:00,183 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:55:03,122 INFO [finetune.py:976] (0/7) Epoch 5, batch 3900, loss[loss=0.1879, simple_loss=0.2399, pruned_loss=0.06802, over 4720.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.2792, pruned_loss=0.08107, over 954407.52 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:55:21,752 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.797e+02 2.146e+02 2.516e+02 4.177e+02, threshold=4.292e+02, percent-clipped=1.0 +2023-03-26 04:55:22,461 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:55:22,535 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8640, 1.1482, 1.7426, 1.7026, 1.5464, 1.5203, 1.5626, 1.6088], + device='cuda:0'), covar=tensor([0.5844, 0.7644, 0.5873, 0.7001, 0.7350, 0.5741, 0.8497, 0.5650], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0248, 0.0255, 0.0259, 0.0242, 0.0218, 0.0274, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:55:57,245 INFO [finetune.py:976] (0/7) Epoch 5, batch 3950, loss[loss=0.1796, simple_loss=0.2504, pruned_loss=0.05435, over 4778.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2747, pruned_loss=0.07869, over 954396.92 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:56:06,702 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:56:12,894 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6214, 1.5078, 1.3297, 1.3801, 1.7997, 1.8859, 1.7389, 1.4045], + device='cuda:0'), covar=tensor([0.0308, 0.0315, 0.0509, 0.0326, 0.0230, 0.0381, 0.0234, 0.0345], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0112, 0.0137, 0.0117, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.7238e-05, 8.7889e-05, 1.0995e-04, 9.2524e-05, 8.2597e-05, 7.4713e-05, + 6.9342e-05, 8.4679e-05], device='cuda:0') +2023-03-26 04:56:29,747 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26904.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:56:30,946 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:56:34,935 INFO [finetune.py:976] (0/7) Epoch 5, batch 4000, loss[loss=0.2985, simple_loss=0.3509, pruned_loss=0.1231, over 4807.00 frames. ], tot_loss[loss=0.2158, simple_loss=0.2739, pruned_loss=0.07891, over 955904.74 frames. ], batch size: 45, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:56:43,169 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 1.662e+02 2.009e+02 2.453e+02 3.802e+02, threshold=4.018e+02, percent-clipped=0.0 +2023-03-26 04:57:05,869 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:57:08,737 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26954.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:57:18,754 INFO [finetune.py:976] (0/7) Epoch 5, batch 4050, loss[loss=0.1939, simple_loss=0.2808, pruned_loss=0.05346, over 4899.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.278, pruned_loss=0.08027, over 956006.63 frames. ], batch size: 37, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:57:19,448 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1387, 3.5776, 3.7768, 4.0222, 3.8427, 3.6682, 4.2072, 1.3703], + device='cuda:0'), covar=tensor([0.0793, 0.0820, 0.0769, 0.0975, 0.1237, 0.1459, 0.0718, 0.4965], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0242, 0.0272, 0.0291, 0.0335, 0.0283, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 04:57:26,826 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26965.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 04:58:10,820 INFO [finetune.py:976] (0/7) Epoch 5, batch 4100, loss[loss=0.2471, simple_loss=0.298, pruned_loss=0.09807, over 4854.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.2827, pruned_loss=0.08249, over 955590.73 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:58:10,952 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27011.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:58:19,550 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.740e+02 2.111e+02 2.577e+02 5.326e+02, threshold=4.223e+02, percent-clipped=3.0 +2023-03-26 04:58:30,245 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3213, 1.4422, 1.4063, 1.6329, 1.4609, 2.9008, 1.2302, 1.5075], + device='cuda:0'), covar=tensor([0.1038, 0.1743, 0.1188, 0.0996, 0.1734, 0.0275, 0.1607, 0.1723], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0082, 0.0077, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 04:58:58,625 INFO [finetune.py:976] (0/7) Epoch 5, batch 4150, loss[loss=0.2282, simple_loss=0.2877, pruned_loss=0.08437, over 4767.00 frames. ], tot_loss[loss=0.2253, simple_loss=0.2837, pruned_loss=0.08352, over 952247.09 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:59:05,622 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27070.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:59:32,487 INFO [finetune.py:976] (0/7) Epoch 5, batch 4200, loss[loss=0.1875, simple_loss=0.2549, pruned_loss=0.06007, over 4800.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2841, pruned_loss=0.08331, over 951031.25 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:59:37,724 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:59:41,614 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.682e+02 1.955e+02 2.295e+02 5.538e+02, threshold=3.911e+02, percent-clipped=3.0 +2023-03-26 04:59:57,812 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:00:00,853 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 05:00:10,425 INFO [finetune.py:976] (0/7) Epoch 5, batch 4250, loss[loss=0.2374, simple_loss=0.2974, pruned_loss=0.08871, over 4786.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2809, pruned_loss=0.08214, over 951014.56 frames. ], batch size: 51, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:00:11,216 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8596, 1.7256, 1.4525, 1.5151, 1.6807, 1.5943, 1.6630, 2.3767], + device='cuda:0'), covar=tensor([0.5790, 0.6159, 0.4464, 0.5879, 0.5142, 0.3297, 0.5508, 0.2112], + device='cuda:0'), in_proj_covar=tensor([0.0280, 0.0256, 0.0219, 0.0282, 0.0237, 0.0200, 0.0241, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:00:23,866 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.16 vs. limit=5.0 +2023-03-26 05:01:08,506 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:01:09,574 INFO [finetune.py:976] (0/7) Epoch 5, batch 4300, loss[loss=0.2248, simple_loss=0.2816, pruned_loss=0.08404, over 4902.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2768, pruned_loss=0.08059, over 951163.60 frames. ], batch size: 35, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:01:18,613 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9730, 1.7551, 1.5721, 1.8866, 1.7061, 1.6797, 1.6769, 2.4808], + device='cuda:0'), covar=tensor([0.6452, 0.7362, 0.5243, 0.7039, 0.6479, 0.3933, 0.7086, 0.2428], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0221, 0.0285, 0.0239, 0.0201, 0.0244, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:01:26,945 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.762e+02 2.023e+02 2.453e+02 1.035e+03, threshold=4.046e+02, percent-clipped=2.0 +2023-03-26 05:01:52,144 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 05:01:59,787 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27260.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:02:00,300 INFO [finetune.py:976] (0/7) Epoch 5, batch 4350, loss[loss=0.2337, simple_loss=0.2828, pruned_loss=0.09234, over 4743.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2732, pruned_loss=0.07872, over 953407.21 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:02:18,884 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 05:02:30,556 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:02:33,996 INFO [finetune.py:976] (0/7) Epoch 5, batch 4400, loss[loss=0.242, simple_loss=0.3107, pruned_loss=0.0867, over 4821.00 frames. ], tot_loss[loss=0.2173, simple_loss=0.275, pruned_loss=0.07984, over 950698.32 frames. ], batch size: 39, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:02:41,212 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.601e+02 1.888e+02 2.389e+02 3.644e+02, threshold=3.775e+02, percent-clipped=0.0 +2023-03-26 05:03:02,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4400, 1.3086, 1.2740, 1.4744, 1.6366, 1.4107, 0.8673, 1.2111], + device='cuda:0'), covar=tensor([0.2493, 0.2515, 0.2198, 0.1991, 0.1846, 0.1419, 0.3161, 0.2079], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0210, 0.0202, 0.0185, 0.0237, 0.0176, 0.0215, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:03:07,514 INFO [finetune.py:976] (0/7) Epoch 5, batch 4450, loss[loss=0.1868, simple_loss=0.2337, pruned_loss=0.06992, over 4018.00 frames. ], tot_loss[loss=0.2225, simple_loss=0.2802, pruned_loss=0.08238, over 946881.78 frames. ], batch size: 17, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:03:12,521 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6860, 1.7104, 1.7741, 1.1692, 1.7745, 2.0195, 1.9567, 1.5358], + device='cuda:0'), covar=tensor([0.0985, 0.0659, 0.0442, 0.0572, 0.0378, 0.0422, 0.0338, 0.0689], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0160, 0.0121, 0.0137, 0.0134, 0.0124, 0.0148, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.8606e-05, 1.1833e-04, 8.7590e-05, 1.0009e-04, 9.6166e-05, 9.1648e-05, + 1.1023e-04, 1.0856e-04], device='cuda:0') +2023-03-26 05:03:18,334 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8544, 1.7707, 1.4286, 1.7032, 1.9073, 1.5610, 2.2231, 1.8832], + device='cuda:0'), covar=tensor([0.1807, 0.2937, 0.3827, 0.3271, 0.2877, 0.2033, 0.3761, 0.2305], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0193, 0.0236, 0.0254, 0.0230, 0.0189, 0.0211, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:03:38,515 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 05:03:40,731 INFO [finetune.py:976] (0/7) Epoch 5, batch 4500, loss[loss=0.1977, simple_loss=0.2667, pruned_loss=0.06437, over 4921.00 frames. ], tot_loss[loss=0.224, simple_loss=0.2815, pruned_loss=0.08324, over 947982.36 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:03:48,439 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.291e+02 1.723e+02 2.077e+02 2.543e+02 6.449e+02, threshold=4.154e+02, percent-clipped=4.0 +2023-03-26 05:04:14,225 INFO [finetune.py:976] (0/7) Epoch 5, batch 4550, loss[loss=0.1989, simple_loss=0.2646, pruned_loss=0.0666, over 4730.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.2823, pruned_loss=0.08306, over 947486.29 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:04:30,957 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0136, 1.8156, 1.5751, 1.7814, 1.7478, 1.7615, 1.7609, 2.5349], + device='cuda:0'), covar=tensor([0.6057, 0.7223, 0.5131, 0.7186, 0.6286, 0.3563, 0.6469, 0.2414], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0255, 0.0218, 0.0281, 0.0236, 0.0199, 0.0241, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:04:42,681 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:04:47,379 INFO [finetune.py:976] (0/7) Epoch 5, batch 4600, loss[loss=0.1958, simple_loss=0.2591, pruned_loss=0.06624, over 4886.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2797, pruned_loss=0.081, over 947295.51 frames. ], batch size: 43, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:04:55,105 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.596e+02 2.009e+02 2.524e+02 8.514e+02, threshold=4.018e+02, percent-clipped=5.0 +2023-03-26 05:05:15,392 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 05:05:20,021 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27560.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:05:20,542 INFO [finetune.py:976] (0/7) Epoch 5, batch 4650, loss[loss=0.2373, simple_loss=0.2888, pruned_loss=0.09284, over 4298.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.2772, pruned_loss=0.07989, over 947537.21 frames. ], batch size: 65, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:05:22,918 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2203, 2.0863, 1.6468, 0.7917, 1.8825, 1.8366, 1.6220, 1.9163], + device='cuda:0'), covar=tensor([0.0902, 0.0667, 0.1455, 0.1902, 0.1286, 0.2251, 0.2153, 0.0854], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0202, 0.0205, 0.0192, 0.0220, 0.0212, 0.0224, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:06:04,731 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1464, 1.6488, 1.8316, 1.9185, 1.7206, 1.7076, 1.8142, 1.8698], + device='cuda:0'), covar=tensor([0.6282, 0.8513, 0.6572, 0.8210, 0.8951, 0.6682, 1.1138, 0.6159], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0248, 0.0255, 0.0258, 0.0241, 0.0218, 0.0275, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:06:06,519 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7003, 1.6481, 1.5864, 0.9211, 1.6752, 1.8171, 1.7562, 1.4798], + device='cuda:0'), covar=tensor([0.0824, 0.0515, 0.0457, 0.0589, 0.0373, 0.0524, 0.0332, 0.0589], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0159, 0.0121, 0.0137, 0.0134, 0.0124, 0.0149, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.8724e-05, 1.1792e-04, 8.7882e-05, 1.0016e-04, 9.6362e-05, 9.1810e-05, + 1.1042e-04, 1.0874e-04], device='cuda:0') +2023-03-26 05:06:07,076 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:06:13,640 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27608.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:06:15,418 INFO [finetune.py:976] (0/7) Epoch 5, batch 4700, loss[loss=0.2362, simple_loss=0.2847, pruned_loss=0.09388, over 4817.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2751, pruned_loss=0.07952, over 949345.24 frames. ], batch size: 41, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:06:27,906 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.549e+02 1.903e+02 2.293e+02 3.137e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 05:06:48,391 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8347, 1.3495, 0.9244, 1.7785, 2.3148, 1.4037, 1.6211, 1.7793], + device='cuda:0'), covar=tensor([0.1538, 0.2215, 0.2272, 0.1297, 0.1881, 0.2096, 0.1554, 0.2118], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0098, 0.0116, 0.0094, 0.0125, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:07:05,705 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:07:14,261 INFO [finetune.py:976] (0/7) Epoch 5, batch 4750, loss[loss=0.2485, simple_loss=0.3024, pruned_loss=0.09736, over 4819.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2746, pruned_loss=0.07942, over 950316.81 frames. ], batch size: 45, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:07:46,486 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.61 vs. limit=2.0 +2023-03-26 05:07:56,926 INFO [finetune.py:976] (0/7) Epoch 5, batch 4800, loss[loss=0.2003, simple_loss=0.2582, pruned_loss=0.07126, over 4747.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2762, pruned_loss=0.0801, over 947732.80 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:07:59,323 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27714.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:08:04,701 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.708e+02 2.066e+02 2.363e+02 4.852e+02, threshold=4.133e+02, percent-clipped=3.0 +2023-03-26 05:08:30,288 INFO [finetune.py:976] (0/7) Epoch 5, batch 4850, loss[loss=0.2436, simple_loss=0.3076, pruned_loss=0.08982, over 4806.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.2798, pruned_loss=0.08077, over 950580.69 frames. ], batch size: 45, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:08:30,390 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5250, 1.7089, 0.9850, 2.2483, 2.5546, 1.8333, 2.1292, 2.1763], + device='cuda:0'), covar=tensor([0.1386, 0.1952, 0.2129, 0.1140, 0.1844, 0.1905, 0.1309, 0.1926], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0098, 0.0115, 0.0094, 0.0125, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:08:39,498 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27775.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:08:50,746 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2334, 2.1547, 1.7086, 2.3781, 2.3406, 1.8319, 2.9146, 2.2535], + device='cuda:0'), covar=tensor([0.1580, 0.3557, 0.4030, 0.3677, 0.2856, 0.1957, 0.3896, 0.2112], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0193, 0.0235, 0.0254, 0.0229, 0.0189, 0.0210, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:08:59,122 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:09:04,314 INFO [finetune.py:976] (0/7) Epoch 5, batch 4900, loss[loss=0.2085, simple_loss=0.2814, pruned_loss=0.06782, over 4841.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2813, pruned_loss=0.08105, over 950698.77 frames. ], batch size: 49, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:09:12,047 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.126e+02 1.628e+02 1.864e+02 2.335e+02 3.818e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-26 05:09:30,670 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:09:37,160 INFO [finetune.py:976] (0/7) Epoch 5, batch 4950, loss[loss=0.2318, simple_loss=0.2899, pruned_loss=0.08685, over 4914.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2816, pruned_loss=0.08078, over 953278.03 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:09:42,193 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 05:09:59,928 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9333, 1.3540, 0.7387, 1.7441, 2.1721, 1.4532, 1.6930, 1.8731], + device='cuda:0'), covar=tensor([0.1449, 0.2039, 0.2350, 0.1205, 0.1927, 0.1933, 0.1355, 0.1890], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0093, 0.0124, 0.0096, 0.0100, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:10:10,462 INFO [finetune.py:976] (0/7) Epoch 5, batch 5000, loss[loss=0.2181, simple_loss=0.2775, pruned_loss=0.07936, over 4758.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2799, pruned_loss=0.08026, over 952986.33 frames. ], batch size: 28, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:10:19,080 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.230e+02 1.618e+02 1.837e+02 2.301e+02 4.829e+02, threshold=3.674e+02, percent-clipped=1.0 +2023-03-26 05:10:43,561 INFO [finetune.py:976] (0/7) Epoch 5, batch 5050, loss[loss=0.1647, simple_loss=0.2213, pruned_loss=0.05402, over 4804.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2773, pruned_loss=0.08009, over 953365.51 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:10:49,779 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8641, 3.3785, 3.4809, 3.7285, 3.6152, 3.3798, 3.9247, 1.2761], + device='cuda:0'), covar=tensor([0.0936, 0.0891, 0.0928, 0.0990, 0.1420, 0.1631, 0.0826, 0.5030], + device='cuda:0'), in_proj_covar=tensor([0.0358, 0.0246, 0.0278, 0.0295, 0.0340, 0.0287, 0.0306, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:10:51,046 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 05:11:12,334 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6319, 1.4504, 1.3962, 1.5044, 1.2066, 3.2999, 1.4188, 1.8490], + device='cuda:0'), covar=tensor([0.4151, 0.3141, 0.2515, 0.2992, 0.1983, 0.0260, 0.2534, 0.1319], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0117, 0.0098, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:11:31,888 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-28000.pt +2023-03-26 05:11:48,643 INFO [finetune.py:976] (0/7) Epoch 5, batch 5100, loss[loss=0.2309, simple_loss=0.2724, pruned_loss=0.09468, over 4777.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2732, pruned_loss=0.07864, over 955172.34 frames. ], batch size: 54, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:12:02,949 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.639e+02 1.875e+02 2.408e+02 3.954e+02, threshold=3.749e+02, percent-clipped=2.0 +2023-03-26 05:12:24,030 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 05:12:24,568 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5361, 1.4219, 1.3193, 1.5411, 1.5437, 1.5327, 0.8235, 1.3044], + device='cuda:0'), covar=tensor([0.1988, 0.2012, 0.1797, 0.1599, 0.1583, 0.1132, 0.2758, 0.1686], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0211, 0.0204, 0.0186, 0.0239, 0.0178, 0.0217, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:12:32,818 INFO [finetune.py:976] (0/7) Epoch 5, batch 5150, loss[loss=0.2679, simple_loss=0.3218, pruned_loss=0.107, over 4908.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2737, pruned_loss=0.07928, over 953361.10 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:12:38,915 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28070.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:13:04,570 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7442, 1.1870, 0.9682, 1.5852, 2.0272, 1.4787, 1.6159, 1.7548], + device='cuda:0'), covar=tensor([0.1589, 0.2206, 0.2103, 0.1285, 0.2080, 0.2132, 0.1396, 0.2023], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0098, 0.0116, 0.0094, 0.0125, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:13:06,309 INFO [finetune.py:976] (0/7) Epoch 5, batch 5200, loss[loss=0.2617, simple_loss=0.3163, pruned_loss=0.1036, over 4851.00 frames. ], tot_loss[loss=0.2195, simple_loss=0.2785, pruned_loss=0.08027, over 955979.67 frames. ], batch size: 31, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:13:11,869 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 05:13:14,538 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.308e+02 1.748e+02 1.996e+02 2.342e+02 5.311e+02, threshold=3.992e+02, percent-clipped=1.0 +2023-03-26 05:13:17,007 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1206, 1.9890, 1.6300, 1.9258, 1.9739, 1.8999, 1.8843, 2.8683], + device='cuda:0'), covar=tensor([0.6080, 0.6587, 0.4927, 0.6483, 0.6110, 0.3522, 0.6522, 0.1991], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0257, 0.0220, 0.0283, 0.0239, 0.0201, 0.0243, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:13:18,805 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7496, 1.5798, 1.4799, 1.3648, 1.8263, 1.5120, 1.7916, 1.7217], + device='cuda:0'), covar=tensor([0.1849, 0.3213, 0.4043, 0.3356, 0.3097, 0.2124, 0.3506, 0.2475], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0193, 0.0236, 0.0254, 0.0230, 0.0189, 0.0211, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:13:39,346 INFO [finetune.py:976] (0/7) Epoch 5, batch 5250, loss[loss=0.2283, simple_loss=0.2856, pruned_loss=0.08556, over 4902.00 frames. ], tot_loss[loss=0.221, simple_loss=0.28, pruned_loss=0.081, over 954877.82 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:13:49,281 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 05:13:54,575 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-26 05:14:12,318 INFO [finetune.py:976] (0/7) Epoch 5, batch 5300, loss[loss=0.2065, simple_loss=0.2726, pruned_loss=0.07019, over 4798.00 frames. ], tot_loss[loss=0.2222, simple_loss=0.2812, pruned_loss=0.0816, over 955423.44 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:14:19,562 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.725e+02 1.957e+02 2.435e+02 6.444e+02, threshold=3.915e+02, percent-clipped=2.0 +2023-03-26 05:14:24,316 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:14:51,819 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28255.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:14:56,275 INFO [finetune.py:976] (0/7) Epoch 5, batch 5350, loss[loss=0.2251, simple_loss=0.2767, pruned_loss=0.08679, over 4833.00 frames. ], tot_loss[loss=0.2212, simple_loss=0.2805, pruned_loss=0.08092, over 954130.23 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:15:13,542 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28287.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:15:15,774 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:15:27,333 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5197, 1.4120, 1.3577, 1.5789, 1.6819, 1.5827, 0.9541, 1.3255], + device='cuda:0'), covar=tensor([0.2224, 0.2172, 0.1881, 0.1713, 0.1825, 0.1213, 0.2927, 0.1886], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0210, 0.0202, 0.0186, 0.0237, 0.0177, 0.0215, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:15:28,416 INFO [finetune.py:976] (0/7) Epoch 5, batch 5400, loss[loss=0.1741, simple_loss=0.2395, pruned_loss=0.05433, over 4921.00 frames. ], tot_loss[loss=0.2173, simple_loss=0.2768, pruned_loss=0.07891, over 954405.78 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:15:31,973 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28316.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:15:36,016 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.900e+01 1.568e+02 1.878e+02 2.260e+02 3.573e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-26 05:15:42,174 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9107, 1.4014, 1.8055, 1.7363, 1.5459, 1.5366, 1.6394, 1.6330], + device='cuda:0'), covar=tensor([0.5785, 0.7656, 0.6028, 0.6941, 0.7978, 0.6198, 0.8992, 0.6005], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0247, 0.0254, 0.0257, 0.0241, 0.0219, 0.0273, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:15:42,227 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.69 vs. limit=2.0 +2023-03-26 05:15:49,135 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0327, 1.8865, 1.5597, 1.9426, 1.8095, 1.7470, 1.7469, 2.5635], + device='cuda:0'), covar=tensor([0.6346, 0.7581, 0.5364, 0.6412, 0.6201, 0.3753, 0.6669, 0.2372], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0283, 0.0239, 0.0201, 0.0243, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:15:53,752 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:16:01,447 INFO [finetune.py:976] (0/7) Epoch 5, batch 5450, loss[loss=0.2037, simple_loss=0.2672, pruned_loss=0.07011, over 4837.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2742, pruned_loss=0.07807, over 956395.95 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:16:12,270 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9704, 1.7411, 1.4864, 1.6240, 1.7131, 1.6723, 1.6507, 2.4893], + device='cuda:0'), covar=tensor([0.6670, 0.7071, 0.5331, 0.6827, 0.5868, 0.3912, 0.6540, 0.2309], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0283, 0.0238, 0.0201, 0.0243, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:16:18,309 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28370.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:16:22,667 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 05:17:02,777 INFO [finetune.py:976] (0/7) Epoch 5, batch 5500, loss[loss=0.2134, simple_loss=0.2691, pruned_loss=0.07889, over 4868.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2712, pruned_loss=0.07768, over 954438.20 frames. ], batch size: 44, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:17:12,775 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28418.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:17:21,150 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.651e+02 2.036e+02 2.478e+02 5.642e+02, threshold=4.072e+02, percent-clipped=3.0 +2023-03-26 05:18:07,648 INFO [finetune.py:976] (0/7) Epoch 5, batch 5550, loss[loss=0.2361, simple_loss=0.283, pruned_loss=0.09455, over 4864.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.273, pruned_loss=0.07863, over 954656.34 frames. ], batch size: 31, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:18:36,677 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:18:45,467 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:18:54,308 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5480, 1.5371, 1.6953, 1.8070, 1.6516, 2.9212, 1.3924, 1.6566], + device='cuda:0'), covar=tensor([0.0913, 0.1609, 0.1343, 0.0898, 0.1408, 0.0316, 0.1323, 0.1479], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0081, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:19:02,060 INFO [finetune.py:976] (0/7) Epoch 5, batch 5600, loss[loss=0.2478, simple_loss=0.302, pruned_loss=0.09684, over 4745.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.277, pruned_loss=0.08004, over 954055.14 frames. ], batch size: 54, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:19:14,573 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.753e+02 2.098e+02 2.591e+02 4.684e+02, threshold=4.196e+02, percent-clipped=2.0 +2023-03-26 05:19:40,170 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:19:47,580 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:19:52,218 INFO [finetune.py:976] (0/7) Epoch 5, batch 5650, loss[loss=0.1853, simple_loss=0.2605, pruned_loss=0.05508, over 4838.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2795, pruned_loss=0.0804, over 954705.91 frames. ], batch size: 47, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:20:06,329 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:20:21,893 INFO [finetune.py:976] (0/7) Epoch 5, batch 5700, loss[loss=0.2022, simple_loss=0.2583, pruned_loss=0.07305, over 3995.00 frames. ], tot_loss[loss=0.2176, simple_loss=0.2758, pruned_loss=0.07971, over 940020.74 frames. ], batch size: 17, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:20:21,934 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28611.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 05:20:29,013 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.580e+02 1.894e+02 2.210e+02 5.665e+02, threshold=3.789e+02, percent-clipped=1.0 +2023-03-26 05:20:29,182 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-03-26 05:20:38,899 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-5.pt +2023-03-26 05:20:53,188 INFO [finetune.py:976] (0/7) Epoch 6, batch 0, loss[loss=0.2267, simple_loss=0.2871, pruned_loss=0.08319, over 4912.00 frames. ], tot_loss[loss=0.2267, simple_loss=0.2871, pruned_loss=0.08319, over 4912.00 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:20:53,189 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 05:20:56,529 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4440, 1.5358, 1.4935, 1.6167, 1.6543, 2.9206, 1.4096, 1.6402], + device='cuda:0'), covar=tensor([0.1025, 0.1746, 0.1118, 0.1015, 0.1491, 0.0339, 0.1422, 0.1668], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0083, 0.0078, 0.0081, 0.0094, 0.0084, 0.0087, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:21:08,973 INFO [finetune.py:1010] (0/7) Epoch 6, validation: loss=0.1659, simple_loss=0.2379, pruned_loss=0.04693, over 2265189.00 frames. +2023-03-26 05:21:08,974 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 05:21:15,235 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:21:59,715 INFO [finetune.py:976] (0/7) Epoch 6, batch 50, loss[loss=0.2561, simple_loss=0.3109, pruned_loss=0.1007, over 4875.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2826, pruned_loss=0.08444, over 216273.13 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:22:30,497 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.667e+02 1.978e+02 2.446e+02 6.098e+02, threshold=3.955e+02, percent-clipped=3.0 +2023-03-26 05:22:41,774 INFO [finetune.py:976] (0/7) Epoch 6, batch 100, loss[loss=0.1881, simple_loss=0.2536, pruned_loss=0.06127, over 4909.00 frames. ], tot_loss[loss=0.2197, simple_loss=0.2768, pruned_loss=0.08131, over 380334.85 frames. ], batch size: 36, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:23:15,383 INFO [finetune.py:976] (0/7) Epoch 6, batch 150, loss[loss=0.2106, simple_loss=0.2681, pruned_loss=0.07655, over 4879.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2714, pruned_loss=0.07919, over 507437.93 frames. ], batch size: 34, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:23:15,558 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-03-26 05:23:37,615 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.593e+02 1.877e+02 2.260e+02 4.734e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-26 05:23:42,565 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 05:23:48,134 INFO [finetune.py:976] (0/7) Epoch 6, batch 200, loss[loss=0.2263, simple_loss=0.275, pruned_loss=0.0888, over 4814.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2693, pruned_loss=0.07864, over 604198.43 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:23:50,546 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:23:55,125 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:03,966 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28861.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:19,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:26,390 INFO [finetune.py:976] (0/7) Epoch 6, batch 250, loss[loss=0.2177, simple_loss=0.2858, pruned_loss=0.07482, over 4718.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2751, pruned_loss=0.08123, over 681390.05 frames. ], batch size: 59, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:24:37,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4660, 2.1602, 1.9730, 2.3032, 2.1900, 2.0906, 2.0315, 3.1463], + device='cuda:0'), covar=tensor([0.5959, 0.8848, 0.5495, 0.7988, 0.6558, 0.3875, 0.7806, 0.2374], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0284, 0.0238, 0.0202, 0.0245, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:24:37,516 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 05:24:47,331 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28911.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:25:02,678 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28922.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:03,152 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.635e+02 1.950e+02 2.422e+02 4.878e+02, threshold=3.900e+02, percent-clipped=5.0 +2023-03-26 05:25:09,281 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:13,903 INFO [finetune.py:976] (0/7) Epoch 6, batch 300, loss[loss=0.2465, simple_loss=0.3019, pruned_loss=0.09553, over 4933.00 frames. ], tot_loss[loss=0.2235, simple_loss=0.2807, pruned_loss=0.08315, over 742259.89 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:25:16,451 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28943.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:27,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3811, 2.2441, 1.8658, 2.4588, 2.4702, 1.9442, 2.9792, 2.3473], + device='cuda:0'), covar=tensor([0.1713, 0.3649, 0.3984, 0.3834, 0.2753, 0.1964, 0.3077, 0.2271], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0193, 0.0236, 0.0255, 0.0231, 0.0190, 0.0211, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:25:28,657 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28959.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:25:47,451 INFO [finetune.py:976] (0/7) Epoch 6, batch 350, loss[loss=0.2463, simple_loss=0.2925, pruned_loss=0.1001, over 4791.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2827, pruned_loss=0.08327, over 789333.86 frames. ], batch size: 51, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:25:49,396 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:26:02,695 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:26:28,284 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 1.836e+02 2.201e+02 2.620e+02 4.241e+02, threshold=4.402e+02, percent-clipped=2.0 +2023-03-26 05:26:38,005 INFO [finetune.py:976] (0/7) Epoch 6, batch 400, loss[loss=0.2485, simple_loss=0.2965, pruned_loss=0.1003, over 4844.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.2818, pruned_loss=0.08222, over 825625.90 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:01,628 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29063.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:27:17,310 INFO [finetune.py:976] (0/7) Epoch 6, batch 450, loss[loss=0.2089, simple_loss=0.269, pruned_loss=0.07434, over 4726.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2784, pruned_loss=0.07999, over 854262.50 frames. ], batch size: 54, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:35,912 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 05:27:45,345 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.708e+02 1.996e+02 2.284e+02 5.200e+02, threshold=3.993e+02, percent-clipped=1.0 +2023-03-26 05:27:55,104 INFO [finetune.py:976] (0/7) Epoch 6, batch 500, loss[loss=0.2078, simple_loss=0.2636, pruned_loss=0.076, over 4709.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2751, pruned_loss=0.07836, over 877381.83 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:57,040 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:01,658 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:28,349 INFO [finetune.py:976] (0/7) Epoch 6, batch 550, loss[loss=0.2626, simple_loss=0.3194, pruned_loss=0.1029, over 4805.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.272, pruned_loss=0.07692, over 896416.50 frames. ], batch size: 45, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:28:28,996 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:31,402 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6786, 1.1769, 0.8187, 1.6183, 2.1337, 1.4028, 1.4278, 1.6737], + device='cuda:0'), covar=tensor([0.1468, 0.2064, 0.1975, 0.1140, 0.1823, 0.1979, 0.1465, 0.1853], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0116, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:28:34,378 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:54,922 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9901, 2.1097, 1.9359, 1.3897, 2.2688, 2.2639, 2.2264, 1.8811], + device='cuda:0'), covar=tensor([0.0666, 0.0623, 0.0792, 0.1016, 0.0509, 0.0631, 0.0589, 0.1002], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0134, 0.0144, 0.0128, 0.0113, 0.0143, 0.0146, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:29:00,092 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:29:04,678 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.638e+02 2.076e+02 2.525e+02 4.090e+02, threshold=4.153e+02, percent-clipped=1.0 +2023-03-26 05:29:18,480 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:29:20,238 INFO [finetune.py:976] (0/7) Epoch 6, batch 600, loss[loss=0.2056, simple_loss=0.2727, pruned_loss=0.06923, over 4912.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2741, pruned_loss=0.07806, over 911223.62 frames. ], batch size: 37, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:29:53,120 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-26 05:29:59,821 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5395, 1.1383, 0.7878, 1.4658, 1.9155, 1.0633, 1.3048, 1.5492], + device='cuda:0'), covar=tensor([0.2253, 0.3177, 0.2687, 0.1746, 0.2695, 0.2852, 0.2193, 0.2853], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:30:24,096 INFO [finetune.py:976] (0/7) Epoch 6, batch 650, loss[loss=0.1623, simple_loss=0.2288, pruned_loss=0.04787, over 4773.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.2769, pruned_loss=0.07895, over 921013.10 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:30:34,097 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:30:34,189 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-03-26 05:31:13,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.735e+02 1.982e+02 2.438e+02 4.583e+02, threshold=3.965e+02, percent-clipped=2.0 +2023-03-26 05:31:33,679 INFO [finetune.py:976] (0/7) Epoch 6, batch 700, loss[loss=0.2363, simple_loss=0.2938, pruned_loss=0.08941, over 4860.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2773, pruned_loss=0.07888, over 927242.08 frames. ], batch size: 34, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:31:35,684 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 05:31:45,811 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:32:00,393 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7784, 0.9141, 1.6565, 1.4963, 1.4307, 1.3922, 1.3859, 1.5000], + device='cuda:0'), covar=tensor([0.5055, 0.7100, 0.5733, 0.6418, 0.7164, 0.5497, 0.7722, 0.5567], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0247, 0.0254, 0.0256, 0.0241, 0.0218, 0.0273, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:32:17,027 INFO [finetune.py:976] (0/7) Epoch 6, batch 750, loss[loss=0.2577, simple_loss=0.3118, pruned_loss=0.1018, over 4886.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2777, pruned_loss=0.07917, over 932666.91 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:32:57,446 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3478, 1.4708, 1.4227, 0.8986, 1.3457, 1.5776, 1.6522, 1.3214], + device='cuda:0'), covar=tensor([0.1039, 0.0590, 0.0513, 0.0520, 0.0486, 0.0623, 0.0321, 0.0683], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0161, 0.0123, 0.0139, 0.0135, 0.0125, 0.0149, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.9464e-05, 1.1885e-04, 8.9112e-05, 1.0124e-04, 9.7035e-05, 9.2586e-05, + 1.1069e-04, 1.0940e-04], device='cuda:0') +2023-03-26 05:32:59,101 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.656e+01 1.814e+02 2.109e+02 2.501e+02 5.044e+02, threshold=4.217e+02, percent-clipped=3.0 +2023-03-26 05:33:26,483 INFO [finetune.py:976] (0/7) Epoch 6, batch 800, loss[loss=0.2204, simple_loss=0.2825, pruned_loss=0.07915, over 4849.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2783, pruned_loss=0.07904, over 938692.79 frames. ], batch size: 44, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:33:40,153 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-26 05:34:10,553 INFO [finetune.py:976] (0/7) Epoch 6, batch 850, loss[loss=0.1827, simple_loss=0.2548, pruned_loss=0.05532, over 4783.00 frames. ], tot_loss[loss=0.2177, simple_loss=0.2774, pruned_loss=0.07896, over 944062.68 frames. ], batch size: 26, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:34:42,987 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1160, 0.8654, 0.9241, 1.1784, 1.2251, 1.2213, 1.0220, 1.0069], + device='cuda:0'), covar=tensor([0.0280, 0.0331, 0.0559, 0.0265, 0.0261, 0.0357, 0.0288, 0.0391], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0112, 0.0138, 0.0117, 0.0103, 0.0100, 0.0091, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.8103e-05, 8.7857e-05, 1.1029e-04, 9.2651e-05, 8.1619e-05, 7.4589e-05, + 6.9603e-05, 8.4474e-05], device='cuda:0') +2023-03-26 05:34:43,561 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29517.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:34:52,273 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.650e+02 2.021e+02 2.394e+02 5.702e+02, threshold=4.042e+02, percent-clipped=1.0 +2023-03-26 05:35:14,735 INFO [finetune.py:976] (0/7) Epoch 6, batch 900, loss[loss=0.2231, simple_loss=0.2767, pruned_loss=0.08474, over 4844.00 frames. ], tot_loss[loss=0.214, simple_loss=0.2734, pruned_loss=0.07729, over 943510.38 frames. ], batch size: 44, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:35:32,170 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:35:46,187 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:05,319 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0082, 1.3260, 1.6447, 1.6899, 1.4889, 1.5253, 1.5989, 1.6162], + device='cuda:0'), covar=tensor([0.7423, 0.9185, 0.7979, 0.8412, 1.0084, 0.7979, 1.0963, 0.7787], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0245, 0.0252, 0.0254, 0.0240, 0.0217, 0.0272, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:36:14,402 INFO [finetune.py:976] (0/7) Epoch 6, batch 950, loss[loss=0.2269, simple_loss=0.2962, pruned_loss=0.07879, over 4810.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2716, pruned_loss=0.07685, over 945678.91 frames. ], batch size: 41, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:36:21,348 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:43,480 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:56,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.661e+02 1.977e+02 2.359e+02 4.237e+02, threshold=3.954e+02, percent-clipped=2.0 +2023-03-26 05:37:07,982 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3936, 2.1828, 1.8231, 0.8314, 1.8818, 1.7641, 1.6528, 1.9408], + device='cuda:0'), covar=tensor([0.1002, 0.0919, 0.1742, 0.2419, 0.1665, 0.2653, 0.2401, 0.1180], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0202, 0.0190, 0.0217, 0.0210, 0.0220, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:37:17,829 INFO [finetune.py:976] (0/7) Epoch 6, batch 1000, loss[loss=0.2065, simple_loss=0.2612, pruned_loss=0.07595, over 4716.00 frames. ], tot_loss[loss=0.2141, simple_loss=0.2735, pruned_loss=0.07734, over 948147.12 frames. ], batch size: 23, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:37:40,145 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:37:47,491 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29662.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:38:20,623 INFO [finetune.py:976] (0/7) Epoch 6, batch 1050, loss[loss=0.1971, simple_loss=0.2569, pruned_loss=0.06859, over 4211.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2758, pruned_loss=0.07813, over 949568.71 frames. ], batch size: 66, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:38:40,529 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 05:38:41,602 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29706.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:39:02,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.795e+02 2.095e+02 2.533e+02 7.754e+02, threshold=4.191e+02, percent-clipped=4.0 +2023-03-26 05:39:03,056 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29723.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 05:39:20,231 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-26 05:39:23,375 INFO [finetune.py:976] (0/7) Epoch 6, batch 1100, loss[loss=0.1761, simple_loss=0.2397, pruned_loss=0.05624, over 4828.00 frames. ], tot_loss[loss=0.218, simple_loss=0.278, pruned_loss=0.07902, over 949942.84 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:39:43,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 05:40:24,912 INFO [finetune.py:976] (0/7) Epoch 6, batch 1150, loss[loss=0.2338, simple_loss=0.3014, pruned_loss=0.08314, over 4896.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2793, pruned_loss=0.07931, over 951164.03 frames. ], batch size: 46, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:40:37,982 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:40:38,614 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1250, 1.9738, 1.6813, 1.7809, 1.8789, 1.8674, 1.8799, 2.6006], + device='cuda:0'), covar=tensor([0.5865, 0.5627, 0.4797, 0.5357, 0.5172, 0.3507, 0.5304, 0.2051], + device='cuda:0'), in_proj_covar=tensor([0.0281, 0.0258, 0.0219, 0.0285, 0.0240, 0.0202, 0.0245, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:41:01,070 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.742e+02 2.057e+02 2.377e+02 6.600e+02, threshold=4.115e+02, percent-clipped=1.0 +2023-03-26 05:41:11,693 INFO [finetune.py:976] (0/7) Epoch 6, batch 1200, loss[loss=0.1857, simple_loss=0.2544, pruned_loss=0.0585, over 4798.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.277, pruned_loss=0.07827, over 953270.14 frames. ], batch size: 45, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:41:28,762 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29863.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:41:32,948 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2630, 2.0189, 2.7541, 1.7072, 2.5254, 2.6427, 1.9820, 2.6872], + device='cuda:0'), covar=tensor([0.1416, 0.2007, 0.1332, 0.2289, 0.0797, 0.1346, 0.2441, 0.0772], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0203, 0.0198, 0.0194, 0.0183, 0.0219, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:41:43,700 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 05:41:45,292 INFO [finetune.py:976] (0/7) Epoch 6, batch 1250, loss[loss=0.195, simple_loss=0.2592, pruned_loss=0.06545, over 4902.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2745, pruned_loss=0.07786, over 954847.40 frames. ], batch size: 36, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:41:47,162 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:41:53,522 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 05:41:57,727 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29907.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:42:07,774 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.638e+02 1.953e+02 2.201e+02 4.150e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 05:42:18,518 INFO [finetune.py:976] (0/7) Epoch 6, batch 1300, loss[loss=0.2251, simple_loss=0.2717, pruned_loss=0.08923, over 4851.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.272, pruned_loss=0.0774, over 954178.25 frames. ], batch size: 49, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:42:19,143 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:42:33,573 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2711, 1.7229, 1.8484, 1.9470, 1.7519, 1.8361, 1.9366, 1.8895], + device='cuda:0'), covar=tensor([0.8555, 1.0934, 0.9190, 0.9597, 1.1017, 0.8017, 1.2917, 0.8258], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0247, 0.0254, 0.0257, 0.0241, 0.0219, 0.0274, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:42:52,025 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0410, 1.8469, 1.8852, 0.8008, 2.0943, 2.3715, 1.8596, 1.8859], + device='cuda:0'), covar=tensor([0.1071, 0.0856, 0.0592, 0.0898, 0.0575, 0.0488, 0.0570, 0.0725], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0159, 0.0121, 0.0138, 0.0133, 0.0123, 0.0147, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.8732e-05, 1.1754e-04, 8.8032e-05, 1.0054e-04, 9.5691e-05, 9.1286e-05, + 1.0920e-04, 1.0829e-04], device='cuda:0') +2023-03-26 05:42:53,738 INFO [finetune.py:976] (0/7) Epoch 6, batch 1350, loss[loss=0.2718, simple_loss=0.3249, pruned_loss=0.1094, over 4819.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2721, pruned_loss=0.07741, over 954791.19 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:43:04,377 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0367, 2.1789, 2.0156, 1.3831, 2.2312, 2.3021, 2.1698, 1.8749], + device='cuda:0'), covar=tensor([0.0676, 0.0589, 0.0802, 0.1063, 0.0522, 0.0723, 0.0680, 0.0993], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0136, 0.0145, 0.0129, 0.0114, 0.0144, 0.0148, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:43:06,176 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-30000.pt +2023-03-26 05:43:22,448 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30018.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:43:25,361 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.660e+02 2.040e+02 2.486e+02 4.804e+02, threshold=4.081e+02, percent-clipped=2.0 +2023-03-26 05:43:35,490 INFO [finetune.py:976] (0/7) Epoch 6, batch 1400, loss[loss=0.2347, simple_loss=0.3063, pruned_loss=0.08158, over 4827.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2769, pruned_loss=0.07943, over 954618.45 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:43:39,910 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7560, 1.5509, 1.9750, 1.3963, 1.8573, 1.9737, 1.4664, 2.1089], + device='cuda:0'), covar=tensor([0.1325, 0.1990, 0.1487, 0.1994, 0.0898, 0.1366, 0.2707, 0.0887], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0205, 0.0199, 0.0195, 0.0184, 0.0220, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:43:52,220 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9351, 1.7608, 1.6036, 2.0282, 2.5270, 1.9499, 1.5635, 1.4840], + device='cuda:0'), covar=tensor([0.2224, 0.2310, 0.2038, 0.1699, 0.1861, 0.1256, 0.2681, 0.1948], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0209, 0.0202, 0.0185, 0.0237, 0.0176, 0.0214, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:44:14,312 INFO [finetune.py:976] (0/7) Epoch 6, batch 1450, loss[loss=0.2399, simple_loss=0.2947, pruned_loss=0.09257, over 4904.00 frames. ], tot_loss[loss=0.2176, simple_loss=0.2774, pruned_loss=0.07889, over 955226.96 frames. ], batch size: 32, lr: 3.91e-03, grad_scale: 64.0 +2023-03-26 05:44:25,282 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-03-26 05:44:27,594 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8165, 1.7522, 1.8029, 1.8175, 1.5557, 3.2954, 1.8266, 2.2885], + device='cuda:0'), covar=tensor([0.2855, 0.2038, 0.1698, 0.1924, 0.1453, 0.0233, 0.2464, 0.1016], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0118, 0.0099, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:44:28,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3142, 1.4913, 1.5118, 0.7567, 1.3972, 1.7360, 1.7648, 1.3665], + device='cuda:0'), covar=tensor([0.0886, 0.0562, 0.0496, 0.0595, 0.0417, 0.0494, 0.0297, 0.0626], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0159, 0.0122, 0.0138, 0.0133, 0.0124, 0.0148, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.8694e-05, 1.1746e-04, 8.8588e-05, 1.0069e-04, 9.5645e-05, 9.1792e-05, + 1.0933e-04, 1.0871e-04], device='cuda:0') +2023-03-26 05:44:58,652 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.825e+02 2.216e+02 2.642e+02 7.386e+02, threshold=4.431e+02, percent-clipped=2.0 +2023-03-26 05:44:59,439 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30125.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:45:16,738 INFO [finetune.py:976] (0/7) Epoch 6, batch 1500, loss[loss=0.2037, simple_loss=0.2689, pruned_loss=0.06923, over 4837.00 frames. ], tot_loss[loss=0.2206, simple_loss=0.2802, pruned_loss=0.08055, over 953077.13 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:45:38,714 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:45:59,647 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:01,366 INFO [finetune.py:976] (0/7) Epoch 6, batch 1550, loss[loss=0.171, simple_loss=0.2315, pruned_loss=0.05523, over 4764.00 frames. ], tot_loss[loss=0.2194, simple_loss=0.2794, pruned_loss=0.07973, over 954276.78 frames. ], batch size: 27, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:46:12,978 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30207.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:14,214 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:19,364 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9704, 1.7104, 1.7737, 1.9403, 1.1160, 4.4441, 1.6936, 2.3164], + device='cuda:0'), covar=tensor([0.3170, 0.2323, 0.1961, 0.2194, 0.1824, 0.0101, 0.2381, 0.1182], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0121, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:46:22,951 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.08 vs. limit=5.0 +2023-03-26 05:46:25,112 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.698e+02 2.038e+02 2.458e+02 3.858e+02, threshold=4.076e+02, percent-clipped=0.0 +2023-03-26 05:46:34,721 INFO [finetune.py:976] (0/7) Epoch 6, batch 1600, loss[loss=0.2068, simple_loss=0.2481, pruned_loss=0.08272, over 4199.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2776, pruned_loss=0.07938, over 955820.33 frames. ], batch size: 66, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:46:44,948 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:56,011 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:47:08,006 INFO [finetune.py:976] (0/7) Epoch 6, batch 1650, loss[loss=0.1493, simple_loss=0.2172, pruned_loss=0.04072, over 4774.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2744, pruned_loss=0.07814, over 956629.11 frames. ], batch size: 28, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:47:27,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30318.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:47:31,616 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.710e+02 1.999e+02 2.408e+02 3.997e+02, threshold=3.998e+02, percent-clipped=0.0 +2023-03-26 05:47:41,359 INFO [finetune.py:976] (0/7) Epoch 6, batch 1700, loss[loss=0.2257, simple_loss=0.2947, pruned_loss=0.07832, over 4863.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2715, pruned_loss=0.07699, over 955594.39 frames. ], batch size: 49, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:47:55,181 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:47:58,770 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30366.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:48:27,033 INFO [finetune.py:976] (0/7) Epoch 6, batch 1750, loss[loss=0.2627, simple_loss=0.3259, pruned_loss=0.09977, over 4822.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2738, pruned_loss=0.07786, over 955221.65 frames. ], batch size: 51, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:48:48,750 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:48:50,966 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.220e+01 1.748e+02 2.265e+02 2.778e+02 4.150e+02, threshold=4.530e+02, percent-clipped=1.0 +2023-03-26 05:49:00,662 INFO [finetune.py:976] (0/7) Epoch 6, batch 1800, loss[loss=0.2552, simple_loss=0.3053, pruned_loss=0.1025, over 4295.00 frames. ], tot_loss[loss=0.217, simple_loss=0.277, pruned_loss=0.07848, over 955913.11 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:49:12,732 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:49:20,053 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-03-26 05:49:28,995 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:49:33,833 INFO [finetune.py:976] (0/7) Epoch 6, batch 1850, loss[loss=0.243, simple_loss=0.3033, pruned_loss=0.0913, over 4906.00 frames. ], tot_loss[loss=0.2177, simple_loss=0.2775, pruned_loss=0.07899, over 954556.50 frames. ], batch size: 36, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:49:44,725 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:50:03,764 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.671e+02 2.010e+02 2.577e+02 4.739e+02, threshold=4.020e+02, percent-clipped=1.0 +2023-03-26 05:50:22,792 INFO [finetune.py:976] (0/7) Epoch 6, batch 1900, loss[loss=0.2409, simple_loss=0.2896, pruned_loss=0.09614, over 4262.00 frames. ], tot_loss[loss=0.2189, simple_loss=0.2793, pruned_loss=0.07924, over 954359.77 frames. ], batch size: 66, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:50:30,766 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 05:50:31,870 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7937, 1.3902, 0.8349, 1.6500, 2.2099, 1.2381, 1.6813, 1.6540], + device='cuda:0'), covar=tensor([0.1652, 0.2193, 0.2236, 0.1217, 0.1947, 0.2102, 0.1415, 0.2159], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0094, 0.0125, 0.0097, 0.0100, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:50:51,721 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:51:20,333 INFO [finetune.py:976] (0/7) Epoch 6, batch 1950, loss[loss=0.2468, simple_loss=0.318, pruned_loss=0.08782, over 4847.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2779, pruned_loss=0.07888, over 954205.56 frames. ], batch size: 47, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:51:28,873 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6405, 1.4605, 1.4760, 1.5486, 0.9926, 3.2283, 1.1231, 1.7362], + device='cuda:0'), covar=tensor([0.3342, 0.2425, 0.2128, 0.2343, 0.1968, 0.0183, 0.2808, 0.1357], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0121, 0.0117, 0.0099, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:51:56,504 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4753, 2.3246, 2.0287, 1.0026, 2.1685, 1.9313, 1.7407, 2.1447], + device='cuda:0'), covar=tensor([0.0912, 0.0751, 0.1363, 0.2036, 0.1309, 0.2087, 0.2043, 0.0928], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0202, 0.0203, 0.0191, 0.0218, 0.0211, 0.0221, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:51:58,194 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.556e+02 1.883e+02 2.215e+02 4.222e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-03-26 05:52:05,989 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4256, 2.1330, 1.6686, 0.7618, 1.8412, 1.9023, 1.7637, 1.8794], + device='cuda:0'), covar=tensor([0.0899, 0.0792, 0.1664, 0.2274, 0.1523, 0.2686, 0.2155, 0.0992], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0202, 0.0203, 0.0191, 0.0217, 0.0211, 0.0221, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:52:09,311 INFO [finetune.py:976] (0/7) Epoch 6, batch 2000, loss[loss=0.2308, simple_loss=0.2902, pruned_loss=0.08574, over 4888.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.2755, pruned_loss=0.07846, over 953905.43 frames. ], batch size: 35, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:52:14,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7493, 1.7306, 2.0257, 1.4859, 1.9314, 1.9428, 1.5822, 2.2293], + device='cuda:0'), covar=tensor([0.1803, 0.2185, 0.1599, 0.2218, 0.1110, 0.1677, 0.2936, 0.1007], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0203, 0.0197, 0.0193, 0.0181, 0.0219, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:52:19,610 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:52:56,172 INFO [finetune.py:976] (0/7) Epoch 6, batch 2050, loss[loss=0.1988, simple_loss=0.2591, pruned_loss=0.06922, over 4787.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2724, pruned_loss=0.07644, over 955804.12 frames. ], batch size: 25, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:52:57,411 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4151, 1.7816, 1.9150, 0.9334, 1.7142, 2.0249, 1.9988, 1.6773], + device='cuda:0'), covar=tensor([0.0941, 0.0561, 0.0445, 0.0566, 0.0491, 0.0522, 0.0323, 0.0596], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0158, 0.0121, 0.0138, 0.0133, 0.0124, 0.0147, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.8620e-05, 1.1719e-04, 8.8037e-05, 1.0089e-04, 9.5586e-05, 9.1968e-05, + 1.0888e-04, 1.0800e-04], device='cuda:0') +2023-03-26 05:53:14,329 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30716.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:53:14,376 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30716.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:53:20,053 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.687e+02 1.889e+02 2.318e+02 4.112e+02, threshold=3.779e+02, percent-clipped=3.0 +2023-03-26 05:53:31,807 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7201, 3.7501, 3.5308, 1.6414, 3.7835, 2.9108, 0.8747, 2.5318], + device='cuda:0'), covar=tensor([0.2659, 0.1862, 0.1628, 0.3529, 0.1008, 0.0919, 0.4610, 0.1608], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0171, 0.0163, 0.0128, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 05:53:40,714 INFO [finetune.py:976] (0/7) Epoch 6, batch 2100, loss[loss=0.2276, simple_loss=0.2813, pruned_loss=0.08695, over 4916.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2704, pruned_loss=0.07541, over 955423.27 frames. ], batch size: 36, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:01,032 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5438, 1.3666, 1.8395, 2.9118, 2.0276, 2.1210, 0.8966, 2.3192], + device='cuda:0'), covar=tensor([0.1748, 0.1609, 0.1324, 0.0643, 0.0845, 0.1301, 0.1959, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0103, 0.0143, 0.0130, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 05:54:08,291 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4344, 1.4052, 1.2204, 1.4180, 1.6932, 1.5859, 1.4859, 1.1994], + device='cuda:0'), covar=tensor([0.0301, 0.0278, 0.0559, 0.0298, 0.0221, 0.0514, 0.0252, 0.0415], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0111, 0.0137, 0.0117, 0.0103, 0.0100, 0.0090, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.7430e-05, 8.7330e-05, 1.0947e-04, 9.2146e-05, 8.1404e-05, 7.3905e-05, + 6.8940e-05, 8.4549e-05], device='cuda:0') +2023-03-26 05:54:13,633 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:54:18,840 INFO [finetune.py:976] (0/7) Epoch 6, batch 2150, loss[loss=0.1658, simple_loss=0.2462, pruned_loss=0.04274, over 4748.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2753, pruned_loss=0.07776, over 955390.02 frames. ], batch size: 27, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:42,042 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.708e+02 2.011e+02 2.625e+02 4.679e+02, threshold=4.022e+02, percent-clipped=7.0 +2023-03-26 05:54:45,608 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30829.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:54:52,232 INFO [finetune.py:976] (0/7) Epoch 6, batch 2200, loss[loss=0.2218, simple_loss=0.2839, pruned_loss=0.07987, over 4918.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.277, pruned_loss=0.07806, over 956233.32 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:55,388 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-26 05:55:16,382 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:55:26,053 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4455, 1.7478, 0.7490, 2.3555, 2.7729, 2.0105, 2.1208, 2.3086], + device='cuda:0'), covar=tensor([0.1360, 0.1943, 0.2446, 0.1106, 0.1660, 0.1872, 0.1322, 0.1939], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0094, 0.0125, 0.0097, 0.0100, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 05:55:37,476 INFO [finetune.py:976] (0/7) Epoch 6, batch 2250, loss[loss=0.214, simple_loss=0.2744, pruned_loss=0.07678, over 4773.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2773, pruned_loss=0.07739, over 958187.35 frames. ], batch size: 27, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:55:53,168 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:10,383 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30913.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:22,247 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.611e+02 1.958e+02 2.302e+02 5.232e+02, threshold=3.915e+02, percent-clipped=2.0 +2023-03-26 05:56:22,981 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:34,301 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:42,903 INFO [finetune.py:976] (0/7) Epoch 6, batch 2300, loss[loss=0.2066, simple_loss=0.2784, pruned_loss=0.06736, over 4824.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2758, pruned_loss=0.07602, over 957590.81 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:57:16,196 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:57:28,849 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1453, 1.4306, 1.5102, 0.6995, 1.2779, 1.6419, 1.6482, 1.4113], + device='cuda:0'), covar=tensor([0.0946, 0.0524, 0.0458, 0.0503, 0.0496, 0.0514, 0.0356, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0157, 0.0119, 0.0137, 0.0132, 0.0123, 0.0146, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.7596e-05, 1.1587e-04, 8.6636e-05, 9.9623e-05, 9.4956e-05, 9.0943e-05, + 1.0826e-04, 1.0673e-04], device='cuda:0') +2023-03-26 05:57:46,780 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:57:49,010 INFO [finetune.py:976] (0/7) Epoch 6, batch 2350, loss[loss=0.2478, simple_loss=0.2946, pruned_loss=0.1005, over 4814.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2744, pruned_loss=0.07633, over 955564.32 frames. ], batch size: 40, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:57:57,470 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:58:01,116 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1269, 2.1889, 2.0255, 1.4632, 2.3642, 2.2834, 2.3252, 2.0012], + device='cuda:0'), covar=tensor([0.0593, 0.0603, 0.0723, 0.0874, 0.0417, 0.0714, 0.0562, 0.0826], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0135, 0.0144, 0.0128, 0.0113, 0.0144, 0.0146, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:58:07,196 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6878, 4.4086, 4.1664, 2.3847, 4.4041, 3.4668, 0.8749, 3.1568], + device='cuda:0'), covar=tensor([0.2715, 0.1503, 0.1433, 0.2829, 0.0775, 0.0791, 0.4408, 0.1284], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0170, 0.0163, 0.0128, 0.0156, 0.0123, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 05:58:19,344 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31011.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:58:19,371 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:58:28,231 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:58:33,048 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.621e+02 1.904e+02 2.250e+02 3.149e+02, threshold=3.808e+02, percent-clipped=0.0 +2023-03-26 05:58:41,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0403, 1.8084, 1.6337, 1.7459, 1.7567, 1.7733, 1.7426, 2.5242], + device='cuda:0'), covar=tensor([0.5296, 0.6310, 0.4307, 0.6043, 0.5516, 0.3185, 0.5641, 0.2267], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0220, 0.0285, 0.0241, 0.0203, 0.0247, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:58:53,274 INFO [finetune.py:976] (0/7) Epoch 6, batch 2400, loss[loss=0.2235, simple_loss=0.2527, pruned_loss=0.09711, over 3765.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2712, pruned_loss=0.07536, over 955749.75 frames. ], batch size: 15, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:59:16,400 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0455, 1.8635, 1.7349, 1.7559, 2.1608, 1.7976, 2.2632, 2.0469], + device='cuda:0'), covar=tensor([0.1620, 0.2876, 0.3671, 0.3106, 0.2836, 0.1893, 0.3330, 0.2163], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0192, 0.0236, 0.0254, 0.0231, 0.0190, 0.0212, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:59:16,908 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:59:18,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1461, 2.0382, 1.9090, 2.0939, 2.8062, 2.1473, 2.0300, 1.5958], + device='cuda:0'), covar=tensor([0.2514, 0.2338, 0.2051, 0.2100, 0.2002, 0.1213, 0.2481, 0.2173], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0209, 0.0203, 0.0186, 0.0237, 0.0176, 0.0214, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:59:24,028 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31072.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:59:34,788 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8080, 1.6601, 1.5636, 1.7011, 1.1136, 3.7778, 1.5759, 2.1236], + device='cuda:0'), covar=tensor([0.3150, 0.2392, 0.2073, 0.2339, 0.1880, 0.0162, 0.2395, 0.1233], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0117, 0.0099, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 05:59:37,283 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9455, 1.8787, 1.6881, 2.0216, 2.4836, 2.0230, 1.6111, 1.5591], + device='cuda:0'), covar=tensor([0.2124, 0.1986, 0.1795, 0.1672, 0.1634, 0.1111, 0.2611, 0.1882], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0209, 0.0204, 0.0186, 0.0237, 0.0176, 0.0214, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 05:59:44,204 INFO [finetune.py:976] (0/7) Epoch 6, batch 2450, loss[loss=0.1922, simple_loss=0.2579, pruned_loss=0.06329, over 4903.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.268, pruned_loss=0.07427, over 954472.71 frames. ], batch size: 43, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 06:00:21,949 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.316e+02 1.762e+02 2.221e+02 2.552e+02 5.044e+02, threshold=4.442e+02, percent-clipped=4.0 +2023-03-26 06:00:30,922 INFO [finetune.py:976] (0/7) Epoch 6, batch 2500, loss[loss=0.2213, simple_loss=0.285, pruned_loss=0.0788, over 4848.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2694, pruned_loss=0.07502, over 956148.65 frames. ], batch size: 49, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:00:31,112 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-03-26 06:00:38,176 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-26 06:01:06,650 INFO [finetune.py:976] (0/7) Epoch 6, batch 2550, loss[loss=0.2528, simple_loss=0.3117, pruned_loss=0.09695, over 4757.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2738, pruned_loss=0.07728, over 954765.51 frames. ], batch size: 54, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:01:50,622 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.680e+02 2.034e+02 2.315e+02 3.655e+02, threshold=4.067e+02, percent-clipped=0.0 +2023-03-26 06:02:04,702 INFO [finetune.py:976] (0/7) Epoch 6, batch 2600, loss[loss=0.2523, simple_loss=0.3025, pruned_loss=0.1011, over 4778.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2757, pruned_loss=0.07755, over 956166.19 frames. ], batch size: 29, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:02:33,424 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:02:59,295 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:09,261 INFO [finetune.py:976] (0/7) Epoch 6, batch 2650, loss[loss=0.2294, simple_loss=0.2839, pruned_loss=0.08748, over 4197.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2793, pruned_loss=0.07954, over 954680.11 frames. ], batch size: 65, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:03:09,907 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:38,130 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:47,924 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.761e+02 2.106e+02 2.462e+02 3.966e+02, threshold=4.213e+02, percent-clipped=0.0 +2023-03-26 06:03:59,094 INFO [finetune.py:976] (0/7) Epoch 6, batch 2700, loss[loss=0.2419, simple_loss=0.2939, pruned_loss=0.09491, over 4921.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.278, pruned_loss=0.07876, over 954689.17 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:04:23,326 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31359.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:04:33,876 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31367.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 06:04:43,293 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2356, 1.3627, 1.5590, 1.0859, 1.2984, 1.4477, 1.2987, 1.5765], + device='cuda:0'), covar=tensor([0.1311, 0.2241, 0.1356, 0.1616, 0.1021, 0.1300, 0.2974, 0.0902], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0207, 0.0199, 0.0196, 0.0185, 0.0222, 0.0219, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:05:04,058 INFO [finetune.py:976] (0/7) Epoch 6, batch 2750, loss[loss=0.1813, simple_loss=0.2382, pruned_loss=0.06219, over 4790.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2748, pruned_loss=0.07761, over 952213.16 frames. ], batch size: 25, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:05:28,059 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31409.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:05:46,888 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.636e+02 1.989e+02 2.265e+02 3.886e+02, threshold=3.977e+02, percent-clipped=0.0 +2023-03-26 06:05:56,330 INFO [finetune.py:976] (0/7) Epoch 6, batch 2800, loss[loss=0.1823, simple_loss=0.2486, pruned_loss=0.058, over 4920.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2721, pruned_loss=0.07666, over 950204.95 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:06:16,513 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:06:42,532 INFO [finetune.py:976] (0/7) Epoch 6, batch 2850, loss[loss=0.3056, simple_loss=0.3404, pruned_loss=0.1354, over 4103.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2723, pruned_loss=0.0774, over 952395.97 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:06:43,220 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2493, 2.8783, 2.7791, 1.1649, 2.9637, 2.2886, 0.6893, 1.8804], + device='cuda:0'), covar=tensor([0.2778, 0.2457, 0.1976, 0.3874, 0.1394, 0.1152, 0.4418, 0.1778], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0171, 0.0163, 0.0128, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 06:06:43,854 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31491.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:07:26,132 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.576e+02 1.974e+02 2.520e+02 5.037e+02, threshold=3.948e+02, percent-clipped=2.0 +2023-03-26 06:07:36,765 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 06:07:42,539 INFO [finetune.py:976] (0/7) Epoch 6, batch 2900, loss[loss=0.2796, simple_loss=0.3409, pruned_loss=0.1092, over 4761.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2745, pruned_loss=0.07787, over 951105.75 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:07:55,738 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31552.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:08:03,221 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8941, 1.8645, 1.6705, 2.0472, 2.4114, 1.9193, 1.5474, 1.5490], + device='cuda:0'), covar=tensor([0.2498, 0.2333, 0.2130, 0.1886, 0.2039, 0.1326, 0.2968, 0.2240], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0209, 0.0202, 0.0186, 0.0236, 0.0176, 0.0212, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:08:04,331 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31559.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:05,132 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 06:08:19,175 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31581.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:27,344 INFO [finetune.py:976] (0/7) Epoch 6, batch 2950, loss[loss=0.2371, simple_loss=0.3012, pruned_loss=0.08651, over 4803.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2768, pruned_loss=0.07815, over 953304.88 frames. ], batch size: 41, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:08:33,411 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31590.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:47,761 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:59,951 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.780e+02 2.105e+02 2.565e+02 4.082e+02, threshold=4.210e+02, percent-clipped=1.0 +2023-03-26 06:09:02,945 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:09:09,343 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31638.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:09:09,892 INFO [finetune.py:976] (0/7) Epoch 6, batch 3000, loss[loss=0.1906, simple_loss=0.2448, pruned_loss=0.06821, over 4359.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2794, pruned_loss=0.07962, over 952436.89 frames. ], batch size: 19, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:09:09,893 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 06:09:14,114 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5210, 1.2649, 1.3118, 1.3316, 1.6461, 1.5960, 1.4750, 1.2300], + device='cuda:0'), covar=tensor([0.0267, 0.0333, 0.0505, 0.0311, 0.0258, 0.0400, 0.0299, 0.0425], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0111, 0.0137, 0.0116, 0.0103, 0.0100, 0.0090, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.7290e-05, 8.7190e-05, 1.0945e-04, 9.1587e-05, 8.1257e-05, 7.4104e-05, + 6.8327e-05, 8.4034e-05], device='cuda:0') +2023-03-26 06:09:23,490 INFO [finetune.py:1010] (0/7) Epoch 6, validation: loss=0.1625, simple_loss=0.2344, pruned_loss=0.04534, over 2265189.00 frames. +2023-03-26 06:09:23,490 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 06:09:55,814 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31667.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:10:19,288 INFO [finetune.py:976] (0/7) Epoch 6, batch 3050, loss[loss=0.213, simple_loss=0.2828, pruned_loss=0.07158, over 4860.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2796, pruned_loss=0.07922, over 953917.52 frames. ], batch size: 31, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:10:28,746 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6446, 1.7084, 1.5161, 1.8611, 2.1750, 1.7799, 1.4370, 1.3801], + device='cuda:0'), covar=tensor([0.2691, 0.2357, 0.2244, 0.2015, 0.2167, 0.1334, 0.2979, 0.2254], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0209, 0.0202, 0.0186, 0.0236, 0.0176, 0.0213, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:10:29,946 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31704.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:10:33,649 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9667, 1.8434, 1.4666, 1.8438, 1.9782, 1.5950, 2.2441, 1.9503], + device='cuda:0'), covar=tensor([0.1749, 0.3090, 0.4073, 0.3527, 0.3067, 0.1995, 0.4041, 0.2429], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0192, 0.0237, 0.0254, 0.0232, 0.0191, 0.0211, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:10:36,650 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31715.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:10:43,061 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.708e+02 2.072e+02 2.420e+02 4.871e+02, threshold=4.144e+02, percent-clipped=1.0 +2023-03-26 06:10:59,685 INFO [finetune.py:976] (0/7) Epoch 6, batch 3100, loss[loss=0.1915, simple_loss=0.2461, pruned_loss=0.06847, over 4826.00 frames. ], tot_loss[loss=0.2162, simple_loss=0.2766, pruned_loss=0.0779, over 955564.32 frames. ], batch size: 30, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:11:20,229 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:20,277 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:35,297 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:36,281 INFO [finetune.py:976] (0/7) Epoch 6, batch 3150, loss[loss=0.1964, simple_loss=0.2605, pruned_loss=0.06613, over 4797.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2725, pruned_loss=0.07634, over 954718.41 frames. ], batch size: 51, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:12:00,897 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:12:01,992 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.748e+01 1.682e+02 2.020e+02 2.535e+02 4.605e+02, threshold=4.040e+02, percent-clipped=1.0 +2023-03-26 06:12:15,951 INFO [finetune.py:976] (0/7) Epoch 6, batch 3200, loss[loss=0.1888, simple_loss=0.2392, pruned_loss=0.06922, over 4793.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2689, pruned_loss=0.07481, over 955205.50 frames. ], batch size: 29, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:12:25,646 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31847.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 06:12:31,912 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:06,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:11,179 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:14,082 INFO [finetune.py:976] (0/7) Epoch 6, batch 3250, loss[loss=0.2481, simple_loss=0.3022, pruned_loss=0.09704, over 4868.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2693, pruned_loss=0.07474, over 955760.24 frames. ], batch size: 34, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:13:51,965 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:01,794 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.695e+02 2.048e+02 2.376e+02 4.231e+02, threshold=4.096e+02, percent-clipped=1.0 +2023-03-26 06:14:17,633 INFO [finetune.py:976] (0/7) Epoch 6, batch 3300, loss[loss=0.1762, simple_loss=0.24, pruned_loss=0.05621, over 4692.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2742, pruned_loss=0.07672, over 956634.52 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:14:17,769 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:29,121 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:43,722 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 06:14:46,579 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:54,838 INFO [finetune.py:976] (0/7) Epoch 6, batch 3350, loss[loss=0.224, simple_loss=0.2982, pruned_loss=0.07492, over 4922.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2758, pruned_loss=0.07648, over 957362.86 frames. ], batch size: 42, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:15:02,168 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-32000.pt +2023-03-26 06:15:11,469 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=32011.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:15:22,687 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.743e+02 2.097e+02 2.540e+02 4.089e+02, threshold=4.194e+02, percent-clipped=0.0 +2023-03-26 06:15:31,463 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1123, 1.0402, 1.8122, 1.6982, 1.6266, 1.6181, 1.5952, 1.7150], + device='cuda:0'), covar=tensor([0.5969, 0.7784, 0.7329, 0.6878, 0.8610, 0.6839, 0.8986, 0.7092], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0246, 0.0255, 0.0257, 0.0243, 0.0220, 0.0274, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:15:41,803 INFO [finetune.py:976] (0/7) Epoch 6, batch 3400, loss[loss=0.1953, simple_loss=0.2661, pruned_loss=0.06224, over 4887.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.2775, pruned_loss=0.07787, over 954544.41 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:16:04,436 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32060.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:16:13,846 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:16:41,402 INFO [finetune.py:976] (0/7) Epoch 6, batch 3450, loss[loss=0.2375, simple_loss=0.3041, pruned_loss=0.08548, over 4896.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2761, pruned_loss=0.07665, over 955569.49 frames. ], batch size: 37, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:16:41,708 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 06:17:05,358 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7819, 1.5036, 1.8463, 1.1063, 1.9139, 1.8617, 1.8119, 1.2843], + device='cuda:0'), covar=tensor([0.0692, 0.0896, 0.0694, 0.1050, 0.0601, 0.0717, 0.0721, 0.1661], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0136, 0.0146, 0.0130, 0.0114, 0.0146, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:17:07,732 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:17:16,933 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.573e+02 1.926e+02 2.516e+02 4.351e+02, threshold=3.853e+02, percent-clipped=2.0 +2023-03-26 06:17:18,880 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6399, 2.4038, 1.9399, 1.0632, 2.1826, 1.9771, 1.8408, 2.0889], + device='cuda:0'), covar=tensor([0.0990, 0.0873, 0.1763, 0.2349, 0.1501, 0.2585, 0.2331, 0.1170], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0203, 0.0201, 0.0190, 0.0217, 0.0209, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:17:25,471 INFO [finetune.py:976] (0/7) Epoch 6, batch 3500, loss[loss=0.2231, simple_loss=0.2856, pruned_loss=0.08028, over 4830.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.2728, pruned_loss=0.07498, over 954941.73 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:17:29,073 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:17:30,891 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32147.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 06:18:09,957 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:18:15,990 INFO [finetune.py:976] (0/7) Epoch 6, batch 3550, loss[loss=0.2041, simple_loss=0.2482, pruned_loss=0.07998, over 4301.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2708, pruned_loss=0.07493, over 955291.26 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:18:19,693 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32195.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:18:22,034 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1420, 1.5629, 0.6679, 1.9798, 2.3294, 1.7670, 1.8176, 1.8354], + device='cuda:0'), covar=tensor([0.1543, 0.2087, 0.2600, 0.1291, 0.2066, 0.1957, 0.1478, 0.2249], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0092, 0.0124, 0.0096, 0.0100, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 06:18:35,988 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0362, 1.6612, 2.1554, 2.1682, 1.8367, 4.3410, 1.5990, 2.0022], + device='cuda:0'), covar=tensor([0.0996, 0.2278, 0.1142, 0.1096, 0.1731, 0.0272, 0.1785, 0.2050], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 06:18:40,463 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.617e+02 1.889e+02 2.318e+02 4.823e+02, threshold=3.777e+02, percent-clipped=2.0 +2023-03-26 06:18:41,218 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1719, 2.0211, 1.6688, 2.1828, 2.2262, 1.8648, 2.5464, 2.1459], + device='cuda:0'), covar=tensor([0.1696, 0.3079, 0.3923, 0.3115, 0.2873, 0.2000, 0.3497, 0.2181], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0192, 0.0235, 0.0253, 0.0231, 0.0191, 0.0212, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:18:52,018 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:19:00,023 INFO [finetune.py:976] (0/7) Epoch 6, batch 3600, loss[loss=0.1619, simple_loss=0.2279, pruned_loss=0.04791, over 4766.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.267, pruned_loss=0.07326, over 954248.63 frames. ], batch size: 28, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:19:36,014 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32271.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:19:57,774 INFO [finetune.py:976] (0/7) Epoch 6, batch 3650, loss[loss=0.2271, simple_loss=0.2863, pruned_loss=0.08392, over 4926.00 frames. ], tot_loss[loss=0.2084, simple_loss=0.2685, pruned_loss=0.07417, over 954128.70 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:20:14,262 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:20:18,591 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5415, 1.4210, 1.8931, 2.8055, 1.9694, 2.2610, 0.9783, 2.2648], + device='cuda:0'), covar=tensor([0.1640, 0.1537, 0.1233, 0.0664, 0.0832, 0.1172, 0.1810, 0.0667], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0118, 0.0136, 0.0167, 0.0102, 0.0142, 0.0130, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 06:20:26,743 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.776e+02 2.129e+02 2.587e+02 5.126e+02, threshold=4.259e+02, percent-clipped=5.0 +2023-03-26 06:20:43,050 INFO [finetune.py:976] (0/7) Epoch 6, batch 3700, loss[loss=0.2272, simple_loss=0.2964, pruned_loss=0.07899, over 4910.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2736, pruned_loss=0.07636, over 952649.04 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:20:53,484 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0072, 1.4624, 1.6167, 1.6869, 1.5165, 1.5665, 1.6673, 1.6934], + device='cuda:0'), covar=tensor([0.6699, 0.7643, 0.6777, 0.7532, 0.8679, 0.6889, 0.9832, 0.6335], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0245, 0.0255, 0.0256, 0.0242, 0.0220, 0.0274, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:20:56,516 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:20:57,803 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9276, 1.9344, 1.6055, 1.9749, 1.8599, 1.8095, 1.7784, 2.5900], + device='cuda:0'), covar=tensor([0.5900, 0.7665, 0.4995, 0.7245, 0.6579, 0.3635, 0.6776, 0.2329], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0283, 0.0240, 0.0203, 0.0245, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:20:59,645 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7571, 1.5529, 1.4436, 1.5110, 1.8849, 1.8801, 1.7456, 1.4471], + device='cuda:0'), covar=tensor([0.0248, 0.0325, 0.0497, 0.0300, 0.0204, 0.0368, 0.0231, 0.0402], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0111, 0.0136, 0.0115, 0.0103, 0.0099, 0.0090, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.7265e-05, 8.7208e-05, 1.0925e-04, 9.0943e-05, 8.1001e-05, 7.3202e-05, + 6.8308e-05, 8.3592e-05], device='cuda:0') +2023-03-26 06:21:16,590 INFO [finetune.py:976] (0/7) Epoch 6, batch 3750, loss[loss=0.1997, simple_loss=0.2728, pruned_loss=0.06329, over 4899.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2758, pruned_loss=0.07725, over 954644.56 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:21:37,052 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:21:48,323 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.626e+02 1.929e+02 2.294e+02 3.909e+02, threshold=3.857e+02, percent-clipped=0.0 +2023-03-26 06:21:58,657 INFO [finetune.py:976] (0/7) Epoch 6, batch 3800, loss[loss=0.1968, simple_loss=0.2703, pruned_loss=0.06167, over 4861.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2761, pruned_loss=0.07657, over 956110.27 frames. ], batch size: 31, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:22:01,776 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:06,173 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 06:22:08,646 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 06:22:10,104 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7668, 3.3881, 3.3461, 1.7691, 3.4946, 2.6070, 1.1087, 2.5755], + device='cuda:0'), covar=tensor([0.2754, 0.1768, 0.1354, 0.3030, 0.1058, 0.1016, 0.3622, 0.1290], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0172, 0.0163, 0.0128, 0.0156, 0.0123, 0.0145, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 06:22:24,417 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:31,248 INFO [finetune.py:976] (0/7) Epoch 6, batch 3850, loss[loss=0.1899, simple_loss=0.2511, pruned_loss=0.06441, over 4824.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.2745, pruned_loss=0.07633, over 954633.62 frames. ], batch size: 33, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:22:33,633 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:36,262 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 06:22:53,999 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.617e+01 1.620e+02 2.056e+02 2.694e+02 5.558e+02, threshold=4.113e+02, percent-clipped=4.0 +2023-03-26 06:22:54,225 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 06:22:55,289 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:57,081 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4103, 2.2484, 1.7753, 0.9848, 2.0435, 1.9276, 1.6989, 1.9685], + device='cuda:0'), covar=tensor([0.0880, 0.0737, 0.1467, 0.2022, 0.1422, 0.2122, 0.2105, 0.1030], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0200, 0.0199, 0.0188, 0.0214, 0.0207, 0.0220, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:23:00,098 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32534.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:04,462 INFO [finetune.py:976] (0/7) Epoch 6, batch 3900, loss[loss=0.1466, simple_loss=0.2236, pruned_loss=0.03481, over 4784.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2725, pruned_loss=0.07582, over 955988.34 frames. ], batch size: 26, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:23:24,748 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:31,332 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32582.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:33,100 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6023, 1.5195, 1.5229, 1.5857, 1.2344, 2.8290, 1.2078, 1.6874], + device='cuda:0'), covar=tensor([0.3222, 0.2425, 0.1936, 0.2218, 0.1793, 0.0306, 0.2515, 0.1205], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0122, 0.0118, 0.0099, 0.0102, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 06:23:36,048 INFO [finetune.py:976] (0/7) Epoch 6, batch 3950, loss[loss=0.1714, simple_loss=0.24, pruned_loss=0.05143, over 4780.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2692, pruned_loss=0.07458, over 954973.04 frames. ], batch size: 27, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:23:53,884 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:24:07,653 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32619.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:24:11,251 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.672e+02 2.141e+02 2.509e+02 4.478e+02, threshold=4.281e+02, percent-clipped=2.0 +2023-03-26 06:24:20,840 INFO [finetune.py:976] (0/7) Epoch 6, batch 4000, loss[loss=0.2043, simple_loss=0.2777, pruned_loss=0.06545, over 4905.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2695, pruned_loss=0.07519, over 954583.84 frames. ], batch size: 35, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:24:26,497 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 06:24:30,951 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:24:31,666 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0681, 1.9888, 1.5818, 2.0658, 1.8544, 1.7466, 1.8515, 2.8354], + device='cuda:0'), covar=tensor([0.6488, 0.7958, 0.5239, 0.6952, 0.6913, 0.3788, 0.7141, 0.2191], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0283, 0.0239, 0.0203, 0.0244, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:25:04,478 INFO [finetune.py:976] (0/7) Epoch 6, batch 4050, loss[loss=0.2701, simple_loss=0.3267, pruned_loss=0.1067, over 4823.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2743, pruned_loss=0.0774, over 954462.24 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:25:28,436 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.831e+02 2.133e+02 2.637e+02 5.226e+02, threshold=4.267e+02, percent-clipped=1.0 +2023-03-26 06:25:42,228 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5987, 1.1807, 0.8908, 1.5182, 1.9035, 1.3699, 1.4503, 1.6622], + device='cuda:0'), covar=tensor([0.1915, 0.2986, 0.2621, 0.1564, 0.2551, 0.2524, 0.1977, 0.2454], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0092, 0.0123, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 06:25:42,718 INFO [finetune.py:976] (0/7) Epoch 6, batch 4100, loss[loss=0.2121, simple_loss=0.2702, pruned_loss=0.07703, over 4906.00 frames. ], tot_loss[loss=0.2189, simple_loss=0.2784, pruned_loss=0.07968, over 954606.91 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:26:28,496 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 06:26:34,198 INFO [finetune.py:976] (0/7) Epoch 6, batch 4150, loss[loss=0.1722, simple_loss=0.2376, pruned_loss=0.0534, over 4790.00 frames. ], tot_loss[loss=0.2199, simple_loss=0.2794, pruned_loss=0.08015, over 954291.87 frames. ], batch size: 29, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:27:18,098 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.203e+02 1.784e+02 2.149e+02 2.523e+02 4.029e+02, threshold=4.299e+02, percent-clipped=0.0 +2023-03-26 06:27:37,294 INFO [finetune.py:976] (0/7) Epoch 6, batch 4200, loss[loss=0.2531, simple_loss=0.3046, pruned_loss=0.1008, over 4883.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2785, pruned_loss=0.079, over 954234.63 frames. ], batch size: 32, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:28:01,957 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4504, 2.1649, 1.7919, 0.9093, 2.0238, 1.9284, 1.6696, 1.9690], + device='cuda:0'), covar=tensor([0.0789, 0.0847, 0.1460, 0.1901, 0.1322, 0.2055, 0.2179, 0.0946], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0201, 0.0200, 0.0188, 0.0215, 0.0207, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:28:34,720 INFO [finetune.py:976] (0/7) Epoch 6, batch 4250, loss[loss=0.2096, simple_loss=0.2622, pruned_loss=0.07849, over 4797.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2755, pruned_loss=0.07779, over 954812.24 frames. ], batch size: 51, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:28:45,363 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8085, 1.6634, 1.5913, 1.7577, 1.4885, 4.1524, 1.8010, 2.4855], + device='cuda:0'), covar=tensor([0.3879, 0.2853, 0.2318, 0.2583, 0.1648, 0.0184, 0.2509, 0.1141], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0115, 0.0118, 0.0123, 0.0118, 0.0099, 0.0102, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 06:29:25,480 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.595e+02 1.920e+02 2.303e+02 3.727e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 06:29:44,594 INFO [finetune.py:976] (0/7) Epoch 6, batch 4300, loss[loss=0.176, simple_loss=0.2561, pruned_loss=0.04799, over 4823.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2726, pruned_loss=0.0764, over 955239.36 frames. ], batch size: 39, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:30:06,911 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9296, 2.6416, 2.1206, 1.2307, 2.2629, 2.3060, 2.0279, 2.2210], + device='cuda:0'), covar=tensor([0.0594, 0.0660, 0.1278, 0.1883, 0.1187, 0.1872, 0.1818, 0.0874], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0200, 0.0200, 0.0188, 0.0215, 0.0207, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:30:17,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=32966.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:30:46,682 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 06:30:47,043 INFO [finetune.py:976] (0/7) Epoch 6, batch 4350, loss[loss=0.1981, simple_loss=0.2594, pruned_loss=0.06836, over 4738.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2694, pruned_loss=0.07491, over 953381.49 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:31:32,019 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.730e+02 2.041e+02 2.589e+02 3.941e+02, threshold=4.082e+02, percent-clipped=1.0 +2023-03-26 06:31:33,361 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:31:50,663 INFO [finetune.py:976] (0/7) Epoch 6, batch 4400, loss[loss=0.2529, simple_loss=0.3113, pruned_loss=0.09724, over 4911.00 frames. ], tot_loss[loss=0.2115, simple_loss=0.2706, pruned_loss=0.07617, over 952278.83 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:32:31,736 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:32:33,601 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1641, 1.7108, 1.2984, 0.5501, 1.5725, 1.7140, 1.3532, 1.6041], + device='cuda:0'), covar=tensor([0.0776, 0.1089, 0.1895, 0.2496, 0.1646, 0.2142, 0.2795, 0.1194], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0199, 0.0199, 0.0188, 0.0214, 0.0206, 0.0218, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:32:42,355 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 06:32:54,588 INFO [finetune.py:976] (0/7) Epoch 6, batch 4450, loss[loss=0.264, simple_loss=0.3277, pruned_loss=0.1002, over 4809.00 frames. ], tot_loss[loss=0.2143, simple_loss=0.274, pruned_loss=0.07728, over 952270.55 frames. ], batch size: 41, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:32:55,878 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3526, 1.2759, 1.6254, 2.4533, 1.6957, 2.0220, 0.8555, 2.0561], + device='cuda:0'), covar=tensor([0.1892, 0.1557, 0.1171, 0.0801, 0.0913, 0.1315, 0.1703, 0.0765], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0118, 0.0135, 0.0165, 0.0102, 0.0141, 0.0128, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 06:33:39,238 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.714e+02 2.125e+02 2.690e+02 4.211e+02, threshold=4.250e+02, percent-clipped=1.0 +2023-03-26 06:33:47,518 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:33:58,508 INFO [finetune.py:976] (0/7) Epoch 6, batch 4500, loss[loss=0.2022, simple_loss=0.2636, pruned_loss=0.07045, over 4826.00 frames. ], tot_loss[loss=0.2158, simple_loss=0.2761, pruned_loss=0.07771, over 952569.82 frames. ], batch size: 39, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:34:19,020 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33155.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:34:52,862 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 06:35:01,062 INFO [finetune.py:976] (0/7) Epoch 6, batch 4550, loss[loss=0.2403, simple_loss=0.2942, pruned_loss=0.09321, over 4912.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2782, pruned_loss=0.0788, over 951905.64 frames. ], batch size: 37, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:35:03,049 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9917, 1.8821, 2.3421, 1.5704, 2.0594, 2.1413, 1.7560, 2.3844], + device='cuda:0'), covar=tensor([0.1618, 0.1994, 0.1510, 0.2088, 0.1010, 0.1827, 0.2570, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0204, 0.0198, 0.0195, 0.0182, 0.0219, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:35:10,303 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7368, 1.6955, 2.1243, 1.3772, 1.9064, 1.9164, 1.5564, 2.1680], + device='cuda:0'), covar=tensor([0.1531, 0.2019, 0.1495, 0.2227, 0.0925, 0.1734, 0.2698, 0.0933], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0204, 0.0198, 0.0195, 0.0182, 0.0219, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:35:33,169 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33216.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:35:44,645 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.755e+02 2.084e+02 2.335e+02 4.621e+02, threshold=4.168e+02, percent-clipped=2.0 +2023-03-26 06:36:05,046 INFO [finetune.py:976] (0/7) Epoch 6, batch 4600, loss[loss=0.2118, simple_loss=0.2722, pruned_loss=0.07566, over 4809.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.2779, pruned_loss=0.0783, over 952629.96 frames. ], batch size: 40, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:36:55,979 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8884, 1.2547, 1.7464, 1.6857, 1.5213, 1.5491, 1.5574, 1.6097], + device='cuda:0'), covar=tensor([0.4666, 0.6275, 0.5316, 0.5530, 0.6837, 0.4822, 0.6913, 0.4866], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0244, 0.0253, 0.0256, 0.0241, 0.0219, 0.0272, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:37:07,549 INFO [finetune.py:976] (0/7) Epoch 6, batch 4650, loss[loss=0.2755, simple_loss=0.3144, pruned_loss=0.1183, over 4828.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2745, pruned_loss=0.07734, over 952371.54 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:37:48,443 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:37:50,591 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.572e+02 1.961e+02 2.434e+02 3.752e+02, threshold=3.921e+02, percent-clipped=0.0 +2023-03-26 06:38:10,951 INFO [finetune.py:976] (0/7) Epoch 6, batch 4700, loss[loss=0.182, simple_loss=0.2363, pruned_loss=0.06387, over 4722.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2711, pruned_loss=0.07569, over 952682.80 frames. ], batch size: 59, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:39:20,540 INFO [finetune.py:976] (0/7) Epoch 6, batch 4750, loss[loss=0.1508, simple_loss=0.2175, pruned_loss=0.0421, over 4896.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2692, pruned_loss=0.07469, over 954793.76 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:40:04,157 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.633e+02 1.917e+02 2.350e+02 3.562e+02, threshold=3.835e+02, percent-clipped=0.0 +2023-03-26 06:40:04,243 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:40:04,898 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2963, 2.2076, 1.8213, 2.5393, 2.3166, 1.9035, 2.9291, 2.3058], + device='cuda:0'), covar=tensor([0.1705, 0.3202, 0.3795, 0.3197, 0.3071, 0.2034, 0.3662, 0.2377], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0191, 0.0236, 0.0255, 0.0232, 0.0192, 0.0212, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:40:24,234 INFO [finetune.py:976] (0/7) Epoch 6, batch 4800, loss[loss=0.2912, simple_loss=0.3437, pruned_loss=0.1193, over 4899.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.2712, pruned_loss=0.07585, over 954618.97 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:41:07,493 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:41:28,610 INFO [finetune.py:976] (0/7) Epoch 6, batch 4850, loss[loss=0.193, simple_loss=0.2487, pruned_loss=0.06869, over 4179.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2742, pruned_loss=0.07657, over 953760.89 frames. ], batch size: 18, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:41:32,574 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4428, 3.7873, 4.0063, 4.2646, 4.1942, 3.9044, 4.5409, 1.4685], + device='cuda:0'), covar=tensor([0.0840, 0.0923, 0.0810, 0.1092, 0.1306, 0.1500, 0.0632, 0.5413], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0241, 0.0274, 0.0294, 0.0334, 0.0283, 0.0302, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:41:41,223 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 06:41:51,257 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5056, 1.5868, 1.1979, 1.3936, 1.7458, 1.6080, 1.4699, 1.2432], + device='cuda:0'), covar=tensor([0.0260, 0.0224, 0.0549, 0.0289, 0.0177, 0.0524, 0.0256, 0.0354], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0110, 0.0137, 0.0116, 0.0103, 0.0099, 0.0090, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.7463e-05, 8.6688e-05, 1.0942e-04, 9.1146e-05, 8.1537e-05, 7.3738e-05, + 6.8507e-05, 8.4334e-05], device='cuda:0') +2023-03-26 06:41:51,884 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9173, 2.0247, 1.3069, 2.0971, 1.9729, 1.6054, 2.9621, 1.9260], + device='cuda:0'), covar=tensor([0.1605, 0.2623, 0.3890, 0.3594, 0.3058, 0.1877, 0.2628, 0.2272], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0237, 0.0256, 0.0233, 0.0192, 0.0213, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:41:59,891 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33511.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:42:00,536 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7532, 1.6898, 1.9238, 2.0146, 2.0292, 4.3819, 1.8121, 2.1184], + device='cuda:0'), covar=tensor([0.1005, 0.1716, 0.1209, 0.1071, 0.1475, 0.0240, 0.1355, 0.1659], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0081, 0.0077, 0.0079, 0.0093, 0.0083, 0.0084, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 06:42:03,571 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7797, 4.0793, 3.9473, 2.1426, 4.1587, 3.0770, 0.6741, 2.9495], + device='cuda:0'), covar=tensor([0.2789, 0.1370, 0.1289, 0.2770, 0.0829, 0.0855, 0.4414, 0.1171], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0173, 0.0163, 0.0129, 0.0156, 0.0123, 0.0146, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 06:42:14,017 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.841e+02 2.130e+02 2.477e+02 4.983e+02, threshold=4.260e+02, percent-clipped=3.0 +2023-03-26 06:42:24,104 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33533.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:42:32,810 INFO [finetune.py:976] (0/7) Epoch 6, batch 4900, loss[loss=0.2091, simple_loss=0.2713, pruned_loss=0.07341, over 4925.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.276, pruned_loss=0.07738, over 955046.50 frames. ], batch size: 29, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:43:28,493 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4252, 1.3503, 1.8666, 2.8589, 1.8907, 2.1916, 0.7777, 2.2813], + device='cuda:0'), covar=tensor([0.1867, 0.1638, 0.1287, 0.0650, 0.0948, 0.1431, 0.1996, 0.0744], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0119, 0.0136, 0.0167, 0.0103, 0.0141, 0.0129, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 06:43:36,449 INFO [finetune.py:976] (0/7) Epoch 6, batch 4950, loss[loss=0.2213, simple_loss=0.2829, pruned_loss=0.07986, over 4769.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.2777, pruned_loss=0.07797, over 955244.07 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:44:20,458 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33622.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:44:22,686 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.562e+01 1.670e+02 2.087e+02 2.382e+02 5.310e+02, threshold=4.173e+02, percent-clipped=3.0 +2023-03-26 06:44:41,931 INFO [finetune.py:976] (0/7) Epoch 6, batch 5000, loss[loss=0.2397, simple_loss=0.2855, pruned_loss=0.09692, over 4892.00 frames. ], tot_loss[loss=0.2143, simple_loss=0.2755, pruned_loss=0.07656, over 956730.20 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:45:14,312 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33670.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:19,059 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:26,106 INFO [finetune.py:976] (0/7) Epoch 6, batch 5050, loss[loss=0.1736, simple_loss=0.2358, pruned_loss=0.05564, over 4844.00 frames. ], tot_loss[loss=0.2118, simple_loss=0.2726, pruned_loss=0.07547, over 957653.08 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:45:44,983 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 06:45:50,298 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.621e+02 1.877e+02 2.380e+02 3.773e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-26 06:45:50,390 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33725.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:58,819 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33738.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:45:59,296 INFO [finetune.py:976] (0/7) Epoch 6, batch 5100, loss[loss=0.2646, simple_loss=0.3085, pruned_loss=0.1104, over 4831.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2695, pruned_loss=0.07497, over 957010.64 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:46:06,663 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9022, 2.5684, 2.1527, 1.1356, 2.3240, 2.1586, 1.8924, 2.1550], + device='cuda:0'), covar=tensor([0.0653, 0.0901, 0.1482, 0.2179, 0.1446, 0.2108, 0.2241, 0.1052], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0203, 0.0203, 0.0190, 0.0218, 0.0210, 0.0222, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:46:10,929 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 06:46:22,563 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:46:32,748 INFO [finetune.py:976] (0/7) Epoch 6, batch 5150, loss[loss=0.2351, simple_loss=0.303, pruned_loss=0.08358, over 4851.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2681, pruned_loss=0.0743, over 955895.36 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:46:46,067 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1105, 1.1958, 1.1673, 1.3283, 1.3109, 2.4636, 1.0850, 1.3921], + device='cuda:0'), covar=tensor([0.1116, 0.2048, 0.1171, 0.1076, 0.1839, 0.0438, 0.1688, 0.1852], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 06:46:48,243 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:46:57,549 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.768e+02 2.107e+02 2.614e+02 3.782e+02, threshold=4.214e+02, percent-clipped=1.0 +2023-03-26 06:46:57,935 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 06:46:59,467 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33828.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:47:06,626 INFO [finetune.py:976] (0/7) Epoch 6, batch 5200, loss[loss=0.2212, simple_loss=0.2821, pruned_loss=0.08016, over 4850.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2717, pruned_loss=0.07487, over 956808.46 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:47:19,506 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:47:25,746 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0172, 1.7951, 1.5730, 1.7506, 1.7551, 1.7702, 1.6933, 2.5003], + device='cuda:0'), covar=tensor([0.5363, 0.5752, 0.4550, 0.5606, 0.5346, 0.3268, 0.5528, 0.2219], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0259, 0.0221, 0.0283, 0.0240, 0.0204, 0.0245, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:47:42,778 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7308, 1.5655, 1.5562, 0.8382, 1.6508, 1.8105, 1.8209, 1.4806], + device='cuda:0'), covar=tensor([0.0982, 0.0696, 0.0442, 0.0620, 0.0405, 0.0589, 0.0343, 0.0670], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0158, 0.0121, 0.0139, 0.0132, 0.0125, 0.0148, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.7353e-05, 1.1694e-04, 8.7653e-05, 1.0129e-04, 9.4938e-05, 9.2041e-05, + 1.0930e-04, 1.0796e-04], device='cuda:0') +2023-03-26 06:47:44,971 INFO [finetune.py:976] (0/7) Epoch 6, batch 5250, loss[loss=0.2242, simple_loss=0.2771, pruned_loss=0.08565, over 4821.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2747, pruned_loss=0.0758, over 958112.55 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:48:03,026 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7142, 1.7479, 2.0631, 1.9677, 2.0188, 4.5318, 1.7505, 2.1447], + device='cuda:0'), covar=tensor([0.0993, 0.1746, 0.1136, 0.1057, 0.1465, 0.0175, 0.1357, 0.1639], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0081, 0.0077, 0.0079, 0.0092, 0.0084, 0.0084, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 06:48:10,025 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.728e+02 2.112e+02 2.559e+02 4.196e+02, threshold=4.224e+02, percent-clipped=0.0 +2023-03-26 06:48:18,714 INFO [finetune.py:976] (0/7) Epoch 6, batch 5300, loss[loss=0.1735, simple_loss=0.2473, pruned_loss=0.04982, over 4782.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2752, pruned_loss=0.07595, over 957404.93 frames. ], batch size: 26, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:48:18,809 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:48:21,968 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-26 06:48:44,315 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7197, 1.6051, 2.1946, 3.5236, 2.4137, 2.3993, 0.9255, 2.7678], + device='cuda:0'), covar=tensor([0.1722, 0.1504, 0.1378, 0.0606, 0.0825, 0.1468, 0.2109, 0.0574], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0119, 0.0137, 0.0167, 0.0103, 0.0141, 0.0129, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 06:48:51,057 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2118, 2.2720, 1.8416, 1.7691, 2.3704, 2.5729, 2.2607, 2.0292], + device='cuda:0'), covar=tensor([0.0332, 0.0343, 0.0543, 0.0403, 0.0362, 0.0577, 0.0349, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0110, 0.0138, 0.0116, 0.0104, 0.0099, 0.0090, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.7764e-05, 8.6799e-05, 1.1006e-04, 9.1354e-05, 8.1958e-05, 7.3562e-05, + 6.8426e-05, 8.4426e-05], device='cuda:0') +2023-03-26 06:48:53,986 INFO [finetune.py:976] (0/7) Epoch 6, batch 5350, loss[loss=0.2121, simple_loss=0.2678, pruned_loss=0.07817, over 4826.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2761, pruned_loss=0.07634, over 957776.46 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:48:58,829 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1905, 1.8724, 1.4359, 0.5634, 1.6740, 1.8382, 1.6417, 1.7346], + device='cuda:0'), covar=tensor([0.0937, 0.0890, 0.1424, 0.2034, 0.1279, 0.2258, 0.2100, 0.0937], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0200, 0.0189, 0.0216, 0.0207, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:49:06,444 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-34000.pt +2023-03-26 06:49:07,548 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34000.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:49:16,327 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 06:49:27,911 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 06:49:40,313 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.730e+02 2.013e+02 2.437e+02 5.230e+02, threshold=4.026e+02, percent-clipped=3.0 +2023-03-26 06:49:50,506 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34033.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:49:59,342 INFO [finetune.py:976] (0/7) Epoch 6, batch 5400, loss[loss=0.1988, simple_loss=0.2577, pruned_loss=0.06991, over 4796.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2727, pruned_loss=0.07535, over 956928.94 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:50:02,510 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:50:42,160 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5712, 1.6078, 1.3132, 1.4343, 1.7740, 1.7470, 1.5970, 1.4025], + device='cuda:0'), covar=tensor([0.0324, 0.0309, 0.0584, 0.0336, 0.0255, 0.0490, 0.0277, 0.0419], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0110, 0.0137, 0.0116, 0.0104, 0.0099, 0.0090, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.7705e-05, 8.6733e-05, 1.0997e-04, 9.1278e-05, 8.1882e-05, 7.3525e-05, + 6.8460e-05, 8.4190e-05], device='cuda:0') +2023-03-26 06:50:48,116 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 06:50:51,575 INFO [finetune.py:976] (0/7) Epoch 6, batch 5450, loss[loss=0.18, simple_loss=0.2422, pruned_loss=0.05889, over 4822.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2702, pruned_loss=0.07462, over 956659.72 frames. ], batch size: 41, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:51:04,655 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34105.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:17,059 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.588e+02 1.777e+02 2.163e+02 4.002e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-03-26 06:51:19,933 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34128.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:27,297 INFO [finetune.py:976] (0/7) Epoch 6, batch 5500, loss[loss=0.2011, simple_loss=0.2605, pruned_loss=0.07086, over 4931.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2671, pruned_loss=0.07348, over 957369.72 frames. ], batch size: 38, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:51:31,015 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34145.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:44,202 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8371, 3.9049, 3.6837, 1.6719, 4.0515, 3.0184, 0.9289, 2.8328], + device='cuda:0'), covar=tensor([0.2479, 0.1736, 0.1460, 0.3332, 0.0881, 0.0937, 0.4346, 0.1396], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0171, 0.0161, 0.0127, 0.0154, 0.0121, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 06:51:50,700 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:52:00,938 INFO [finetune.py:976] (0/7) Epoch 6, batch 5550, loss[loss=0.1947, simple_loss=0.2596, pruned_loss=0.06491, over 4768.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2698, pruned_loss=0.07484, over 955334.06 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:52:15,683 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:52:37,389 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.703e+02 1.979e+02 2.376e+02 4.570e+02, threshold=3.959e+02, percent-clipped=2.0 +2023-03-26 06:52:55,409 INFO [finetune.py:976] (0/7) Epoch 6, batch 5600, loss[loss=0.2187, simple_loss=0.2877, pruned_loss=0.07488, over 4863.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.273, pruned_loss=0.07556, over 952748.05 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:53:35,037 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 06:53:54,498 INFO [finetune.py:976] (0/7) Epoch 6, batch 5650, loss[loss=0.2351, simple_loss=0.2979, pruned_loss=0.0862, over 4797.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2764, pruned_loss=0.07646, over 952792.76 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:53:58,438 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34295.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:53:59,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2708, 2.1747, 1.8108, 2.2765, 2.2397, 1.9793, 2.6412, 2.2278], + device='cuda:0'), covar=tensor([0.1442, 0.2788, 0.3643, 0.2969, 0.2680, 0.1812, 0.3351, 0.2150], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0236, 0.0256, 0.0234, 0.0192, 0.0213, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:54:35,325 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.805e+01 1.675e+02 2.029e+02 2.481e+02 4.265e+02, threshold=4.057e+02, percent-clipped=3.0 +2023-03-26 06:54:40,153 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:54:48,432 INFO [finetune.py:976] (0/7) Epoch 6, batch 5700, loss[loss=0.1213, simple_loss=0.178, pruned_loss=0.03224, over 4136.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2703, pruned_loss=0.07521, over 932032.95 frames. ], batch size: 18, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:55:10,893 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6924, 2.3323, 2.5300, 2.4434, 2.1910, 2.2538, 2.3939, 2.3880], + device='cuda:0'), covar=tensor([0.4129, 0.5540, 0.5089, 0.5751, 0.7124, 0.5190, 0.7540, 0.4660], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0244, 0.0254, 0.0255, 0.0242, 0.0219, 0.0272, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:55:19,478 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-6.pt +2023-03-26 06:55:38,389 INFO [finetune.py:976] (0/7) Epoch 7, batch 0, loss[loss=0.1809, simple_loss=0.2512, pruned_loss=0.05528, over 4718.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2512, pruned_loss=0.05528, over 4718.00 frames. ], batch size: 27, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:55:38,390 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 06:55:55,928 INFO [finetune.py:1010] (0/7) Epoch 7, validation: loss=0.165, simple_loss=0.2365, pruned_loss=0.04677, over 2265189.00 frames. +2023-03-26 06:55:55,929 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 06:56:14,684 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:56:16,536 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1436, 2.0443, 1.6335, 2.0876, 1.9704, 1.8438, 1.9071, 2.8500], + device='cuda:0'), covar=tensor([0.5590, 0.6811, 0.4729, 0.6445, 0.6047, 0.3393, 0.6681, 0.2105], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0282, 0.0240, 0.0204, 0.0244, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 06:56:36,606 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:56:59,197 INFO [finetune.py:976] (0/7) Epoch 7, batch 50, loss[loss=0.2091, simple_loss=0.2656, pruned_loss=0.0763, over 4855.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2799, pruned_loss=0.07883, over 216822.23 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:57:09,544 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.600e+02 2.022e+02 2.565e+02 5.766e+02, threshold=4.045e+02, percent-clipped=4.0 +2023-03-26 06:57:28,320 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-03-26 06:57:49,954 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-26 06:58:05,579 INFO [finetune.py:976] (0/7) Epoch 7, batch 100, loss[loss=0.2278, simple_loss=0.2716, pruned_loss=0.09198, over 4750.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2693, pruned_loss=0.07319, over 380695.91 frames. ], batch size: 54, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:58:45,254 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:59:06,468 INFO [finetune.py:976] (0/7) Epoch 7, batch 150, loss[loss=0.177, simple_loss=0.2512, pruned_loss=0.05138, over 4931.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2649, pruned_loss=0.07263, over 509408.87 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:59:17,659 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.586e+02 1.902e+02 2.328e+02 6.438e+02, threshold=3.804e+02, percent-clipped=3.0 +2023-03-26 07:00:10,780 INFO [finetune.py:976] (0/7) Epoch 7, batch 200, loss[loss=0.2168, simple_loss=0.2698, pruned_loss=0.08188, over 4891.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2646, pruned_loss=0.07305, over 609236.22 frames. ], batch size: 35, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 07:00:42,237 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:00:43,443 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:13,421 INFO [finetune.py:976] (0/7) Epoch 7, batch 250, loss[loss=0.2173, simple_loss=0.2843, pruned_loss=0.07517, over 4932.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2698, pruned_loss=0.0752, over 684540.86 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:01:22,724 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34622.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:24,454 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.737e+02 2.023e+02 2.550e+02 3.958e+02, threshold=4.047e+02, percent-clipped=1.0 +2023-03-26 07:01:25,176 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34626.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:45,022 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:53,159 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4921, 1.3945, 1.9991, 3.1801, 2.1034, 2.2908, 0.9283, 2.4940], + device='cuda:0'), covar=tensor([0.1830, 0.1561, 0.1361, 0.0592, 0.0876, 0.1377, 0.2059, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0118, 0.0136, 0.0166, 0.0102, 0.0141, 0.0129, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 07:01:56,245 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8557, 1.5926, 2.3553, 3.8141, 2.5380, 2.5837, 0.7171, 3.0111], + device='cuda:0'), covar=tensor([0.1781, 0.1526, 0.1359, 0.0608, 0.0806, 0.1739, 0.2185, 0.0519], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0136, 0.0166, 0.0102, 0.0141, 0.0129, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 07:01:57,384 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:02:14,154 INFO [finetune.py:976] (0/7) Epoch 7, batch 300, loss[loss=0.2036, simple_loss=0.2794, pruned_loss=0.06389, over 4836.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2742, pruned_loss=0.07581, over 745188.31 frames. ], batch size: 47, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:02:14,860 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2258, 3.6641, 3.8469, 4.0373, 3.9600, 3.7256, 4.2898, 1.7161], + device='cuda:0'), covar=tensor([0.0750, 0.0952, 0.0728, 0.0853, 0.1175, 0.1301, 0.0658, 0.4423], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0244, 0.0275, 0.0295, 0.0336, 0.0284, 0.0304, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:02:34,486 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34683.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:02:36,964 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34687.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:02:43,339 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.07 vs. limit=2.0 +2023-03-26 07:02:55,194 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34700.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:03:05,995 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9927, 2.0292, 1.9158, 1.3865, 2.1492, 2.1235, 2.0931, 1.7322], + device='cuda:0'), covar=tensor([0.0614, 0.0555, 0.0756, 0.0928, 0.0470, 0.0701, 0.0593, 0.0996], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0135, 0.0145, 0.0129, 0.0114, 0.0146, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:03:06,626 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34710.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:03:15,678 INFO [finetune.py:976] (0/7) Epoch 7, batch 350, loss[loss=0.24, simple_loss=0.2967, pruned_loss=0.09162, over 4751.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2768, pruned_loss=0.07717, over 790223.40 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:03:27,161 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.664e+02 2.043e+02 2.496e+02 5.690e+02, threshold=4.087e+02, percent-clipped=3.0 +2023-03-26 07:03:54,737 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:04:16,438 INFO [finetune.py:976] (0/7) Epoch 7, batch 400, loss[loss=0.1956, simple_loss=0.2543, pruned_loss=0.06842, over 4798.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2768, pruned_loss=0.07653, over 828497.39 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:04:24,030 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34771.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 07:04:55,054 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:05:14,340 INFO [finetune.py:976] (0/7) Epoch 7, batch 450, loss[loss=0.228, simple_loss=0.2869, pruned_loss=0.08459, over 4831.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2749, pruned_loss=0.07552, over 857357.16 frames. ], batch size: 39, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:05:25,983 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.133e+01 1.634e+02 1.827e+02 2.211e+02 3.915e+02, threshold=3.654e+02, percent-clipped=0.0 +2023-03-26 07:05:50,914 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:06:11,591 INFO [finetune.py:976] (0/7) Epoch 7, batch 500, loss[loss=0.1897, simple_loss=0.2504, pruned_loss=0.06447, over 4913.00 frames. ], tot_loss[loss=0.2119, simple_loss=0.273, pruned_loss=0.07545, over 878643.16 frames. ], batch size: 37, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:07:15,211 INFO [finetune.py:976] (0/7) Epoch 7, batch 550, loss[loss=0.1993, simple_loss=0.2668, pruned_loss=0.06592, over 4818.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2699, pruned_loss=0.0745, over 895028.83 frames. ], batch size: 30, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:07:26,487 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.634e+02 2.013e+02 2.383e+02 4.182e+02, threshold=4.026e+02, percent-clipped=3.0 +2023-03-26 07:07:53,596 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:08:14,352 INFO [finetune.py:976] (0/7) Epoch 7, batch 600, loss[loss=0.257, simple_loss=0.3184, pruned_loss=0.09782, over 4871.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2709, pruned_loss=0.07539, over 909623.38 frames. ], batch size: 44, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:08:31,409 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34978.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:08:34,781 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34982.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:08:41,349 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 07:09:16,193 INFO [finetune.py:976] (0/7) Epoch 7, batch 650, loss[loss=0.1946, simple_loss=0.2625, pruned_loss=0.06335, over 4924.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2745, pruned_loss=0.07656, over 919570.08 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:09:27,411 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.303e+02 1.723e+02 2.031e+02 2.472e+02 3.902e+02, threshold=4.061e+02, percent-clipped=0.0 +2023-03-26 07:10:17,350 INFO [finetune.py:976] (0/7) Epoch 7, batch 700, loss[loss=0.2313, simple_loss=0.2715, pruned_loss=0.0955, over 3959.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2758, pruned_loss=0.07635, over 926935.97 frames. ], batch size: 17, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:10:17,414 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35066.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:10:18,806 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 07:10:22,905 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 07:10:46,500 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:11:11,733 INFO [finetune.py:976] (0/7) Epoch 7, batch 750, loss[loss=0.1999, simple_loss=0.2518, pruned_loss=0.074, over 4796.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2768, pruned_loss=0.07652, over 930239.07 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:11:21,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4288, 2.8469, 2.2759, 1.8245, 2.8728, 2.7286, 2.6542, 2.4519], + device='cuda:0'), covar=tensor([0.0688, 0.0574, 0.0894, 0.0976, 0.0393, 0.0785, 0.0706, 0.0896], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0133, 0.0143, 0.0127, 0.0113, 0.0144, 0.0144, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:11:23,577 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.641e+02 2.027e+02 2.429e+02 3.682e+02, threshold=4.054e+02, percent-clipped=0.0 +2023-03-26 07:11:33,043 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1553, 1.3310, 0.7156, 2.0605, 2.3329, 1.7744, 1.6867, 2.0029], + device='cuda:0'), covar=tensor([0.1450, 0.2196, 0.2414, 0.1124, 0.1962, 0.1884, 0.1429, 0.1835], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0093, 0.0124, 0.0096, 0.0101, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 07:12:01,753 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35153.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:12:14,612 INFO [finetune.py:976] (0/7) Epoch 7, batch 800, loss[loss=0.1937, simple_loss=0.2581, pruned_loss=0.06469, over 4796.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2769, pruned_loss=0.07647, over 936947.22 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:12:26,872 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1032, 2.1521, 1.7362, 1.5999, 2.3543, 2.3421, 2.1232, 1.9292], + device='cuda:0'), covar=tensor([0.0306, 0.0300, 0.0560, 0.0358, 0.0266, 0.0469, 0.0251, 0.0362], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0112, 0.0140, 0.0117, 0.0105, 0.0101, 0.0091, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.9361e-05, 8.8117e-05, 1.1162e-04, 9.2568e-05, 8.3129e-05, 7.5057e-05, + 6.9173e-05, 8.5793e-05], device='cuda:0') +2023-03-26 07:13:17,535 INFO [finetune.py:976] (0/7) Epoch 7, batch 850, loss[loss=0.2164, simple_loss=0.2593, pruned_loss=0.08675, over 4322.00 frames. ], tot_loss[loss=0.2124, simple_loss=0.2742, pruned_loss=0.07533, over 941651.78 frames. ], batch size: 18, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:13:27,729 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.645e+02 1.948e+02 2.227e+02 3.525e+02, threshold=3.897e+02, percent-clipped=0.0 +2023-03-26 07:13:31,078 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6960, 2.3370, 2.0350, 1.0478, 2.2413, 1.9866, 1.8181, 2.1108], + device='cuda:0'), covar=tensor([0.0871, 0.1058, 0.1776, 0.2433, 0.1693, 0.2460, 0.2323, 0.1191], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0202, 0.0200, 0.0189, 0.0217, 0.0207, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:13:55,199 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35249.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:14:14,655 INFO [finetune.py:976] (0/7) Epoch 7, batch 900, loss[loss=0.1464, simple_loss=0.2182, pruned_loss=0.03726, over 4766.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2709, pruned_loss=0.07437, over 943298.40 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:14:16,044 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 07:14:32,294 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35278.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:14:35,276 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:14:55,279 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:16,601 INFO [finetune.py:976] (0/7) Epoch 7, batch 950, loss[loss=0.1804, simple_loss=0.2389, pruned_loss=0.06097, over 4778.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2689, pruned_loss=0.0737, over 945490.48 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:15:31,371 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.515e+02 1.811e+02 2.306e+02 3.628e+02, threshold=3.621e+02, percent-clipped=0.0 +2023-03-26 07:15:31,440 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:33,889 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:40,734 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9046, 1.3044, 1.7589, 1.7157, 1.5087, 1.5419, 1.6176, 1.6090], + device='cuda:0'), covar=tensor([0.4499, 0.5841, 0.4998, 0.5426, 0.6469, 0.4861, 0.6821, 0.4486], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0243, 0.0254, 0.0254, 0.0242, 0.0218, 0.0271, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:15:50,904 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35345.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:16:13,221 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-03-26 07:16:18,536 INFO [finetune.py:976] (0/7) Epoch 7, batch 1000, loss[loss=0.2234, simple_loss=0.2833, pruned_loss=0.08175, over 4940.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2721, pruned_loss=0.07496, over 947996.19 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:16:18,649 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35366.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:16:27,461 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:17:08,111 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35406.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:17:17,935 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35414.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:17:19,085 INFO [finetune.py:976] (0/7) Epoch 7, batch 1050, loss[loss=0.2241, simple_loss=0.2821, pruned_loss=0.08311, over 4897.00 frames. ], tot_loss[loss=0.2123, simple_loss=0.274, pruned_loss=0.07534, over 950663.95 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:17:30,142 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.696e+02 1.924e+02 2.371e+02 5.787e+02, threshold=3.848e+02, percent-clipped=4.0 +2023-03-26 07:17:41,403 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:17:50,924 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7667, 1.7502, 1.8334, 1.0931, 2.0227, 1.8619, 1.8032, 1.5716], + device='cuda:0'), covar=tensor([0.0613, 0.0745, 0.0730, 0.0999, 0.0575, 0.0780, 0.0667, 0.1113], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0134, 0.0143, 0.0126, 0.0113, 0.0144, 0.0145, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:17:59,811 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:18:02,361 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 07:18:20,377 INFO [finetune.py:976] (0/7) Epoch 7, batch 1100, loss[loss=0.2052, simple_loss=0.2601, pruned_loss=0.07518, over 4800.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2751, pruned_loss=0.07621, over 951463.76 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:18:22,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8097, 1.5768, 1.4653, 1.2157, 1.5904, 1.5779, 1.5213, 2.0792], + device='cuda:0'), covar=tensor([0.5500, 0.6021, 0.4265, 0.5106, 0.4666, 0.3173, 0.5098, 0.2355], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0281, 0.0239, 0.0204, 0.0244, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:19:22,492 INFO [finetune.py:976] (0/7) Epoch 7, batch 1150, loss[loss=0.2501, simple_loss=0.3116, pruned_loss=0.09435, over 4818.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2743, pruned_loss=0.07566, over 950182.10 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:19:33,760 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.764e+02 2.068e+02 2.432e+02 4.937e+02, threshold=4.137e+02, percent-clipped=2.0 +2023-03-26 07:19:41,298 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:20:24,974 INFO [finetune.py:976] (0/7) Epoch 7, batch 1200, loss[loss=0.2278, simple_loss=0.2886, pruned_loss=0.08355, over 4901.00 frames. ], tot_loss[loss=0.213, simple_loss=0.274, pruned_loss=0.07597, over 951003.83 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:20:51,621 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:21:20,132 INFO [finetune.py:976] (0/7) Epoch 7, batch 1250, loss[loss=0.2738, simple_loss=0.303, pruned_loss=0.1223, over 4887.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2706, pruned_loss=0.07433, over 951456.71 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:21:31,425 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.688e+02 2.018e+02 2.672e+02 1.298e+03, threshold=4.035e+02, percent-clipped=4.0 +2023-03-26 07:22:16,797 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1400, 1.3216, 1.4041, 0.6470, 1.1995, 1.5405, 1.6049, 1.2778], + device='cuda:0'), covar=tensor([0.0946, 0.0565, 0.0430, 0.0626, 0.0490, 0.0550, 0.0323, 0.0688], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0156, 0.0120, 0.0137, 0.0132, 0.0124, 0.0146, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.6864e-05, 1.1546e-04, 8.6634e-05, 9.9502e-05, 9.4578e-05, 9.1764e-05, + 1.0752e-04, 1.0703e-04], device='cuda:0') +2023-03-26 07:22:18,864 INFO [finetune.py:976] (0/7) Epoch 7, batch 1300, loss[loss=0.181, simple_loss=0.255, pruned_loss=0.05348, over 4898.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2672, pruned_loss=0.07233, over 954282.96 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:23:05,606 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35701.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:23:25,710 INFO [finetune.py:976] (0/7) Epoch 7, batch 1350, loss[loss=0.1918, simple_loss=0.266, pruned_loss=0.05878, over 4864.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2669, pruned_loss=0.07221, over 951257.82 frames. ], batch size: 31, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:23:38,296 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.109e+01 1.659e+02 1.871e+02 2.249e+02 4.421e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-26 07:23:40,798 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:06,295 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:25,447 INFO [finetune.py:976] (0/7) Epoch 7, batch 1400, loss[loss=0.2327, simple_loss=0.2927, pruned_loss=0.08637, over 4810.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2698, pruned_loss=0.0732, over 951072.38 frames. ], batch size: 40, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:24:33,474 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:34,784 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.64 vs. limit=5.0 +2023-03-26 07:25:01,857 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35796.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:25:17,189 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:25:23,859 INFO [finetune.py:976] (0/7) Epoch 7, batch 1450, loss[loss=0.2146, simple_loss=0.2865, pruned_loss=0.07139, over 4827.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2732, pruned_loss=0.07433, over 951686.40 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:25:33,656 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.737e+02 2.011e+02 2.560e+02 4.083e+02, threshold=4.021e+02, percent-clipped=3.0 +2023-03-26 07:25:43,665 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:26:24,786 INFO [finetune.py:976] (0/7) Epoch 7, batch 1500, loss[loss=0.2497, simple_loss=0.3109, pruned_loss=0.09427, over 4797.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2745, pruned_loss=0.07494, over 952875.71 frames. ], batch size: 45, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:26:26,301 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-26 07:26:32,647 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:26:47,871 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:27:11,196 INFO [finetune.py:976] (0/7) Epoch 7, batch 1550, loss[loss=0.2125, simple_loss=0.2768, pruned_loss=0.07406, over 4748.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2746, pruned_loss=0.07486, over 954925.98 frames. ], batch size: 59, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:27:17,686 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.940e+01 1.555e+02 1.902e+02 2.352e+02 4.828e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-03-26 07:27:32,039 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 07:27:43,326 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5011, 1.5071, 1.9117, 1.8857, 1.6248, 3.4452, 1.3072, 1.5108], + device='cuda:0'), covar=tensor([0.0997, 0.1731, 0.1105, 0.0984, 0.1595, 0.0259, 0.1510, 0.1795], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 07:27:45,066 INFO [finetune.py:976] (0/7) Epoch 7, batch 1600, loss[loss=0.1759, simple_loss=0.2414, pruned_loss=0.05521, over 4755.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2713, pruned_loss=0.07386, over 954577.44 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:27:52,568 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 07:28:24,111 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-36000.pt +2023-03-26 07:28:25,870 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36001.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:28:35,456 INFO [finetune.py:976] (0/7) Epoch 7, batch 1650, loss[loss=0.1745, simple_loss=0.2442, pruned_loss=0.05238, over 4908.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2678, pruned_loss=0.07269, over 953044.63 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:28:41,915 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.569e+02 1.867e+02 2.342e+02 3.778e+02, threshold=3.734e+02, percent-clipped=0.0 +2023-03-26 07:28:50,701 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:29:09,234 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36049.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:29:11,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2018, 2.1669, 2.3273, 0.9160, 2.5919, 2.7311, 2.3205, 2.1132], + device='cuda:0'), covar=tensor([0.0842, 0.0573, 0.0352, 0.0812, 0.0396, 0.0448, 0.0436, 0.0617], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0157, 0.0120, 0.0137, 0.0132, 0.0124, 0.0145, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.6855e-05, 1.1560e-04, 8.7115e-05, 9.9725e-05, 9.4577e-05, 9.1317e-05, + 1.0736e-04, 1.0718e-04], device='cuda:0') +2023-03-26 07:29:25,212 INFO [finetune.py:976] (0/7) Epoch 7, batch 1700, loss[loss=0.2611, simple_loss=0.3073, pruned_loss=0.1074, over 4833.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2669, pruned_loss=0.073, over 953884.99 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:29:39,775 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:30:15,625 INFO [finetune.py:976] (0/7) Epoch 7, batch 1750, loss[loss=0.2007, simple_loss=0.2547, pruned_loss=0.07335, over 4722.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2689, pruned_loss=0.0738, over 953400.37 frames. ], batch size: 23, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:30:27,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.647e+02 1.960e+02 2.452e+02 4.962e+02, threshold=3.920e+02, percent-clipped=3.0 +2023-03-26 07:30:28,490 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36127.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:18,546 INFO [finetune.py:976] (0/7) Epoch 7, batch 1800, loss[loss=0.1974, simple_loss=0.2691, pruned_loss=0.06291, over 4806.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2724, pruned_loss=0.07454, over 953852.73 frames. ], batch size: 51, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:31:19,215 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:19,361 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 07:31:38,148 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:47,171 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:32:21,146 INFO [finetune.py:976] (0/7) Epoch 7, batch 1850, loss[loss=0.192, simple_loss=0.2456, pruned_loss=0.06916, over 4193.00 frames. ], tot_loss[loss=0.2119, simple_loss=0.2733, pruned_loss=0.07522, over 955239.59 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:32:21,268 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5529, 2.4417, 2.8882, 1.9328, 2.7960, 3.0014, 2.2395, 2.9676], + device='cuda:0'), covar=tensor([0.1631, 0.1888, 0.1666, 0.2460, 0.1055, 0.1823, 0.2454, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0203, 0.0196, 0.0195, 0.0180, 0.0219, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:32:23,080 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 07:32:33,456 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.736e+02 2.131e+02 2.651e+02 6.216e+02, threshold=4.263e+02, percent-clipped=3.0 +2023-03-26 07:32:40,446 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:32:50,348 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36242.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:33:21,418 INFO [finetune.py:976] (0/7) Epoch 7, batch 1900, loss[loss=0.2216, simple_loss=0.2876, pruned_loss=0.07782, over 4888.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2741, pruned_loss=0.07499, over 954875.62 frames. ], batch size: 35, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:34:25,473 INFO [finetune.py:976] (0/7) Epoch 7, batch 1950, loss[loss=0.1959, simple_loss=0.2541, pruned_loss=0.06884, over 4822.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2717, pruned_loss=0.07409, over 955061.09 frames. ], batch size: 41, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:34:35,936 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 07:34:36,925 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.969e+01 1.685e+02 2.051e+02 2.475e+02 4.640e+02, threshold=4.103e+02, percent-clipped=3.0 +2023-03-26 07:34:46,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1586, 3.5712, 3.7193, 4.0155, 3.8950, 3.6769, 4.2038, 1.2051], + device='cuda:0'), covar=tensor([0.0806, 0.0857, 0.0874, 0.0962, 0.1269, 0.1440, 0.0737, 0.5402], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0240, 0.0270, 0.0288, 0.0329, 0.0280, 0.0299, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:35:28,761 INFO [finetune.py:976] (0/7) Epoch 7, batch 2000, loss[loss=0.1919, simple_loss=0.2593, pruned_loss=0.06224, over 4771.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2684, pruned_loss=0.07283, over 955018.91 frames. ], batch size: 54, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:36:30,438 INFO [finetune.py:976] (0/7) Epoch 7, batch 2050, loss[loss=0.2561, simple_loss=0.3076, pruned_loss=0.1023, over 4857.00 frames. ], tot_loss[loss=0.2038, simple_loss=0.2648, pruned_loss=0.07133, over 954684.33 frames. ], batch size: 44, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:36:43,670 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.910e+01 1.532e+02 1.893e+02 2.218e+02 7.941e+02, threshold=3.786e+02, percent-clipped=2.0 +2023-03-26 07:36:44,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36427.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:09,328 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 07:37:34,745 INFO [finetune.py:976] (0/7) Epoch 7, batch 2100, loss[loss=0.2212, simple_loss=0.2842, pruned_loss=0.0791, over 4823.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2648, pruned_loss=0.07093, over 954349.99 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:37:35,463 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:45,790 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:47,809 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.29 vs. limit=5.0 +2023-03-26 07:38:36,450 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:38:36,990 INFO [finetune.py:976] (0/7) Epoch 7, batch 2150, loss[loss=0.2753, simple_loss=0.3287, pruned_loss=0.1109, over 4803.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2683, pruned_loss=0.07259, over 954957.15 frames. ], batch size: 51, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:38:48,004 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.787e+02 2.211e+02 2.590e+02 5.595e+02, threshold=4.423e+02, percent-clipped=4.0 +2023-03-26 07:39:00,122 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36537.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:39:35,927 INFO [finetune.py:976] (0/7) Epoch 7, batch 2200, loss[loss=0.2029, simple_loss=0.2783, pruned_loss=0.06377, over 4827.00 frames. ], tot_loss[loss=0.2084, simple_loss=0.2703, pruned_loss=0.0733, over 952802.38 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:40:25,401 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:27,769 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36608.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:30,219 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:38,177 INFO [finetune.py:976] (0/7) Epoch 7, batch 2250, loss[loss=0.166, simple_loss=0.2376, pruned_loss=0.04724, over 4768.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2736, pruned_loss=0.07491, over 955034.28 frames. ], batch size: 26, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:40:49,910 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.735e+02 1.950e+02 2.446e+02 5.153e+02, threshold=3.899e+02, percent-clipped=1.0 +2023-03-26 07:41:41,215 INFO [finetune.py:976] (0/7) Epoch 7, batch 2300, loss[loss=0.1552, simple_loss=0.2326, pruned_loss=0.0389, over 4814.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2748, pruned_loss=0.07508, over 955217.70 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:41:41,330 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:41:43,634 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:41:51,486 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36673.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:42:16,666 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 07:42:18,147 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36700.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:42:36,646 INFO [finetune.py:976] (0/7) Epoch 7, batch 2350, loss[loss=0.228, simple_loss=0.2783, pruned_loss=0.08887, over 4816.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.274, pruned_loss=0.07524, over 956723.05 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:42:49,194 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.509e+02 1.866e+02 2.321e+02 4.735e+02, threshold=3.732e+02, percent-clipped=2.0 +2023-03-26 07:42:57,840 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9244, 1.2938, 0.9245, 1.8239, 2.3531, 1.4594, 1.5840, 1.8103], + device='cuda:0'), covar=tensor([0.1533, 0.2206, 0.2138, 0.1236, 0.1840, 0.2005, 0.1518, 0.2010], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0113, 0.0092, 0.0124, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 07:43:08,375 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2415, 1.2921, 1.6282, 1.0873, 1.3095, 1.4340, 1.3302, 1.5003], + device='cuda:0'), covar=tensor([0.1348, 0.2376, 0.1352, 0.1552, 0.1001, 0.1302, 0.3058, 0.0969], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0203, 0.0196, 0.0195, 0.0180, 0.0219, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:43:29,343 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36761.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:43:37,712 INFO [finetune.py:976] (0/7) Epoch 7, batch 2400, loss[loss=0.1712, simple_loss=0.2419, pruned_loss=0.05028, over 4867.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2725, pruned_loss=0.07543, over 956579.40 frames. ], batch size: 34, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:44:28,028 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36805.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:44:40,195 INFO [finetune.py:976] (0/7) Epoch 7, batch 2450, loss[loss=0.173, simple_loss=0.242, pruned_loss=0.05197, over 4764.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2698, pruned_loss=0.07443, over 956000.59 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:44:51,706 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.798e+02 2.140e+02 2.594e+02 4.660e+02, threshold=4.281e+02, percent-clipped=3.0 +2023-03-26 07:45:10,260 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36837.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:45:31,471 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9748, 1.7262, 1.5571, 1.6868, 1.6850, 1.6660, 1.6528, 2.5154], + device='cuda:0'), covar=tensor([0.5695, 0.6722, 0.4629, 0.6299, 0.5548, 0.3318, 0.6226, 0.2182], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0258, 0.0220, 0.0281, 0.0240, 0.0205, 0.0244, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:45:49,129 INFO [finetune.py:976] (0/7) Epoch 7, batch 2500, loss[loss=0.2414, simple_loss=0.2993, pruned_loss=0.09172, over 4811.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2713, pruned_loss=0.07531, over 952762.38 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:45:49,274 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36866.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:46:12,542 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:46:32,261 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8676, 4.2367, 4.0538, 2.2524, 4.3648, 3.3230, 0.7217, 2.8948], + device='cuda:0'), covar=tensor([0.2625, 0.1550, 0.1314, 0.3106, 0.0757, 0.0877, 0.4715, 0.1458], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0171, 0.0160, 0.0127, 0.0152, 0.0121, 0.0144, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 07:46:44,094 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:46:52,300 INFO [finetune.py:976] (0/7) Epoch 7, batch 2550, loss[loss=0.2636, simple_loss=0.3107, pruned_loss=0.1082, over 4895.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2754, pruned_loss=0.07682, over 953453.87 frames. ], batch size: 35, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:46:54,938 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 07:47:03,264 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 1.620e+02 1.912e+02 2.307e+02 6.491e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-03-26 07:47:33,238 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5594, 1.0442, 0.7690, 1.4685, 2.0603, 0.7406, 1.2570, 1.4461], + device='cuda:0'), covar=tensor([0.1552, 0.2279, 0.1957, 0.1263, 0.1953, 0.2058, 0.1584, 0.1942], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0124, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 07:47:45,966 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36958.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:47,752 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36961.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:55,195 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:56,363 INFO [finetune.py:976] (0/7) Epoch 7, batch 2600, loss[loss=0.2365, simple_loss=0.2889, pruned_loss=0.09201, over 4818.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2753, pruned_loss=0.07609, over 954995.64 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:47:57,650 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:04,524 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4388, 1.4400, 1.7181, 1.6879, 1.5610, 3.4045, 1.2998, 1.5365], + device='cuda:0'), covar=tensor([0.1051, 0.1882, 0.1228, 0.1077, 0.1704, 0.0222, 0.1576, 0.1839], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0076, 0.0079, 0.0092, 0.0083, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 07:48:05,149 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:08,926 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-03-26 07:48:41,468 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:53,702 INFO [finetune.py:976] (0/7) Epoch 7, batch 2650, loss[loss=0.2155, simple_loss=0.2859, pruned_loss=0.0725, over 4859.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2763, pruned_loss=0.07603, over 955072.48 frames. ], batch size: 44, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:48:56,097 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37019.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:49:05,454 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.627e+02 1.954e+02 2.393e+02 3.704e+02, threshold=3.907e+02, percent-clipped=0.0 +2023-03-26 07:49:37,770 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 07:49:41,683 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37056.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:49:47,719 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:49:48,223 INFO [finetune.py:976] (0/7) Epoch 7, batch 2700, loss[loss=0.1784, simple_loss=0.2502, pruned_loss=0.05335, over 4840.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2743, pruned_loss=0.07506, over 955815.40 frames. ], batch size: 47, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:49:57,361 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4490, 1.5751, 1.8208, 1.7180, 1.6651, 3.5108, 1.4381, 1.6458], + device='cuda:0'), covar=tensor([0.1089, 0.1702, 0.1106, 0.1102, 0.1533, 0.0224, 0.1452, 0.1684], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0076, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 07:50:10,162 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1830, 1.7350, 1.1825, 2.2175, 2.5497, 1.7510, 1.9361, 2.1348], + device='cuda:0'), covar=tensor([0.1233, 0.1864, 0.1965, 0.0973, 0.1725, 0.1681, 0.1392, 0.1639], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0113, 0.0092, 0.0123, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 07:50:16,810 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9637, 2.1844, 1.7860, 1.5810, 2.2889, 2.3167, 2.2156, 1.9032], + device='cuda:0'), covar=tensor([0.0385, 0.0355, 0.0666, 0.0393, 0.0318, 0.0792, 0.0358, 0.0424], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0112, 0.0140, 0.0117, 0.0105, 0.0101, 0.0091, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.9714e-05, 8.7758e-05, 1.1201e-04, 9.1954e-05, 8.2717e-05, 7.4755e-05, + 6.9157e-05, 8.5753e-05], device='cuda:0') +2023-03-26 07:50:22,104 INFO [finetune.py:976] (0/7) Epoch 7, batch 2750, loss[loss=0.1852, simple_loss=0.2593, pruned_loss=0.05553, over 4822.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2709, pruned_loss=0.07398, over 955844.06 frames. ], batch size: 40, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:50:28,698 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.629e+02 1.991e+02 2.307e+02 4.303e+02, threshold=3.983e+02, percent-clipped=1.0 +2023-03-26 07:50:58,202 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37161.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:51:03,500 INFO [finetune.py:976] (0/7) Epoch 7, batch 2800, loss[loss=0.1581, simple_loss=0.2215, pruned_loss=0.04739, over 4800.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2669, pruned_loss=0.07211, over 956409.84 frames. ], batch size: 51, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:52:09,300 INFO [finetune.py:976] (0/7) Epoch 7, batch 2850, loss[loss=0.2014, simple_loss=0.2651, pruned_loss=0.06886, over 4764.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2658, pruned_loss=0.07212, over 955926.86 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:52:20,862 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.582e+02 1.929e+02 2.327e+02 4.539e+02, threshold=3.857e+02, percent-clipped=3.0 +2023-03-26 07:53:00,965 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37261.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:02,780 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37264.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:03,916 INFO [finetune.py:976] (0/7) Epoch 7, batch 2900, loss[loss=0.2502, simple_loss=0.3135, pruned_loss=0.09349, over 4894.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2676, pruned_loss=0.07244, over 956029.00 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:53:09,378 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37267.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:09,978 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:10,018 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:03,735 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.9764, 4.2703, 4.5655, 4.8469, 4.7460, 4.3880, 5.0905, 1.6557], + device='cuda:0'), covar=tensor([0.0676, 0.0778, 0.0755, 0.0806, 0.1072, 0.1535, 0.0537, 0.5333], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0275, 0.0295, 0.0332, 0.0285, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:54:04,930 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37309.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:06,755 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:07,986 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37314.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:13,846 INFO [finetune.py:976] (0/7) Epoch 7, batch 2950, loss[loss=0.1934, simple_loss=0.2638, pruned_loss=0.06148, over 4830.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2711, pruned_loss=0.07304, over 957935.48 frames. ], batch size: 33, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:54:13,904 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37316.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:17,027 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0187, 1.9215, 1.5679, 2.0618, 2.0319, 1.7226, 2.3801, 2.0556], + device='cuda:0'), covar=tensor([0.1730, 0.3086, 0.3844, 0.3189, 0.3048, 0.2067, 0.3763, 0.2243], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0190, 0.0234, 0.0255, 0.0233, 0.0192, 0.0211, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:54:25,441 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.702e+02 2.045e+02 2.514e+02 5.908e+02, threshold=4.090e+02, percent-clipped=3.0 +2023-03-26 07:54:27,398 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:59,207 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2663, 3.7101, 3.9528, 4.1401, 4.0318, 3.7757, 4.3533, 1.4738], + device='cuda:0'), covar=tensor([0.0645, 0.0715, 0.0778, 0.0820, 0.1029, 0.1339, 0.0582, 0.4576], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0276, 0.0295, 0.0333, 0.0285, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:55:00,455 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37356.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:55:05,692 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:55:08,106 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37364.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:55:09,239 INFO [finetune.py:976] (0/7) Epoch 7, batch 3000, loss[loss=0.2138, simple_loss=0.2668, pruned_loss=0.08044, over 4803.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2726, pruned_loss=0.074, over 957634.92 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:55:09,240 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 07:55:11,012 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3851, 1.5660, 1.5201, 1.6303, 1.6593, 3.0288, 1.4512, 1.6801], + device='cuda:0'), covar=tensor([0.0990, 0.1796, 0.1066, 0.0962, 0.1494, 0.0281, 0.1422, 0.1655], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0076, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 07:55:14,959 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8322, 3.4052, 3.5496, 3.7284, 3.5473, 3.3926, 3.8813, 1.3663], + device='cuda:0'), covar=tensor([0.0869, 0.0863, 0.0820, 0.0934, 0.1426, 0.1576, 0.0764, 0.4725], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0243, 0.0275, 0.0295, 0.0333, 0.0284, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:55:17,519 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8581, 1.3320, 0.9356, 1.6787, 2.0866, 1.1318, 1.4607, 1.7141], + device='cuda:0'), covar=tensor([0.1365, 0.1935, 0.1861, 0.1127, 0.1995, 0.2031, 0.1329, 0.1838], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0113, 0.0093, 0.0124, 0.0096, 0.0101, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 07:55:18,096 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6647, 1.5789, 2.0836, 1.4095, 1.8580, 1.8977, 1.5112, 2.0954], + device='cuda:0'), covar=tensor([0.1336, 0.2117, 0.1203, 0.1891, 0.0853, 0.1333, 0.2809, 0.0793], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0204, 0.0197, 0.0196, 0.0181, 0.0221, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:55:18,324 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7571, 1.6505, 2.1245, 1.4357, 1.8890, 1.9761, 1.5976, 2.1149], + device='cuda:0'), covar=tensor([0.1289, 0.2258, 0.1259, 0.1928, 0.0954, 0.1310, 0.2784, 0.0704], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0204, 0.0197, 0.0196, 0.0181, 0.0221, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 07:55:25,742 INFO [finetune.py:1010] (0/7) Epoch 7, validation: loss=0.161, simple_loss=0.2327, pruned_loss=0.04464, over 2265189.00 frames. +2023-03-26 07:55:25,743 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 07:55:51,417 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7176, 1.5688, 1.6546, 1.7204, 1.1884, 3.5907, 1.4168, 2.0630], + device='cuda:0'), covar=tensor([0.3403, 0.2489, 0.2054, 0.2287, 0.1856, 0.0175, 0.2462, 0.1227], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0118, 0.0122, 0.0116, 0.0098, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 07:55:54,503 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37404.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:56:02,292 INFO [finetune.py:976] (0/7) Epoch 7, batch 3050, loss[loss=0.2159, simple_loss=0.2833, pruned_loss=0.07431, over 4825.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2732, pruned_loss=0.07427, over 956036.08 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:56:11,535 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:56:12,025 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.581e+02 1.871e+02 2.387e+02 4.591e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-26 07:56:48,964 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37461.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:56:57,054 INFO [finetune.py:976] (0/7) Epoch 7, batch 3100, loss[loss=0.215, simple_loss=0.261, pruned_loss=0.08451, over 4929.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2722, pruned_loss=0.07401, over 956089.80 frames. ], batch size: 33, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:57:52,635 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37509.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:58:01,946 INFO [finetune.py:976] (0/7) Epoch 7, batch 3150, loss[loss=0.1996, simple_loss=0.2676, pruned_loss=0.06583, over 4902.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2705, pruned_loss=0.07337, over 955070.43 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:58:04,544 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-03-26 07:58:13,088 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.704e+02 2.041e+02 2.515e+02 5.799e+02, threshold=4.081e+02, percent-clipped=3.0 +2023-03-26 07:59:05,775 INFO [finetune.py:976] (0/7) Epoch 7, batch 3200, loss[loss=0.2191, simple_loss=0.2809, pruned_loss=0.07868, over 4856.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2668, pruned_loss=0.0719, over 953554.51 frames. ], batch size: 44, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:59:06,469 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:59:25,396 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8196, 1.1406, 1.7772, 1.6159, 1.5075, 1.4630, 1.5514, 1.5267], + device='cuda:0'), covar=tensor([0.4363, 0.5742, 0.4833, 0.5349, 0.6233, 0.5006, 0.6638, 0.4604], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0243, 0.0256, 0.0256, 0.0245, 0.0221, 0.0274, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:00:08,875 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37614.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:09,466 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37615.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:14,949 INFO [finetune.py:976] (0/7) Epoch 7, batch 3250, loss[loss=0.2039, simple_loss=0.257, pruned_loss=0.07536, over 4711.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2674, pruned_loss=0.07274, over 955007.13 frames. ], batch size: 23, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:00:19,897 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:26,116 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.664e+02 1.918e+02 2.274e+02 4.430e+02, threshold=3.836e+02, percent-clipped=1.0 +2023-03-26 08:00:59,741 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 08:01:09,573 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37660.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:01:10,702 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37662.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:01:18,272 INFO [finetune.py:976] (0/7) Epoch 7, batch 3300, loss[loss=0.2015, simple_loss=0.2506, pruned_loss=0.07616, over 4775.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2734, pruned_loss=0.07598, over 956969.81 frames. ], batch size: 23, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:01:52,595 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6688, 2.2487, 1.9005, 0.9353, 2.1319, 2.0807, 1.6976, 2.0945], + device='cuda:0'), covar=tensor([0.0719, 0.0897, 0.1318, 0.1955, 0.1231, 0.2113, 0.2208, 0.0857], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0199, 0.0187, 0.0217, 0.0205, 0.0220, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:02:12,450 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:02:22,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0591, 1.6632, 2.4699, 1.7376, 2.3082, 2.3289, 1.7494, 2.3582], + device='cuda:0'), covar=tensor([0.1407, 0.2232, 0.1380, 0.2025, 0.0993, 0.1403, 0.2930, 0.1017], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0203, 0.0197, 0.0195, 0.0180, 0.0219, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:02:22,966 INFO [finetune.py:976] (0/7) Epoch 7, batch 3350, loss[loss=0.1727, simple_loss=0.2416, pruned_loss=0.05187, over 4744.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2756, pruned_loss=0.07664, over 956230.59 frames. ], batch size: 27, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:02:25,456 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:02:34,110 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.767e+02 2.019e+02 2.457e+02 5.992e+02, threshold=4.038e+02, percent-clipped=4.0 +2023-03-26 08:02:45,334 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4216, 1.4110, 1.6640, 1.6951, 1.6545, 3.3071, 1.3711, 1.5709], + device='cuda:0'), covar=tensor([0.1080, 0.1889, 0.1153, 0.1054, 0.1631, 0.0313, 0.1570, 0.1827], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 08:03:16,427 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3342, 1.3609, 1.3596, 0.7648, 1.5434, 1.4380, 1.3807, 1.2631], + device='cuda:0'), covar=tensor([0.0585, 0.0682, 0.0723, 0.0968, 0.0695, 0.0688, 0.0648, 0.1140], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0133, 0.0143, 0.0126, 0.0112, 0.0144, 0.0146, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:03:28,096 INFO [finetune.py:976] (0/7) Epoch 7, batch 3400, loss[loss=0.2561, simple_loss=0.311, pruned_loss=0.1006, over 4726.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2789, pruned_loss=0.07847, over 955593.09 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:04:32,120 INFO [finetune.py:976] (0/7) Epoch 7, batch 3450, loss[loss=0.2128, simple_loss=0.2806, pruned_loss=0.07248, over 4863.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2778, pruned_loss=0.07715, over 954265.60 frames. ], batch size: 34, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:04:43,341 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.715e+02 1.991e+02 2.496e+02 6.747e+02, threshold=3.982e+02, percent-clipped=3.0 +2023-03-26 08:05:36,354 INFO [finetune.py:976] (0/7) Epoch 7, batch 3500, loss[loss=0.1795, simple_loss=0.2396, pruned_loss=0.05967, over 4827.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2744, pruned_loss=0.07604, over 951863.66 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:05:45,662 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 08:05:54,124 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9558, 1.7356, 2.2843, 1.6963, 2.1471, 2.2160, 1.7553, 2.2970], + device='cuda:0'), covar=tensor([0.1018, 0.1541, 0.1228, 0.1492, 0.0677, 0.1049, 0.2020, 0.0575], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0204, 0.0198, 0.0195, 0.0181, 0.0221, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:06:28,204 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1232, 1.9510, 1.6206, 1.9318, 2.1161, 1.7659, 2.3870, 2.0554], + device='cuda:0'), covar=tensor([0.1543, 0.2829, 0.3686, 0.3284, 0.2776, 0.1893, 0.3797, 0.2144], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0190, 0.0235, 0.0255, 0.0234, 0.0193, 0.0212, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:06:40,844 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 08:06:41,166 INFO [finetune.py:976] (0/7) Epoch 7, batch 3550, loss[loss=0.2043, simple_loss=0.2687, pruned_loss=0.06993, over 4797.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2705, pruned_loss=0.07437, over 952103.23 frames. ], batch size: 25, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:06:51,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:06:58,533 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.559e+02 1.846e+02 2.185e+02 4.242e+02, threshold=3.693e+02, percent-clipped=1.0 +2023-03-26 08:07:01,290 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 08:07:52,117 INFO [finetune.py:976] (0/7) Epoch 7, batch 3600, loss[loss=0.2122, simple_loss=0.2732, pruned_loss=0.07559, over 4840.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.268, pruned_loss=0.07308, over 954438.67 frames. ], batch size: 47, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:07:56,327 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:08:06,512 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37979.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:08:35,841 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-38000.pt +2023-03-26 08:08:59,065 INFO [finetune.py:976] (0/7) Epoch 7, batch 3650, loss[loss=0.2998, simple_loss=0.3486, pruned_loss=0.1254, over 4747.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.271, pruned_loss=0.07438, over 954397.16 frames. ], batch size: 59, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:09:07,213 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:09:10,802 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.699e+02 2.068e+02 2.418e+02 4.148e+02, threshold=4.136e+02, percent-clipped=4.0 +2023-03-26 08:09:20,545 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3408, 1.9410, 1.7145, 0.8524, 1.9645, 1.7793, 1.4225, 1.8782], + device='cuda:0'), covar=tensor([0.1057, 0.1094, 0.1905, 0.2648, 0.1648, 0.2434, 0.2856, 0.1354], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0199, 0.0188, 0.0218, 0.0206, 0.0221, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:09:27,926 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:09:46,799 INFO [finetune.py:976] (0/7) Epoch 7, batch 3700, loss[loss=0.1754, simple_loss=0.2466, pruned_loss=0.05207, over 4762.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.2732, pruned_loss=0.07478, over 953340.84 frames. ], batch size: 27, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:09:48,564 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38068.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:10:19,924 INFO [finetune.py:976] (0/7) Epoch 7, batch 3750, loss[loss=0.2358, simple_loss=0.2913, pruned_loss=0.09014, over 4917.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2748, pruned_loss=0.07551, over 954712.23 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:10:26,924 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.737e+01 1.627e+02 1.982e+02 2.503e+02 4.763e+02, threshold=3.965e+02, percent-clipped=1.0 +2023-03-26 08:10:57,173 INFO [finetune.py:976] (0/7) Epoch 7, batch 3800, loss[loss=0.1968, simple_loss=0.2687, pruned_loss=0.06249, over 4785.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2766, pruned_loss=0.07623, over 953904.90 frames. ], batch size: 51, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:11:28,188 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 08:11:30,369 INFO [finetune.py:976] (0/7) Epoch 7, batch 3850, loss[loss=0.195, simple_loss=0.2684, pruned_loss=0.06076, over 4716.00 frames. ], tot_loss[loss=0.2115, simple_loss=0.2737, pruned_loss=0.07468, over 950317.80 frames. ], batch size: 59, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:11:43,043 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.610e+02 2.090e+02 2.406e+02 4.877e+02, threshold=4.181e+02, percent-clipped=2.0 +2023-03-26 08:11:43,302 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 08:12:25,293 INFO [finetune.py:976] (0/7) Epoch 7, batch 3900, loss[loss=0.194, simple_loss=0.2498, pruned_loss=0.06911, over 4908.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.271, pruned_loss=0.07432, over 952276.27 frames. ], batch size: 46, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:13:28,060 INFO [finetune.py:976] (0/7) Epoch 7, batch 3950, loss[loss=0.215, simple_loss=0.2804, pruned_loss=0.07476, over 4922.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2678, pruned_loss=0.07292, over 953547.91 frames. ], batch size: 37, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:13:45,513 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.693e+02 1.988e+02 2.374e+02 4.679e+02, threshold=3.976e+02, percent-clipped=1.0 +2023-03-26 08:13:56,358 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38335.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:14:21,111 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9210, 1.7039, 1.6040, 1.9385, 2.5681, 1.9236, 1.7489, 1.4277], + device='cuda:0'), covar=tensor([0.2607, 0.2537, 0.2295, 0.2012, 0.2098, 0.1416, 0.2692, 0.2194], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0210, 0.0205, 0.0188, 0.0240, 0.0179, 0.0215, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:14:21,705 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:14:25,157 INFO [finetune.py:976] (0/7) Epoch 7, batch 4000, loss[loss=0.2103, simple_loss=0.2726, pruned_loss=0.07401, over 4902.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2673, pruned_loss=0.07295, over 954347.39 frames. ], batch size: 35, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:15:29,792 INFO [finetune.py:976] (0/7) Epoch 7, batch 4050, loss[loss=0.2181, simple_loss=0.2748, pruned_loss=0.08068, over 4937.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2717, pruned_loss=0.07519, over 954900.03 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:15:33,981 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:15:42,068 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.778e+02 2.129e+02 2.625e+02 5.238e+02, threshold=4.258e+02, percent-clipped=5.0 +2023-03-26 08:16:32,427 INFO [finetune.py:976] (0/7) Epoch 7, batch 4100, loss[loss=0.1959, simple_loss=0.2729, pruned_loss=0.05943, over 4753.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2732, pruned_loss=0.07541, over 955151.74 frames. ], batch size: 54, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:16:51,229 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2938, 2.8953, 3.0239, 3.2295, 3.0356, 2.9046, 3.3109, 0.9818], + device='cuda:0'), covar=tensor([0.0966, 0.0925, 0.0925, 0.0947, 0.1519, 0.1563, 0.1072, 0.4824], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0275, 0.0295, 0.0333, 0.0284, 0.0305, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:17:30,624 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-26 08:17:31,590 INFO [finetune.py:976] (0/7) Epoch 7, batch 4150, loss[loss=0.2367, simple_loss=0.2879, pruned_loss=0.09278, over 4854.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2746, pruned_loss=0.07571, over 956668.21 frames. ], batch size: 31, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:17:43,680 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.723e+02 2.145e+02 2.598e+02 6.605e+02, threshold=4.291e+02, percent-clipped=2.0 +2023-03-26 08:18:34,212 INFO [finetune.py:976] (0/7) Epoch 7, batch 4200, loss[loss=0.2084, simple_loss=0.2795, pruned_loss=0.06869, over 4806.00 frames. ], tot_loss[loss=0.213, simple_loss=0.2752, pruned_loss=0.07542, over 956152.89 frames. ], batch size: 41, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:19:34,109 INFO [finetune.py:976] (0/7) Epoch 7, batch 4250, loss[loss=0.1922, simple_loss=0.2633, pruned_loss=0.06055, over 4904.00 frames. ], tot_loss[loss=0.2115, simple_loss=0.2736, pruned_loss=0.07472, over 957002.25 frames. ], batch size: 46, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:19:44,841 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.606e+02 1.980e+02 2.259e+02 5.740e+02, threshold=3.960e+02, percent-clipped=2.0 +2023-03-26 08:20:02,306 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:20:38,695 INFO [finetune.py:976] (0/7) Epoch 7, batch 4300, loss[loss=0.212, simple_loss=0.2717, pruned_loss=0.07614, over 4823.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2695, pruned_loss=0.07308, over 957185.25 frames. ], batch size: 41, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:20:59,375 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38683.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:21:12,971 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4821, 2.2448, 1.8789, 2.4983, 2.4748, 2.0407, 2.9169, 2.4096], + device='cuda:0'), covar=tensor([0.1629, 0.3482, 0.3885, 0.3609, 0.2895, 0.1904, 0.3926, 0.2223], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0190, 0.0235, 0.0253, 0.0233, 0.0193, 0.0212, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:21:34,983 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8244, 1.6182, 1.6353, 1.8430, 2.5022, 1.8613, 1.7822, 1.4579], + device='cuda:0'), covar=tensor([0.2455, 0.2614, 0.2178, 0.1950, 0.1973, 0.1528, 0.2603, 0.2196], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0209, 0.0204, 0.0186, 0.0239, 0.0177, 0.0214, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:21:41,238 INFO [finetune.py:976] (0/7) Epoch 7, batch 4350, loss[loss=0.2182, simple_loss=0.2711, pruned_loss=0.08262, over 4914.00 frames. ], tot_loss[loss=0.2038, simple_loss=0.2656, pruned_loss=0.07096, over 958867.46 frames. ], batch size: 36, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:21:41,305 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38716.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:21:51,837 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0092, 4.5974, 4.3520, 2.6598, 4.6181, 3.5095, 0.8237, 3.2087], + device='cuda:0'), covar=tensor([0.2383, 0.2154, 0.1377, 0.2930, 0.0912, 0.0912, 0.4752, 0.1523], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0174, 0.0162, 0.0129, 0.0154, 0.0123, 0.0146, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 08:21:52,343 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.679e+02 1.871e+02 2.197e+02 5.866e+02, threshold=3.741e+02, percent-clipped=4.0 +2023-03-26 08:21:52,468 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7196, 1.6781, 1.6218, 1.1115, 1.8407, 1.7295, 1.7757, 1.4619], + device='cuda:0'), covar=tensor([0.0579, 0.0664, 0.0800, 0.0949, 0.0585, 0.0813, 0.0619, 0.1106], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0135, 0.0144, 0.0127, 0.0114, 0.0147, 0.0147, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:22:02,908 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2478, 1.4092, 1.6576, 1.1029, 1.2936, 1.3713, 1.3619, 1.5980], + device='cuda:0'), covar=tensor([0.1405, 0.2205, 0.1211, 0.1563, 0.1045, 0.1427, 0.2828, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0204, 0.0198, 0.0195, 0.0182, 0.0221, 0.0216, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:22:43,656 INFO [finetune.py:976] (0/7) Epoch 7, batch 4400, loss[loss=0.1971, simple_loss=0.2644, pruned_loss=0.06487, over 4791.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2672, pruned_loss=0.07191, over 958727.48 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:23:34,450 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 08:23:42,200 INFO [finetune.py:976] (0/7) Epoch 7, batch 4450, loss[loss=0.2701, simple_loss=0.3271, pruned_loss=0.1065, over 4766.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2712, pruned_loss=0.07383, over 956923.66 frames. ], batch size: 54, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:23:51,919 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:23:53,601 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.712e+02 1.965e+02 2.330e+02 4.727e+02, threshold=3.929e+02, percent-clipped=4.0 +2023-03-26 08:24:44,783 INFO [finetune.py:976] (0/7) Epoch 7, batch 4500, loss[loss=0.2747, simple_loss=0.3291, pruned_loss=0.1102, over 4812.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2733, pruned_loss=0.07456, over 956504.63 frames. ], batch size: 41, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:24:57,925 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 08:25:06,330 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:25:08,167 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3322, 1.4509, 1.3941, 1.5694, 1.5344, 3.0550, 1.2912, 1.5532], + device='cuda:0'), covar=tensor([0.0988, 0.1758, 0.1189, 0.1019, 0.1686, 0.0303, 0.1545, 0.1747], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0083, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 08:25:49,098 INFO [finetune.py:976] (0/7) Epoch 7, batch 4550, loss[loss=0.2304, simple_loss=0.2789, pruned_loss=0.09093, over 4833.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2755, pruned_loss=0.07512, over 957116.91 frames. ], batch size: 47, lr: 3.86e-03, grad_scale: 64.0 +2023-03-26 08:25:59,509 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.671e+02 2.000e+02 2.524e+02 3.434e+02, threshold=4.000e+02, percent-clipped=0.0 +2023-03-26 08:26:05,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1549, 1.3086, 1.0818, 1.3032, 1.3897, 2.5177, 1.2074, 1.4526], + device='cuda:0'), covar=tensor([0.1036, 0.1954, 0.1232, 0.1033, 0.1808, 0.0341, 0.1608, 0.1788], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 08:26:13,063 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7544, 1.5680, 1.5960, 1.7799, 1.2941, 3.7593, 1.3641, 2.1191], + device='cuda:0'), covar=tensor([0.3247, 0.2436, 0.2073, 0.2163, 0.1698, 0.0149, 0.2657, 0.1246], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0123, 0.0117, 0.0098, 0.0101, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 08:26:47,278 INFO [finetune.py:976] (0/7) Epoch 7, batch 4600, loss[loss=0.1621, simple_loss=0.235, pruned_loss=0.04465, over 4760.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2727, pruned_loss=0.07282, over 957299.85 frames. ], batch size: 28, lr: 3.86e-03, grad_scale: 64.0 +2023-03-26 08:27:15,234 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.53 vs. limit=5.0 +2023-03-26 08:27:54,985 INFO [finetune.py:976] (0/7) Epoch 7, batch 4650, loss[loss=0.1656, simple_loss=0.2287, pruned_loss=0.05123, over 4711.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2693, pruned_loss=0.07171, over 954484.62 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:27:55,088 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:28:06,170 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.761e+01 1.505e+02 1.924e+02 2.345e+02 4.238e+02, threshold=3.847e+02, percent-clipped=2.0 +2023-03-26 08:28:07,542 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9537, 1.2517, 1.7062, 1.6990, 1.5429, 1.5647, 1.6568, 1.6523], + device='cuda:0'), covar=tensor([0.6328, 0.7505, 0.6985, 0.6825, 0.8571, 0.6465, 0.8713, 0.6190], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0242, 0.0254, 0.0254, 0.0244, 0.0220, 0.0272, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:28:50,938 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:28:57,994 INFO [finetune.py:976] (0/7) Epoch 7, batch 4700, loss[loss=0.1693, simple_loss=0.2395, pruned_loss=0.04952, over 4789.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2663, pruned_loss=0.07094, over 954848.14 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:29:09,885 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6703, 3.9789, 4.2096, 4.4411, 4.3925, 4.1397, 4.7129, 1.4906], + device='cuda:0'), covar=tensor([0.0598, 0.0789, 0.0665, 0.0691, 0.1015, 0.1248, 0.0579, 0.4995], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0274, 0.0292, 0.0331, 0.0282, 0.0302, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:29:56,018 INFO [finetune.py:976] (0/7) Epoch 7, batch 4750, loss[loss=0.173, simple_loss=0.236, pruned_loss=0.05504, over 4778.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2642, pruned_loss=0.07074, over 955679.12 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:30:08,811 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.609e+02 1.819e+02 2.206e+02 4.512e+02, threshold=3.638e+02, percent-clipped=2.0 +2023-03-26 08:30:17,944 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6526, 1.4862, 1.5289, 1.6259, 1.1101, 3.3914, 1.2890, 1.7685], + device='cuda:0'), covar=tensor([0.3349, 0.2421, 0.2092, 0.2352, 0.1925, 0.0199, 0.2723, 0.1416], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0119, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 08:30:38,542 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7163, 1.5075, 1.5559, 1.6172, 1.1282, 3.5387, 1.3423, 1.8787], + device='cuda:0'), covar=tensor([0.3084, 0.2320, 0.1984, 0.2111, 0.1643, 0.0171, 0.2520, 0.1298], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 08:30:58,729 INFO [finetune.py:976] (0/7) Epoch 7, batch 4800, loss[loss=0.2461, simple_loss=0.3021, pruned_loss=0.09503, over 4776.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2684, pruned_loss=0.07308, over 953731.56 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:31:12,881 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:31:22,018 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39185.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:31:57,083 INFO [finetune.py:976] (0/7) Epoch 7, batch 4850, loss[loss=0.2105, simple_loss=0.2757, pruned_loss=0.07261, over 4836.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2721, pruned_loss=0.07423, over 955731.46 frames. ], batch size: 47, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:32:02,622 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7059, 1.8999, 1.5137, 1.6284, 2.0950, 1.9309, 1.8483, 1.7380], + device='cuda:0'), covar=tensor([0.0376, 0.0368, 0.0522, 0.0294, 0.0279, 0.0816, 0.0390, 0.0382], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0112, 0.0140, 0.0116, 0.0104, 0.0101, 0.0092, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.9501e-05, 8.7717e-05, 1.1228e-04, 9.1085e-05, 8.2277e-05, 7.5203e-05, + 6.9292e-05, 8.5743e-05], device='cuda:0') +2023-03-26 08:32:06,049 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.691e+02 2.004e+02 2.499e+02 4.240e+02, threshold=4.008e+02, percent-clipped=2.0 +2023-03-26 08:32:18,164 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:32:30,567 INFO [finetune.py:976] (0/7) Epoch 7, batch 4900, loss[loss=0.1866, simple_loss=0.2477, pruned_loss=0.06276, over 4716.00 frames. ], tot_loss[loss=0.213, simple_loss=0.275, pruned_loss=0.07553, over 954007.14 frames. ], batch size: 23, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:32:43,211 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:32:52,747 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 08:33:03,517 INFO [finetune.py:976] (0/7) Epoch 7, batch 4950, loss[loss=0.2448, simple_loss=0.2999, pruned_loss=0.09485, over 4797.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2746, pruned_loss=0.07487, over 954527.64 frames. ], batch size: 45, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:33:12,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.619e+02 1.980e+02 2.423e+02 3.796e+02, threshold=3.961e+02, percent-clipped=0.0 +2023-03-26 08:33:24,236 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:33:37,224 INFO [finetune.py:976] (0/7) Epoch 7, batch 5000, loss[loss=0.1904, simple_loss=0.2503, pruned_loss=0.06528, over 4929.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2717, pruned_loss=0.07311, over 952015.24 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:34:10,935 INFO [finetune.py:976] (0/7) Epoch 7, batch 5050, loss[loss=0.2141, simple_loss=0.2555, pruned_loss=0.08634, over 4255.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2681, pruned_loss=0.07161, over 953738.02 frames. ], batch size: 65, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:34:16,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9094, 2.6310, 2.4206, 3.1082, 2.7214, 2.6736, 2.6208, 3.6524], + device='cuda:0'), covar=tensor([0.4460, 0.5866, 0.4058, 0.4474, 0.4617, 0.2818, 0.4733, 0.1704], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0220, 0.0281, 0.0241, 0.0207, 0.0246, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:34:19,592 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.623e+02 1.955e+02 2.404e+02 3.498e+02, threshold=3.910e+02, percent-clipped=0.0 +2023-03-26 08:34:24,576 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 08:34:52,730 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:34:54,406 INFO [finetune.py:976] (0/7) Epoch 7, batch 5100, loss[loss=0.2305, simple_loss=0.2873, pruned_loss=0.08688, over 4815.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2646, pruned_loss=0.07042, over 954523.22 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:35:04,715 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1763, 4.9720, 4.6722, 2.9009, 5.0466, 3.7921, 0.9916, 3.5302], + device='cuda:0'), covar=tensor([0.2257, 0.1786, 0.1410, 0.3042, 0.0807, 0.0869, 0.5243, 0.1411], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0172, 0.0162, 0.0129, 0.0155, 0.0123, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 08:35:05,361 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:32,525 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4347, 1.4559, 1.2095, 1.4630, 1.7289, 1.6132, 1.4625, 1.2415], + device='cuda:0'), covar=tensor([0.0328, 0.0332, 0.0575, 0.0285, 0.0224, 0.0438, 0.0306, 0.0416], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0111, 0.0140, 0.0116, 0.0104, 0.0101, 0.0091, 0.0110], + device='cuda:0'), out_proj_covar=tensor([6.9358e-05, 8.7539e-05, 1.1161e-04, 9.1366e-05, 8.1587e-05, 7.4974e-05, + 6.8587e-05, 8.5320e-05], device='cuda:0') +2023-03-26 08:35:38,645 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-03-26 08:35:39,560 INFO [finetune.py:976] (0/7) Epoch 7, batch 5150, loss[loss=0.2143, simple_loss=0.282, pruned_loss=0.0733, over 4809.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2651, pruned_loss=0.07088, over 950443.43 frames. ], batch size: 41, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:35:47,013 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:48,750 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:49,815 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.405e+01 1.710e+02 2.016e+02 2.412e+02 5.054e+02, threshold=4.032e+02, percent-clipped=2.0 +2023-03-26 08:35:57,882 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0388, 1.6776, 2.5850, 1.3823, 2.1931, 2.2546, 1.6476, 2.3689], + device='cuda:0'), covar=tensor([0.1694, 0.2229, 0.1506, 0.2576, 0.1154, 0.1837, 0.2841, 0.1242], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0205, 0.0199, 0.0196, 0.0182, 0.0222, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:36:08,539 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39541.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:36:24,641 INFO [finetune.py:976] (0/7) Epoch 7, batch 5200, loss[loss=0.2132, simple_loss=0.2772, pruned_loss=0.07461, over 4794.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2663, pruned_loss=0.07117, over 946992.44 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:36:48,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7440, 3.9028, 3.7138, 1.8677, 3.9447, 3.0773, 0.7270, 2.6132], + device='cuda:0'), covar=tensor([0.2474, 0.1907, 0.1587, 0.3303, 0.1056, 0.1010, 0.4801, 0.1674], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0171, 0.0160, 0.0128, 0.0154, 0.0122, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 08:36:59,642 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6102, 2.3934, 1.9287, 0.9927, 2.0672, 1.9728, 1.9145, 2.1006], + device='cuda:0'), covar=tensor([0.0860, 0.0860, 0.1839, 0.2313, 0.1593, 0.2479, 0.2160, 0.1096], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0200, 0.0201, 0.0187, 0.0217, 0.0206, 0.0222, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:37:07,830 INFO [finetune.py:976] (0/7) Epoch 7, batch 5250, loss[loss=0.2205, simple_loss=0.2646, pruned_loss=0.08815, over 4444.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2693, pruned_loss=0.07255, over 947237.79 frames. ], batch size: 19, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:37:15,021 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.709e+02 2.070e+02 2.577e+02 5.953e+02, threshold=4.140e+02, percent-clipped=1.0 +2023-03-26 08:37:25,980 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39639.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 08:37:26,590 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:37:43,681 INFO [finetune.py:976] (0/7) Epoch 7, batch 5300, loss[loss=0.2074, simple_loss=0.2647, pruned_loss=0.07508, over 4314.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2718, pruned_loss=0.07399, over 947879.06 frames. ], batch size: 65, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:38:17,441 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39700.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 08:38:32,613 INFO [finetune.py:976] (0/7) Epoch 7, batch 5350, loss[loss=0.2371, simple_loss=0.2809, pruned_loss=0.0966, over 4793.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2714, pruned_loss=0.07288, over 949833.98 frames. ], batch size: 29, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:38:40,830 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.565e+02 1.855e+02 2.323e+02 5.491e+02, threshold=3.710e+02, percent-clipped=1.0 +2023-03-26 08:39:15,939 INFO [finetune.py:976] (0/7) Epoch 7, batch 5400, loss[loss=0.1723, simple_loss=0.2269, pruned_loss=0.05882, over 4744.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2691, pruned_loss=0.07205, over 950909.41 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:39:16,642 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39767.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:39:50,827 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39815.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:39:51,311 INFO [finetune.py:976] (0/7) Epoch 7, batch 5450, loss[loss=0.1904, simple_loss=0.252, pruned_loss=0.0644, over 4790.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2665, pruned_loss=0.07115, over 952765.51 frames. ], batch size: 29, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:39:53,200 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:03,634 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.339e+01 1.599e+02 1.928e+02 2.299e+02 3.698e+02, threshold=3.856e+02, percent-clipped=0.0 +2023-03-26 08:40:03,759 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39828.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:16,807 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39841.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:54,098 INFO [finetune.py:976] (0/7) Epoch 7, batch 5500, loss[loss=0.1438, simple_loss=0.2167, pruned_loss=0.03546, over 4819.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2629, pruned_loss=0.06975, over 953825.65 frames. ], batch size: 25, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:41:05,350 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39876.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:41:18,432 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39889.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:41:57,092 INFO [finetune.py:976] (0/7) Epoch 7, batch 5550, loss[loss=0.1686, simple_loss=0.2326, pruned_loss=0.05233, over 4777.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2671, pruned_loss=0.07223, over 954514.85 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:42:09,838 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.647e+02 1.995e+02 2.278e+02 3.177e+02, threshold=3.991e+02, percent-clipped=0.0 +2023-03-26 08:42:28,005 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:42:58,490 INFO [finetune.py:976] (0/7) Epoch 7, batch 5600, loss[loss=0.2511, simple_loss=0.3088, pruned_loss=0.09669, over 4815.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2728, pruned_loss=0.07435, over 954746.90 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:43:20,784 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:43:29,077 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7366, 1.5402, 1.3776, 1.5770, 1.9429, 1.9018, 1.7322, 1.4246], + device='cuda:0'), covar=tensor([0.0297, 0.0335, 0.0548, 0.0315, 0.0222, 0.0365, 0.0264, 0.0400], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0112, 0.0139, 0.0115, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.9489e-05, 8.7956e-05, 1.1126e-04, 9.0751e-05, 8.1925e-05, 7.4955e-05, + 6.8941e-05, 8.4879e-05], device='cuda:0') +2023-03-26 08:43:30,215 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39995.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 08:43:33,211 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-40000.pt +2023-03-26 08:43:42,196 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3068, 2.9120, 3.0472, 3.2582, 3.0923, 2.8775, 3.3514, 0.9752], + device='cuda:0'), covar=tensor([0.1194, 0.1012, 0.1117, 0.1301, 0.1737, 0.1886, 0.1143, 0.5316], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0276, 0.0294, 0.0333, 0.0282, 0.0305, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:43:52,203 INFO [finetune.py:976] (0/7) Epoch 7, batch 5650, loss[loss=0.1989, simple_loss=0.266, pruned_loss=0.06585, over 4750.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.2745, pruned_loss=0.07416, over 955057.10 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:44:09,061 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.647e+02 1.995e+02 2.469e+02 4.643e+02, threshold=3.989e+02, percent-clipped=3.0 +2023-03-26 08:44:50,697 INFO [finetune.py:976] (0/7) Epoch 7, batch 5700, loss[loss=0.2322, simple_loss=0.2745, pruned_loss=0.09499, over 4131.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2693, pruned_loss=0.07279, over 937873.30 frames. ], batch size: 18, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:45:23,201 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-7.pt +2023-03-26 08:45:42,073 INFO [finetune.py:976] (0/7) Epoch 8, batch 0, loss[loss=0.2426, simple_loss=0.2954, pruned_loss=0.09487, over 4915.00 frames. ], tot_loss[loss=0.2426, simple_loss=0.2954, pruned_loss=0.09487, over 4915.00 frames. ], batch size: 41, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:45:42,074 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 08:45:54,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6103, 1.3142, 1.4373, 1.4374, 1.7126, 1.6714, 1.5577, 1.3939], + device='cuda:0'), covar=tensor([0.0347, 0.0319, 0.0554, 0.0278, 0.0337, 0.0388, 0.0336, 0.0433], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0113, 0.0140, 0.0116, 0.0105, 0.0101, 0.0092, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.0022e-05, 8.8500e-05, 1.1228e-04, 9.1282e-05, 8.2386e-05, 7.5241e-05, + 6.9340e-05, 8.5588e-05], device='cuda:0') +2023-03-26 08:45:57,866 INFO [finetune.py:1010] (0/7) Epoch 8, validation: loss=0.1624, simple_loss=0.234, pruned_loss=0.04544, over 2265189.00 frames. +2023-03-26 08:45:57,866 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 08:46:09,356 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 08:46:20,338 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:46:26,527 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40123.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:46:29,510 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.580e+02 2.018e+02 2.508e+02 5.130e+02, threshold=4.036e+02, percent-clipped=1.0 +2023-03-26 08:46:41,216 INFO [finetune.py:976] (0/7) Epoch 8, batch 50, loss[loss=0.2116, simple_loss=0.2535, pruned_loss=0.08489, over 4738.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.2779, pruned_loss=0.07766, over 217739.00 frames. ], batch size: 54, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:46:42,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5192, 1.3560, 1.3390, 1.5801, 1.7226, 1.5615, 0.9093, 1.3102], + device='cuda:0'), covar=tensor([0.2252, 0.2186, 0.1900, 0.1739, 0.1606, 0.1278, 0.2675, 0.1931], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0209, 0.0204, 0.0188, 0.0239, 0.0178, 0.0214, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:47:04,281 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:10,680 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40171.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:13,925 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 08:47:26,491 INFO [finetune.py:976] (0/7) Epoch 8, batch 100, loss[loss=0.2138, simple_loss=0.2771, pruned_loss=0.07526, over 4819.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2692, pruned_loss=0.07327, over 382881.67 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:47:27,221 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:39,835 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3768, 1.2317, 1.2147, 1.3334, 1.5336, 1.4302, 1.3227, 1.1483], + device='cuda:0'), covar=tensor([0.0314, 0.0254, 0.0550, 0.0257, 0.0219, 0.0559, 0.0277, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0112, 0.0140, 0.0115, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.9898e-05, 8.7818e-05, 1.1161e-04, 9.0712e-05, 8.2141e-05, 7.4776e-05, + 6.9050e-05, 8.4917e-05], device='cuda:0') +2023-03-26 08:47:42,293 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40218.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:48,308 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.570e+02 1.832e+02 2.394e+02 3.868e+02, threshold=3.663e+02, percent-clipped=0.0 +2023-03-26 08:47:58,790 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8427, 1.2099, 0.8311, 1.7414, 2.0995, 1.5004, 1.5659, 1.7691], + device='cuda:0'), covar=tensor([0.1474, 0.2211, 0.2231, 0.1214, 0.2039, 0.1991, 0.1498, 0.2016], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0092, 0.0123, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 08:47:59,301 INFO [finetune.py:976] (0/7) Epoch 8, batch 150, loss[loss=0.1819, simple_loss=0.2495, pruned_loss=0.05716, over 4905.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2603, pruned_loss=0.06986, over 509877.99 frames. ], batch size: 37, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:48:07,718 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40256.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:18,389 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:19,035 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4495, 2.1698, 2.8513, 1.6989, 2.6741, 2.6493, 2.0712, 2.8326], + device='cuda:0'), covar=tensor([0.1682, 0.2066, 0.1581, 0.2428, 0.1018, 0.1752, 0.2813, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0206, 0.0200, 0.0198, 0.0183, 0.0223, 0.0220, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:48:22,639 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40279.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:33,049 INFO [finetune.py:976] (0/7) Epoch 8, batch 200, loss[loss=0.2059, simple_loss=0.2703, pruned_loss=0.07078, over 4843.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2617, pruned_loss=0.07172, over 610040.56 frames. ], batch size: 44, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:48:33,740 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40295.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 08:48:55,738 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.652e+02 1.957e+02 2.371e+02 3.958e+02, threshold=3.914e+02, percent-clipped=3.0 +2023-03-26 08:48:57,103 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:58,974 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:49:05,956 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40343.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 08:49:06,465 INFO [finetune.py:976] (0/7) Epoch 8, batch 250, loss[loss=0.1767, simple_loss=0.24, pruned_loss=0.05674, over 4771.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2651, pruned_loss=0.07145, over 688967.36 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:49:07,293 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 08:49:08,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.9013, 4.2165, 4.3984, 4.6982, 4.5656, 4.3787, 4.9723, 1.5310], + device='cuda:0'), covar=tensor([0.0762, 0.0861, 0.0712, 0.0910, 0.1390, 0.1475, 0.0559, 0.5788], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0242, 0.0274, 0.0292, 0.0331, 0.0280, 0.0303, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:49:37,956 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40391.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:49:40,040 INFO [finetune.py:976] (0/7) Epoch 8, batch 300, loss[loss=0.2065, simple_loss=0.2747, pruned_loss=0.06913, over 4858.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2667, pruned_loss=0.0713, over 747567.06 frames. ], batch size: 34, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:49:59,866 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:50:07,993 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.680e+02 2.022e+02 2.440e+02 4.521e+02, threshold=4.043e+02, percent-clipped=1.0 +2023-03-26 08:50:10,546 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 08:50:27,979 INFO [finetune.py:976] (0/7) Epoch 8, batch 350, loss[loss=0.1661, simple_loss=0.233, pruned_loss=0.04963, over 4826.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2682, pruned_loss=0.07165, over 793593.24 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:51:01,402 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:01,450 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:26,223 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:27,362 INFO [finetune.py:976] (0/7) Epoch 8, batch 400, loss[loss=0.2018, simple_loss=0.2738, pruned_loss=0.06489, over 4758.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.269, pruned_loss=0.07157, over 827960.06 frames. ], batch size: 28, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:51:52,935 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40519.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:58,841 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.662e+02 2.008e+02 2.590e+02 4.107e+02, threshold=4.016e+02, percent-clipped=2.0 +2023-03-26 08:52:11,108 INFO [finetune.py:976] (0/7) Epoch 8, batch 450, loss[loss=0.1849, simple_loss=0.2424, pruned_loss=0.06373, over 4815.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2687, pruned_loss=0.07189, over 854867.70 frames. ], batch size: 39, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:52:21,158 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:21,916 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 08:52:22,480 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:38,695 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4187, 2.3762, 1.8381, 2.3791, 2.3118, 2.0321, 2.8744, 2.5242], + device='cuda:0'), covar=tensor([0.1495, 0.2667, 0.3526, 0.3367, 0.2925, 0.1841, 0.4462, 0.2070], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0189, 0.0235, 0.0255, 0.0235, 0.0193, 0.0211, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:52:41,172 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40574.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:54,278 INFO [finetune.py:976] (0/7) Epoch 8, batch 500, loss[loss=0.1901, simple_loss=0.2511, pruned_loss=0.06449, over 4826.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2658, pruned_loss=0.07051, over 879368.49 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:53:17,852 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.648e+02 1.946e+02 2.379e+02 4.476e+02, threshold=3.892e+02, percent-clipped=1.0 +2023-03-26 08:53:17,930 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40628.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:53:28,115 INFO [finetune.py:976] (0/7) Epoch 8, batch 550, loss[loss=0.172, simple_loss=0.2437, pruned_loss=0.05014, over 4789.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2622, pruned_loss=0.06901, over 894907.73 frames. ], batch size: 29, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:53:32,505 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:53:56,753 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:54:01,538 INFO [finetune.py:976] (0/7) Epoch 8, batch 600, loss[loss=0.1954, simple_loss=0.2528, pruned_loss=0.06903, over 4822.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.264, pruned_loss=0.07052, over 909810.90 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:54:14,585 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40712.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:54:15,273 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 08:54:24,591 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.757e+02 2.080e+02 2.524e+02 4.426e+02, threshold=4.160e+02, percent-clipped=1.0 +2023-03-26 08:54:31,197 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8543, 2.1361, 1.5384, 1.6803, 2.2124, 2.0453, 1.9914, 1.7618], + device='cuda:0'), covar=tensor([0.0311, 0.0241, 0.0523, 0.0285, 0.0242, 0.0605, 0.0277, 0.0343], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0111, 0.0140, 0.0116, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([6.9993e-05, 8.7293e-05, 1.1184e-04, 9.1341e-05, 8.1936e-05, 7.4972e-05, + 6.8618e-05, 8.4793e-05], device='cuda:0') +2023-03-26 08:54:34,712 INFO [finetune.py:976] (0/7) Epoch 8, batch 650, loss[loss=0.1979, simple_loss=0.2589, pruned_loss=0.06841, over 4760.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2688, pruned_loss=0.07225, over 920473.14 frames. ], batch size: 27, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:54:40,856 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 08:54:51,305 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5397, 1.4310, 1.3586, 1.5782, 1.5675, 1.6329, 0.8853, 1.3345], + device='cuda:0'), covar=tensor([0.2212, 0.2043, 0.1871, 0.1710, 0.1696, 0.1212, 0.2741, 0.1880], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0209, 0.0205, 0.0188, 0.0240, 0.0178, 0.0215, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:55:08,422 INFO [finetune.py:976] (0/7) Epoch 8, batch 700, loss[loss=0.1947, simple_loss=0.2674, pruned_loss=0.06098, over 4848.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2712, pruned_loss=0.07353, over 927774.51 frames. ], batch size: 44, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:55:16,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.5161, 3.9415, 4.1519, 4.3702, 4.2567, 3.9814, 4.6566, 1.3880], + device='cuda:0'), covar=tensor([0.0757, 0.0729, 0.0720, 0.0848, 0.1213, 0.1535, 0.0641, 0.5329], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0240, 0.0272, 0.0288, 0.0329, 0.0278, 0.0300, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:55:25,044 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0482, 4.5736, 4.4090, 2.6678, 4.6813, 3.4720, 1.2320, 3.2470], + device='cuda:0'), covar=tensor([0.2099, 0.1282, 0.1123, 0.2495, 0.0686, 0.0857, 0.3915, 0.1138], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0174, 0.0163, 0.0131, 0.0157, 0.0124, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 08:55:31,876 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.702e+02 1.948e+02 2.422e+02 4.930e+02, threshold=3.896e+02, percent-clipped=3.0 +2023-03-26 08:55:51,723 INFO [finetune.py:976] (0/7) Epoch 8, batch 750, loss[loss=0.2441, simple_loss=0.3009, pruned_loss=0.09366, over 4822.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2722, pruned_loss=0.07374, over 935093.48 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:55:54,208 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:00,054 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 08:56:01,141 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:20,360 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:32,133 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40874.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:55,830 INFO [finetune.py:976] (0/7) Epoch 8, batch 800, loss[loss=0.1404, simple_loss=0.2131, pruned_loss=0.03389, over 4798.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2711, pruned_loss=0.07248, over 941134.41 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:57:03,988 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:04,636 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:23,553 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40922.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:26,521 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40926.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:27,573 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.606e+02 1.985e+02 2.397e+02 9.945e+02, threshold=3.971e+02, percent-clipped=3.0 +2023-03-26 08:57:27,674 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40928.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:45,605 INFO [finetune.py:976] (0/7) Epoch 8, batch 850, loss[loss=0.2048, simple_loss=0.262, pruned_loss=0.07375, over 4828.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2687, pruned_loss=0.07211, over 944438.48 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:00,384 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40961.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 08:58:06,566 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 08:58:10,967 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:18,103 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:22,822 INFO [finetune.py:976] (0/7) Epoch 8, batch 900, loss[loss=0.1748, simple_loss=0.2363, pruned_loss=0.05668, over 4755.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2652, pruned_loss=0.07085, over 946469.08 frames. ], batch size: 23, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:25,181 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:31,500 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:39,350 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3926, 2.2548, 1.8622, 0.9473, 2.0433, 1.8349, 1.6470, 1.9966], + device='cuda:0'), covar=tensor([0.0867, 0.0728, 0.1592, 0.1995, 0.1376, 0.2103, 0.2310, 0.1022], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0199, 0.0201, 0.0187, 0.0217, 0.0205, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:58:46,165 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.525e+02 1.868e+02 2.283e+02 3.598e+02, threshold=3.736e+02, percent-clipped=0.0 +2023-03-26 08:58:50,349 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41034.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:53,953 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9860, 1.9447, 1.5892, 1.9537, 1.8022, 1.7734, 1.8653, 2.5204], + device='cuda:0'), covar=tensor([0.4942, 0.5363, 0.3928, 0.5145, 0.4845, 0.2811, 0.5012, 0.1957], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0258, 0.0221, 0.0280, 0.0242, 0.0206, 0.0245, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:58:56,854 INFO [finetune.py:976] (0/7) Epoch 8, batch 950, loss[loss=0.2183, simple_loss=0.2677, pruned_loss=0.08446, over 4916.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2638, pruned_loss=0.07036, over 948382.94 frames. ], batch size: 38, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:57,569 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41045.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:06,036 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:12,504 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9290, 4.0692, 3.8280, 1.9562, 4.1606, 3.1189, 0.8957, 2.7287], + device='cuda:0'), covar=tensor([0.2344, 0.1683, 0.1414, 0.3157, 0.0844, 0.0971, 0.4283, 0.1476], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0173, 0.0162, 0.0131, 0.0157, 0.0123, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 08:59:23,031 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7494, 1.6113, 1.4277, 1.3663, 1.7885, 1.4813, 1.7867, 1.6921], + device='cuda:0'), covar=tensor([0.1562, 0.2254, 0.3450, 0.2653, 0.2764, 0.1910, 0.2977, 0.2022], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0191, 0.0236, 0.0256, 0.0237, 0.0194, 0.0212, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 08:59:30,583 INFO [finetune.py:976] (0/7) Epoch 8, batch 1000, loss[loss=0.1959, simple_loss=0.2615, pruned_loss=0.06521, over 4795.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2659, pruned_loss=0.07135, over 948598.93 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:59:36,007 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:38,377 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:52,970 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.652e+02 2.000e+02 2.359e+02 4.809e+02, threshold=4.000e+02, percent-clipped=2.0 +2023-03-26 09:00:04,080 INFO [finetune.py:976] (0/7) Epoch 8, batch 1050, loss[loss=0.1827, simple_loss=0.2502, pruned_loss=0.05756, over 4895.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2675, pruned_loss=0.07166, over 951040.78 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:00:06,628 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:12,013 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7480, 1.5549, 2.0686, 3.3414, 2.3953, 2.4236, 1.0476, 2.7207], + device='cuda:0'), covar=tensor([0.1701, 0.1538, 0.1419, 0.0603, 0.0780, 0.1283, 0.1792, 0.0614], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0138, 0.0126, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 09:00:16,272 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:37,444 INFO [finetune.py:976] (0/7) Epoch 8, batch 1100, loss[loss=0.252, simple_loss=0.3152, pruned_loss=0.09436, over 4718.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2682, pruned_loss=0.07175, over 951876.06 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:00:38,722 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:41,931 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 09:00:50,091 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4830, 1.3344, 1.2535, 1.5361, 1.6501, 1.5408, 0.8574, 1.2726], + device='cuda:0'), covar=tensor([0.2474, 0.2410, 0.2162, 0.1826, 0.1736, 0.1326, 0.2978, 0.2014], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0211, 0.0207, 0.0189, 0.0241, 0.0180, 0.0216, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:00:54,974 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41221.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:57,433 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7291, 3.7276, 3.5858, 1.8662, 3.9114, 2.7642, 0.7626, 2.5580], + device='cuda:0'), covar=tensor([0.2419, 0.1533, 0.1534, 0.3083, 0.0906, 0.1036, 0.4337, 0.1485], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0173, 0.0162, 0.0131, 0.0157, 0.0124, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:00:59,684 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.750e+02 2.155e+02 2.664e+02 4.791e+02, threshold=4.309e+02, percent-clipped=2.0 +2023-03-26 09:01:17,464 INFO [finetune.py:976] (0/7) Epoch 8, batch 1150, loss[loss=0.2412, simple_loss=0.3058, pruned_loss=0.08832, over 4898.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2699, pruned_loss=0.07247, over 951759.22 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:01:31,996 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41256.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 09:02:04,131 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 09:02:15,112 INFO [finetune.py:976] (0/7) Epoch 8, batch 1200, loss[loss=0.1898, simple_loss=0.2525, pruned_loss=0.06357, over 4873.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2684, pruned_loss=0.07181, over 954219.77 frames. ], batch size: 32, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:02:24,087 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:02:37,270 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.297e+01 1.635e+02 1.914e+02 2.289e+02 4.123e+02, threshold=3.829e+02, percent-clipped=0.0 +2023-03-26 09:02:38,628 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9645, 1.0382, 1.8533, 1.7751, 1.6245, 1.5828, 1.6175, 1.6826], + device='cuda:0'), covar=tensor([0.3940, 0.5218, 0.4257, 0.4467, 0.5692, 0.4294, 0.5642, 0.4171], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0241, 0.0254, 0.0254, 0.0245, 0.0222, 0.0272, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:02:39,195 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6579, 1.5347, 2.1717, 1.9527, 1.8425, 4.2391, 1.6687, 1.9118], + device='cuda:0'), covar=tensor([0.1108, 0.2090, 0.1180, 0.1159, 0.1695, 0.0288, 0.1513, 0.1839], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:02:41,054 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 09:02:42,648 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5863, 1.3496, 1.9641, 2.8164, 2.0051, 2.1779, 1.1041, 2.2825], + device='cuda:0'), covar=tensor([0.1589, 0.1474, 0.1115, 0.0589, 0.0695, 0.1494, 0.1440, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0166, 0.0102, 0.0140, 0.0127, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 09:02:43,464 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 09:02:51,370 INFO [finetune.py:976] (0/7) Epoch 8, batch 1250, loss[loss=0.1771, simple_loss=0.2579, pruned_loss=0.0481, over 4832.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2652, pruned_loss=0.07021, over 954140.73 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:03:02,102 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7571, 1.2383, 0.9433, 1.5423, 2.1530, 1.2412, 1.4912, 1.5874], + device='cuda:0'), covar=tensor([0.1507, 0.2246, 0.2020, 0.1272, 0.1967, 0.2045, 0.1524, 0.2021], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0097, 0.0114, 0.0092, 0.0124, 0.0096, 0.0101, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 09:03:02,675 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:03,907 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:17,124 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 09:03:32,985 INFO [finetune.py:976] (0/7) Epoch 8, batch 1300, loss[loss=0.1996, simple_loss=0.2545, pruned_loss=0.07232, over 4823.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2616, pruned_loss=0.06853, over 954180.65 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:03:37,902 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41401.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:48,489 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6214, 2.3809, 1.8950, 0.8419, 2.0976, 1.9383, 1.8229, 2.0554], + device='cuda:0'), covar=tensor([0.0828, 0.0874, 0.1487, 0.2149, 0.1421, 0.2189, 0.2151, 0.1002], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0203, 0.0189, 0.0219, 0.0207, 0.0224, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:03:56,242 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.674e+02 1.900e+02 2.309e+02 4.379e+02, threshold=3.799e+02, percent-clipped=1.0 +2023-03-26 09:04:05,874 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 09:04:06,253 INFO [finetune.py:976] (0/7) Epoch 8, batch 1350, loss[loss=0.1663, simple_loss=0.2308, pruned_loss=0.05092, over 4820.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2614, pruned_loss=0.06857, over 952754.98 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:04:10,436 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7269, 1.6344, 1.4053, 1.4382, 1.8004, 1.5017, 1.8270, 1.7162], + device='cuda:0'), covar=tensor([0.1467, 0.2394, 0.3314, 0.2675, 0.2729, 0.1748, 0.3361, 0.1893], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0189, 0.0234, 0.0254, 0.0235, 0.0193, 0.0211, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:04:15,880 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-26 09:04:16,292 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:04:19,213 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8207, 1.6053, 1.3973, 1.4803, 1.5316, 1.5570, 1.5869, 2.2945], + device='cuda:0'), covar=tensor([0.5275, 0.5562, 0.4251, 0.4939, 0.5036, 0.2993, 0.4972, 0.2114], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0257, 0.0221, 0.0280, 0.0241, 0.0205, 0.0245, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:04:20,407 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9730, 2.0094, 1.9452, 1.3552, 2.0311, 2.0612, 2.0823, 1.5825], + device='cuda:0'), covar=tensor([0.0659, 0.0619, 0.0765, 0.0962, 0.0595, 0.0767, 0.0667, 0.1206], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0133, 0.0144, 0.0126, 0.0114, 0.0144, 0.0145, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:04:39,875 INFO [finetune.py:976] (0/7) Epoch 8, batch 1400, loss[loss=0.2451, simple_loss=0.3039, pruned_loss=0.09315, over 4927.00 frames. ], tot_loss[loss=0.2037, simple_loss=0.2664, pruned_loss=0.07044, over 953101.62 frames. ], batch size: 33, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:04:58,585 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:05:02,219 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5288, 1.7604, 1.4217, 1.3989, 1.8231, 1.7134, 1.6960, 1.6701], + device='cuda:0'), covar=tensor([0.0385, 0.0270, 0.0421, 0.0349, 0.0305, 0.0615, 0.0305, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0112, 0.0141, 0.0117, 0.0105, 0.0102, 0.0092, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.0729e-05, 8.8100e-05, 1.1285e-04, 9.2435e-05, 8.2816e-05, 7.5376e-05, + 6.9343e-05, 8.6015e-05], device='cuda:0') +2023-03-26 09:05:03,328 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.704e+02 2.004e+02 2.444e+02 3.700e+02, threshold=4.008e+02, percent-clipped=0.0 +2023-03-26 09:05:12,589 INFO [finetune.py:976] (0/7) Epoch 8, batch 1450, loss[loss=0.2219, simple_loss=0.2721, pruned_loss=0.08585, over 4805.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2677, pruned_loss=0.07051, over 953522.54 frames. ], batch size: 25, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:05:21,958 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41556.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:05:22,099 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 09:05:30,811 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:05:46,315 INFO [finetune.py:976] (0/7) Epoch 8, batch 1500, loss[loss=0.2191, simple_loss=0.2812, pruned_loss=0.07855, over 4808.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2699, pruned_loss=0.0717, over 953877.70 frames. ], batch size: 45, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:05:51,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9029, 0.9513, 1.7541, 1.6573, 1.5684, 1.5045, 1.5000, 1.6405], + device='cuda:0'), covar=tensor([0.4568, 0.6004, 0.5090, 0.5156, 0.6296, 0.4807, 0.6274, 0.4624], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0241, 0.0253, 0.0253, 0.0244, 0.0221, 0.0272, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:05:53,540 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:06:03,488 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5116, 1.3406, 1.3143, 1.3618, 1.1142, 3.2061, 1.2547, 1.6784], + device='cuda:0'), covar=tensor([0.4368, 0.3255, 0.2599, 0.3100, 0.2003, 0.0283, 0.2717, 0.1415], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:06:10,831 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.635e+02 1.924e+02 2.365e+02 3.634e+02, threshold=3.848e+02, percent-clipped=0.0 +2023-03-26 09:06:19,421 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2170, 3.6212, 3.8316, 4.0715, 3.9610, 3.6871, 4.2861, 1.5176], + device='cuda:0'), covar=tensor([0.0706, 0.0745, 0.0779, 0.0848, 0.1160, 0.1396, 0.0657, 0.4836], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0276, 0.0294, 0.0332, 0.0283, 0.0303, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:06:22,437 INFO [finetune.py:976] (0/7) Epoch 8, batch 1550, loss[loss=0.1844, simple_loss=0.2493, pruned_loss=0.05974, over 4811.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2693, pruned_loss=0.07141, over 953032.66 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:06:34,449 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:06:35,067 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:06:38,632 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2181, 1.7474, 2.0520, 2.0998, 1.8167, 1.8543, 1.9639, 1.9049], + device='cuda:0'), covar=tensor([0.4775, 0.5930, 0.4455, 0.5243, 0.6086, 0.4816, 0.6941, 0.4396], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0241, 0.0253, 0.0253, 0.0244, 0.0221, 0.0272, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:07:05,613 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7682, 1.9514, 1.5704, 1.4527, 2.1476, 2.1177, 1.9240, 1.8189], + device='cuda:0'), covar=tensor([0.0376, 0.0326, 0.0472, 0.0379, 0.0341, 0.0633, 0.0318, 0.0390], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0112, 0.0140, 0.0117, 0.0105, 0.0101, 0.0091, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.0721e-05, 8.7791e-05, 1.1189e-04, 9.2112e-05, 8.2522e-05, 7.5037e-05, + 6.9188e-05, 8.5342e-05], device='cuda:0') +2023-03-26 09:07:19,719 INFO [finetune.py:976] (0/7) Epoch 8, batch 1600, loss[loss=0.2071, simple_loss=0.2739, pruned_loss=0.07014, over 4907.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2671, pruned_loss=0.07066, over 953581.70 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:07:25,526 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:07:25,552 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:07:39,585 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41715.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 09:07:48,911 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.581e+02 1.949e+02 2.490e+02 4.755e+02, threshold=3.899e+02, percent-clipped=2.0 +2023-03-26 09:07:58,438 INFO [finetune.py:976] (0/7) Epoch 8, batch 1650, loss[loss=0.2272, simple_loss=0.2748, pruned_loss=0.08986, over 4938.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2633, pruned_loss=0.06951, over 954843.21 frames. ], batch size: 33, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:08:01,525 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41749.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:08:03,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7829, 3.7117, 3.5461, 1.8233, 3.7633, 2.8163, 0.8433, 2.5900], + device='cuda:0'), covar=tensor([0.2513, 0.1960, 0.1530, 0.3379, 0.1092, 0.1140, 0.4511, 0.1654], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0172, 0.0161, 0.0130, 0.0156, 0.0123, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:08:09,849 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41758.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:08:42,690 INFO [finetune.py:976] (0/7) Epoch 8, batch 1700, loss[loss=0.2291, simple_loss=0.2855, pruned_loss=0.08636, over 4759.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2615, pruned_loss=0.06858, over 955275.59 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:08:49,888 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 09:08:50,302 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:09:05,089 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 09:09:06,690 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.763e+02 2.034e+02 2.335e+02 4.675e+02, threshold=4.069e+02, percent-clipped=2.0 +2023-03-26 09:09:13,295 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7365, 1.5929, 1.3990, 1.2620, 1.7354, 1.4515, 1.7756, 1.6434], + device='cuda:0'), covar=tensor([0.1619, 0.2537, 0.3756, 0.3091, 0.3261, 0.2070, 0.3858, 0.2288], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0190, 0.0235, 0.0255, 0.0236, 0.0194, 0.0212, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:09:16,758 INFO [finetune.py:976] (0/7) Epoch 8, batch 1750, loss[loss=0.2441, simple_loss=0.303, pruned_loss=0.09259, over 4873.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2653, pruned_loss=0.07033, over 955246.84 frames. ], batch size: 34, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:09:50,589 INFO [finetune.py:976] (0/7) Epoch 8, batch 1800, loss[loss=0.2167, simple_loss=0.2722, pruned_loss=0.08061, over 4766.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2688, pruned_loss=0.07162, over 956491.74 frames. ], batch size: 28, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:09:57,969 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:10:10,749 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3265, 2.8524, 2.6581, 1.3197, 2.8740, 2.3013, 2.1417, 2.4263], + device='cuda:0'), covar=tensor([0.0909, 0.0897, 0.1602, 0.2279, 0.1423, 0.1961, 0.2114, 0.1236], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0201, 0.0203, 0.0189, 0.0219, 0.0208, 0.0223, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:10:13,608 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.792e+02 2.103e+02 2.633e+02 4.479e+02, threshold=4.207e+02, percent-clipped=2.0 +2023-03-26 09:10:23,653 INFO [finetune.py:976] (0/7) Epoch 8, batch 1850, loss[loss=0.1865, simple_loss=0.2581, pruned_loss=0.05751, over 4859.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2695, pruned_loss=0.07134, over 956158.11 frames. ], batch size: 31, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:10:26,670 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41948.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:10:30,550 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 09:10:38,713 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41967.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:10:57,329 INFO [finetune.py:976] (0/7) Epoch 8, batch 1900, loss[loss=0.194, simple_loss=0.2721, pruned_loss=0.05797, over 4914.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2708, pruned_loss=0.07172, over 954958.81 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:11:01,614 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-42000.pt +2023-03-26 09:11:02,719 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4916, 0.9961, 0.7842, 1.3245, 1.9020, 0.7348, 1.1672, 1.2442], + device='cuda:0'), covar=tensor([0.1757, 0.2632, 0.2020, 0.1490, 0.2275, 0.2194, 0.1929, 0.2344], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0098, 0.0115, 0.0093, 0.0125, 0.0097, 0.0102, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 09:11:08,249 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42009.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:11:08,792 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42010.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:11:22,121 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.556e+02 1.920e+02 2.218e+02 3.872e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 09:11:32,121 INFO [finetune.py:976] (0/7) Epoch 8, batch 1950, loss[loss=0.1747, simple_loss=0.2417, pruned_loss=0.05387, over 4875.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2689, pruned_loss=0.07079, over 953706.12 frames. ], batch size: 34, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:11:34,009 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0344, 2.0518, 1.9565, 1.3513, 2.1734, 2.1415, 2.0820, 1.7179], + device='cuda:0'), covar=tensor([0.0601, 0.0638, 0.0771, 0.0972, 0.0519, 0.0769, 0.0677, 0.1115], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0131, 0.0141, 0.0124, 0.0112, 0.0141, 0.0143, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:12:30,959 INFO [finetune.py:976] (0/7) Epoch 8, batch 2000, loss[loss=0.1885, simple_loss=0.242, pruned_loss=0.06751, over 4902.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2662, pruned_loss=0.07022, over 953679.54 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:12:56,438 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.506e+02 1.840e+02 2.176e+02 3.856e+02, threshold=3.679e+02, percent-clipped=1.0 +2023-03-26 09:13:06,571 INFO [finetune.py:976] (0/7) Epoch 8, batch 2050, loss[loss=0.1911, simple_loss=0.2478, pruned_loss=0.06723, over 4748.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2618, pruned_loss=0.06822, over 955651.38 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:13:13,557 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.0820, 4.3687, 4.6029, 4.8281, 4.7582, 4.5816, 5.1709, 1.4323], + device='cuda:0'), covar=tensor([0.0799, 0.0806, 0.0789, 0.1204, 0.1342, 0.1248, 0.0570, 0.5720], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0242, 0.0276, 0.0293, 0.0333, 0.0282, 0.0302, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:13:53,279 INFO [finetune.py:976] (0/7) Epoch 8, batch 2100, loss[loss=0.1855, simple_loss=0.2481, pruned_loss=0.06146, over 4757.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2606, pruned_loss=0.06758, over 956734.27 frames. ], batch size: 26, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:14:05,644 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 09:14:08,581 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0324, 1.9918, 1.6250, 1.9657, 1.8978, 1.8785, 1.8805, 2.6399], + device='cuda:0'), covar=tensor([0.5347, 0.6607, 0.4452, 0.6237, 0.6049, 0.3086, 0.6129, 0.2015], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0259, 0.0222, 0.0282, 0.0242, 0.0206, 0.0245, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:14:16,279 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.710e+02 1.945e+02 2.376e+02 4.149e+02, threshold=3.889e+02, percent-clipped=2.0 +2023-03-26 09:14:21,730 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9966, 2.0883, 1.5112, 2.1221, 1.9438, 1.7935, 1.8548, 2.7506], + device='cuda:0'), covar=tensor([0.5752, 0.6245, 0.4737, 0.5798, 0.5801, 0.3202, 0.6308, 0.2018], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0259, 0.0221, 0.0281, 0.0241, 0.0206, 0.0245, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:14:26,997 INFO [finetune.py:976] (0/7) Epoch 8, batch 2150, loss[loss=0.1757, simple_loss=0.2506, pruned_loss=0.05045, over 4910.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2626, pruned_loss=0.06784, over 956130.49 frames. ], batch size: 37, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:14:38,876 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42262.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:15:18,053 INFO [finetune.py:976] (0/7) Epoch 8, batch 2200, loss[loss=0.2034, simple_loss=0.268, pruned_loss=0.06934, over 4820.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2669, pruned_loss=0.06955, over 956641.98 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:15:25,343 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42304.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:15:27,349 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-26 09:15:29,021 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42310.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:15:45,962 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.543e+02 1.921e+02 2.479e+02 5.347e+02, threshold=3.843e+02, percent-clipped=1.0 +2023-03-26 09:16:06,876 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8393, 1.6530, 1.6282, 1.7392, 1.0636, 4.3113, 1.6446, 2.2478], + device='cuda:0'), covar=tensor([0.3343, 0.2445, 0.2099, 0.2229, 0.2033, 0.0115, 0.2544, 0.1326], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0116, 0.0098, 0.0100, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:16:07,395 INFO [finetune.py:976] (0/7) Epoch 8, batch 2250, loss[loss=0.1882, simple_loss=0.2755, pruned_loss=0.05046, over 4916.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2688, pruned_loss=0.07045, over 956149.42 frames. ], batch size: 42, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:16:14,876 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4444, 1.4656, 1.2372, 1.3480, 1.6977, 1.6542, 1.4083, 1.2397], + device='cuda:0'), covar=tensor([0.0287, 0.0242, 0.0478, 0.0286, 0.0188, 0.0364, 0.0354, 0.0348], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0111, 0.0139, 0.0116, 0.0104, 0.0101, 0.0090, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.0111e-05, 8.7343e-05, 1.1111e-04, 9.1437e-05, 8.1752e-05, 7.5017e-05, + 6.8355e-05, 8.4770e-05], device='cuda:0') +2023-03-26 09:16:27,424 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:16:43,620 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3974, 2.3013, 1.9322, 0.9479, 2.0614, 1.8478, 1.7053, 2.1068], + device='cuda:0'), covar=tensor([0.0909, 0.0652, 0.1372, 0.1977, 0.1378, 0.2233, 0.1884, 0.0846], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0203, 0.0188, 0.0219, 0.0207, 0.0223, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:16:47,918 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:17:07,514 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 09:17:09,028 INFO [finetune.py:976] (0/7) Epoch 8, batch 2300, loss[loss=0.1616, simple_loss=0.2222, pruned_loss=0.05047, over 4086.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2697, pruned_loss=0.07026, over 956433.59 frames. ], batch size: 17, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:17:15,814 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7114, 3.6064, 3.3882, 1.8897, 3.7062, 2.7485, 0.9551, 2.5870], + device='cuda:0'), covar=tensor([0.2344, 0.2240, 0.1695, 0.3252, 0.1168, 0.1207, 0.4570, 0.1506], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0171, 0.0160, 0.0128, 0.0155, 0.0122, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:17:51,929 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7657, 1.6517, 2.0620, 1.4129, 1.8373, 1.9987, 1.6598, 2.2787], + device='cuda:0'), covar=tensor([0.1335, 0.1831, 0.1381, 0.1873, 0.1063, 0.1509, 0.2571, 0.0849], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0203, 0.0197, 0.0195, 0.0180, 0.0219, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:17:57,227 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.935e+01 1.473e+02 1.816e+02 2.175e+02 3.275e+02, threshold=3.633e+02, percent-clipped=0.0 +2023-03-26 09:18:06,502 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:18:09,278 INFO [finetune.py:976] (0/7) Epoch 8, batch 2350, loss[loss=0.2243, simple_loss=0.2801, pruned_loss=0.08428, over 4881.00 frames. ], tot_loss[loss=0.203, simple_loss=0.267, pruned_loss=0.06944, over 956196.75 frames. ], batch size: 43, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:18:40,036 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7609, 2.5431, 2.1123, 1.1246, 2.2955, 2.0809, 1.9306, 2.3335], + device='cuda:0'), covar=tensor([0.0773, 0.0878, 0.1418, 0.2149, 0.1499, 0.2243, 0.2035, 0.0959], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0203, 0.0188, 0.0219, 0.0207, 0.0223, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:18:49,078 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8199, 1.7835, 1.4613, 1.6170, 2.0023, 2.0189, 1.7682, 1.4856], + device='cuda:0'), covar=tensor([0.0252, 0.0285, 0.0528, 0.0312, 0.0203, 0.0336, 0.0282, 0.0410], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0111, 0.0139, 0.0116, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.0068e-05, 8.7079e-05, 1.1133e-04, 9.1347e-05, 8.1824e-05, 7.4771e-05, + 6.8501e-05, 8.4775e-05], device='cuda:0') +2023-03-26 09:18:51,871 INFO [finetune.py:976] (0/7) Epoch 8, batch 2400, loss[loss=0.1536, simple_loss=0.222, pruned_loss=0.04259, over 4036.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2639, pruned_loss=0.06866, over 954905.59 frames. ], batch size: 17, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:19:02,409 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:19:25,682 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.517e+02 1.798e+02 2.223e+02 5.682e+02, threshold=3.597e+02, percent-clipped=2.0 +2023-03-26 09:19:35,402 INFO [finetune.py:976] (0/7) Epoch 8, batch 2450, loss[loss=0.1884, simple_loss=0.2511, pruned_loss=0.06285, over 4771.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2612, pruned_loss=0.06805, over 956069.59 frames. ], batch size: 26, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:19:47,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42562.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:19:51,987 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:04,302 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 09:20:08,910 INFO [finetune.py:976] (0/7) Epoch 8, batch 2500, loss[loss=0.2364, simple_loss=0.3028, pruned_loss=0.08496, over 4304.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2625, pruned_loss=0.06865, over 954197.67 frames. ], batch size: 65, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:20:16,184 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42604.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:20:20,726 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:33,723 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.709e+02 1.979e+02 2.316e+02 5.134e+02, threshold=3.959e+02, percent-clipped=4.0 +2023-03-26 09:20:35,812 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 09:20:42,882 INFO [finetune.py:976] (0/7) Epoch 8, batch 2550, loss[loss=0.1956, simple_loss=0.2638, pruned_loss=0.06371, over 4834.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2679, pruned_loss=0.07057, over 954841.05 frames. ], batch size: 49, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:20:46,523 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:53,412 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42652.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:20:55,286 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0846, 2.1825, 2.0341, 1.4934, 2.3103, 2.3453, 2.3036, 1.8958], + device='cuda:0'), covar=tensor([0.0715, 0.0683, 0.0908, 0.1063, 0.0512, 0.0769, 0.0720, 0.1048], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0125, 0.0114, 0.0143, 0.0144, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:20:56,614 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 09:21:04,736 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 09:21:14,123 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-26 09:21:25,366 INFO [finetune.py:976] (0/7) Epoch 8, batch 2600, loss[loss=0.2014, simple_loss=0.2614, pruned_loss=0.07069, over 4821.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2692, pruned_loss=0.07144, over 955106.09 frames. ], batch size: 30, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:21:35,573 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1077, 0.9943, 1.0621, 0.3187, 0.8418, 1.1338, 1.2126, 0.9750], + device='cuda:0'), covar=tensor([0.0868, 0.0493, 0.0426, 0.0583, 0.0524, 0.0550, 0.0374, 0.0620], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0156, 0.0120, 0.0135, 0.0132, 0.0125, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5473e-05, 1.1468e-04, 8.6315e-05, 9.8468e-05, 9.4264e-05, 9.1482e-05, + 1.0583e-04, 1.0820e-04], device='cuda:0') +2023-03-26 09:21:36,664 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42710.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:21:49,607 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.771e+02 2.168e+02 2.787e+02 4.495e+02, threshold=4.337e+02, percent-clipped=4.0 +2023-03-26 09:21:53,818 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:21:59,151 INFO [finetune.py:976] (0/7) Epoch 8, batch 2650, loss[loss=0.2352, simple_loss=0.2978, pruned_loss=0.08628, over 4910.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2708, pruned_loss=0.07201, over 954625.56 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:22:03,614 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 09:22:30,408 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2854, 2.0811, 1.8637, 2.1130, 2.0313, 2.0166, 2.0408, 2.8249], + device='cuda:0'), covar=tensor([0.4715, 0.5712, 0.4017, 0.4645, 0.4504, 0.2950, 0.4702, 0.1769], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0259, 0.0221, 0.0281, 0.0241, 0.0207, 0.0245, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:22:40,396 INFO [finetune.py:976] (0/7) Epoch 8, batch 2700, loss[loss=0.1968, simple_loss=0.2535, pruned_loss=0.07011, over 4765.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2686, pruned_loss=0.07027, over 954331.15 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:23:27,185 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7608, 3.6721, 3.4883, 1.7703, 3.7141, 2.8397, 0.6457, 2.5727], + device='cuda:0'), covar=tensor([0.2130, 0.1699, 0.1444, 0.3066, 0.0994, 0.1033, 0.4369, 0.1316], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0169, 0.0158, 0.0128, 0.0154, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:23:27,701 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.604e+02 1.897e+02 2.218e+02 3.599e+02, threshold=3.793e+02, percent-clipped=0.0 +2023-03-26 09:23:36,581 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 09:23:46,886 INFO [finetune.py:976] (0/7) Epoch 8, batch 2750, loss[loss=0.203, simple_loss=0.2616, pruned_loss=0.07217, over 4795.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2662, pruned_loss=0.06996, over 954470.19 frames. ], batch size: 29, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:23:55,958 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:02,856 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42862.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:14,753 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42879.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:31,685 INFO [finetune.py:976] (0/7) Epoch 8, batch 2800, loss[loss=0.2351, simple_loss=0.2878, pruned_loss=0.09119, over 4731.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2629, pruned_loss=0.06853, over 956919.98 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:24:48,567 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:53,410 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 09:24:54,864 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.137e+02 1.632e+02 1.941e+02 2.379e+02 3.960e+02, threshold=3.882e+02, percent-clipped=2.0 +2023-03-26 09:25:02,612 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:25:04,920 INFO [finetune.py:976] (0/7) Epoch 8, batch 2850, loss[loss=0.2249, simple_loss=0.2864, pruned_loss=0.08165, over 4833.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2636, pruned_loss=0.06948, over 955969.90 frames. ], batch size: 39, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:25:16,480 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8526, 1.6886, 1.6141, 1.7282, 1.2701, 3.3668, 1.3378, 1.8471], + device='cuda:0'), covar=tensor([0.3081, 0.2373, 0.1978, 0.2232, 0.1819, 0.0226, 0.2463, 0.1230], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0118, 0.0123, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:25:38,296 INFO [finetune.py:976] (0/7) Epoch 8, batch 2900, loss[loss=0.2052, simple_loss=0.2698, pruned_loss=0.07027, over 4824.00 frames. ], tot_loss[loss=0.2054, simple_loss=0.2685, pruned_loss=0.07112, over 955473.85 frames. ], batch size: 33, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:25:45,628 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43005.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:03,398 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.926e+01 1.731e+02 1.970e+02 2.373e+02 5.777e+02, threshold=3.941e+02, percent-clipped=2.0 +2023-03-26 09:26:05,322 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0506, 0.8229, 0.8654, 0.2465, 0.7424, 0.9554, 1.0264, 0.8250], + device='cuda:0'), covar=tensor([0.0797, 0.0527, 0.0448, 0.0565, 0.0528, 0.0671, 0.0400, 0.0612], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0156, 0.0120, 0.0135, 0.0132, 0.0125, 0.0145, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5358e-05, 1.1456e-04, 8.6378e-05, 9.8422e-05, 9.4300e-05, 9.1496e-05, + 1.0621e-04, 1.0819e-04], device='cuda:0') +2023-03-26 09:26:13,021 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43035.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:16,085 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-26 09:26:18,884 INFO [finetune.py:976] (0/7) Epoch 8, batch 2950, loss[loss=0.2282, simple_loss=0.3046, pruned_loss=0.07593, over 4807.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2712, pruned_loss=0.0726, over 954550.29 frames. ], batch size: 40, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:26:44,521 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:52,571 INFO [finetune.py:976] (0/7) Epoch 8, batch 3000, loss[loss=0.2278, simple_loss=0.2953, pruned_loss=0.0802, over 4915.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2721, pruned_loss=0.07268, over 955973.11 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:26:52,572 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 09:26:58,430 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6669, 1.5171, 1.5203, 1.5621, 1.0825, 3.0215, 1.0981, 1.6744], + device='cuda:0'), covar=tensor([0.3251, 0.2373, 0.2005, 0.2160, 0.1847, 0.0250, 0.2696, 0.1293], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:26:58,536 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4012, 1.5238, 1.3781, 1.6496, 1.6282, 3.0333, 1.4177, 1.5896], + device='cuda:0'), covar=tensor([0.0979, 0.1842, 0.1082, 0.0927, 0.1572, 0.0261, 0.1480, 0.1693], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0082, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 09:27:10,878 INFO [finetune.py:1010] (0/7) Epoch 8, validation: loss=0.16, simple_loss=0.2311, pruned_loss=0.04446, over 2265189.00 frames. +2023-03-26 09:27:10,878 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 09:27:49,850 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.686e+02 2.049e+02 2.426e+02 3.920e+02, threshold=4.099e+02, percent-clipped=0.0 +2023-03-26 09:28:00,451 INFO [finetune.py:976] (0/7) Epoch 8, batch 3050, loss[loss=0.2262, simple_loss=0.2944, pruned_loss=0.07895, over 4842.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2724, pruned_loss=0.07274, over 955330.11 frames. ], batch size: 44, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:28:03,005 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2545, 2.1100, 1.6654, 2.1957, 2.2562, 1.9297, 2.5237, 2.2131], + device='cuda:0'), covar=tensor([0.1602, 0.2830, 0.3721, 0.3245, 0.2780, 0.1948, 0.3455, 0.2334], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0188, 0.0234, 0.0254, 0.0235, 0.0193, 0.0211, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:28:13,421 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43162.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:28:36,035 INFO [finetune.py:976] (0/7) Epoch 8, batch 3100, loss[loss=0.17, simple_loss=0.232, pruned_loss=0.05404, over 4764.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.269, pruned_loss=0.07133, over 955843.87 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:28:52,822 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43210.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:29:01,136 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:29:03,053 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1925, 1.6906, 2.1084, 2.0112, 1.7577, 1.8227, 1.8877, 1.9111], + device='cuda:0'), covar=tensor([0.4862, 0.6083, 0.4546, 0.5592, 0.6615, 0.4945, 0.7272, 0.4356], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0241, 0.0253, 0.0254, 0.0244, 0.0222, 0.0272, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:29:14,670 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.652e+02 1.930e+02 2.337e+02 4.149e+02, threshold=3.860e+02, percent-clipped=1.0 +2023-03-26 09:29:23,814 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:29:34,859 INFO [finetune.py:976] (0/7) Epoch 8, batch 3150, loss[loss=0.1827, simple_loss=0.2514, pruned_loss=0.05704, over 4765.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2656, pruned_loss=0.06974, over 954812.71 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:29:47,709 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1396, 1.9257, 1.6480, 1.8477, 1.8679, 1.7941, 1.8491, 2.5760], + device='cuda:0'), covar=tensor([0.5068, 0.5624, 0.4351, 0.5367, 0.5281, 0.3210, 0.5122, 0.2109], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0257, 0.0220, 0.0278, 0.0239, 0.0204, 0.0243, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:30:09,010 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9353, 1.8151, 1.9066, 1.2632, 2.0136, 2.0347, 1.8701, 1.6933], + device='cuda:0'), covar=tensor([0.0588, 0.0676, 0.0739, 0.0937, 0.0812, 0.0605, 0.0653, 0.1057], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0132, 0.0143, 0.0124, 0.0114, 0.0142, 0.0144, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:30:24,889 INFO [finetune.py:976] (0/7) Epoch 8, batch 3200, loss[loss=0.2039, simple_loss=0.266, pruned_loss=0.07088, over 4897.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2624, pruned_loss=0.06902, over 955623.47 frames. ], batch size: 32, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:30:32,643 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43305.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:30:49,477 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.668e+02 2.087e+02 2.541e+02 1.424e+03, threshold=4.174e+02, percent-clipped=3.0 +2023-03-26 09:30:56,528 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 09:31:03,658 INFO [finetune.py:976] (0/7) Epoch 8, batch 3250, loss[loss=0.2637, simple_loss=0.3185, pruned_loss=0.1044, over 4799.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2629, pruned_loss=0.06913, over 954601.46 frames. ], batch size: 51, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:31:06,209 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-26 09:31:15,391 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:31:59,776 INFO [finetune.py:976] (0/7) Epoch 8, batch 3300, loss[loss=0.2005, simple_loss=0.2778, pruned_loss=0.0616, over 4764.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2675, pruned_loss=0.07083, over 954146.99 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:32:45,203 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 09:32:45,626 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.691e+02 2.029e+02 2.462e+02 4.055e+02, threshold=4.059e+02, percent-clipped=0.0 +2023-03-26 09:32:45,855 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-26 09:32:46,382 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2338, 2.5412, 2.4873, 1.3077, 2.6587, 2.2170, 1.8126, 2.2780], + device='cuda:0'), covar=tensor([0.0795, 0.1301, 0.2030, 0.2576, 0.2080, 0.2402, 0.2777, 0.1598], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0202, 0.0203, 0.0189, 0.0220, 0.0208, 0.0225, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:33:04,667 INFO [finetune.py:976] (0/7) Epoch 8, batch 3350, loss[loss=0.1987, simple_loss=0.26, pruned_loss=0.06867, over 4797.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2696, pruned_loss=0.07166, over 953485.03 frames. ], batch size: 29, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:33:07,171 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5729, 1.3272, 1.9053, 3.1071, 2.1463, 2.2298, 0.7024, 2.3641], + device='cuda:0'), covar=tensor([0.1913, 0.1666, 0.1485, 0.0669, 0.0878, 0.1465, 0.2231, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0165, 0.0101, 0.0138, 0.0127, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 09:33:51,119 INFO [finetune.py:976] (0/7) Epoch 8, batch 3400, loss[loss=0.236, simple_loss=0.2939, pruned_loss=0.08902, over 4153.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2716, pruned_loss=0.07234, over 952633.38 frames. ], batch size: 66, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:34:04,480 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:05,088 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:10,991 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6906, 1.5926, 1.5799, 1.7020, 1.2147, 3.6070, 1.4114, 2.0511], + device='cuda:0'), covar=tensor([0.3402, 0.2482, 0.2109, 0.2436, 0.1935, 0.0183, 0.2628, 0.1208], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:34:15,396 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.653e+02 1.866e+02 2.233e+02 4.638e+02, threshold=3.733e+02, percent-clipped=1.0 +2023-03-26 09:34:19,627 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43535.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:25,008 INFO [finetune.py:976] (0/7) Epoch 8, batch 3450, loss[loss=0.2162, simple_loss=0.2885, pruned_loss=0.07194, over 4864.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2712, pruned_loss=0.07183, over 953022.88 frames. ], batch size: 34, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:34:28,263 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 09:34:43,301 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:55,920 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43575.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:35:02,191 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43583.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:35:08,915 INFO [finetune.py:976] (0/7) Epoch 8, batch 3500, loss[loss=0.2202, simple_loss=0.2668, pruned_loss=0.0868, over 4742.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2686, pruned_loss=0.07138, over 952229.46 frames. ], batch size: 59, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:35:34,549 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.655e+02 2.028e+02 2.395e+02 4.370e+02, threshold=4.057e+02, percent-clipped=4.0 +2023-03-26 09:35:44,695 INFO [finetune.py:976] (0/7) Epoch 8, batch 3550, loss[loss=0.1397, simple_loss=0.2023, pruned_loss=0.03851, over 4772.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2663, pruned_loss=0.07078, over 952892.58 frames. ], batch size: 26, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:35:44,899 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-26 09:36:12,256 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-03-26 09:36:13,464 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 09:36:18,004 INFO [finetune.py:976] (0/7) Epoch 8, batch 3600, loss[loss=0.1887, simple_loss=0.2528, pruned_loss=0.06232, over 4765.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2626, pruned_loss=0.06937, over 951676.21 frames. ], batch size: 26, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:36:18,672 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9806, 4.4734, 4.2444, 2.3686, 4.5737, 3.4345, 0.7661, 3.2375], + device='cuda:0'), covar=tensor([0.2389, 0.1565, 0.1286, 0.3040, 0.0718, 0.0845, 0.4584, 0.1225], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0171, 0.0160, 0.0129, 0.0155, 0.0122, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:36:38,845 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 09:36:40,280 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.641e+02 1.836e+02 2.142e+02 3.900e+02, threshold=3.673e+02, percent-clipped=0.0 +2023-03-26 09:36:55,752 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7358, 1.1521, 0.9302, 1.5677, 2.1157, 1.1653, 1.4029, 1.5497], + device='cuda:0'), covar=tensor([0.1725, 0.2492, 0.2028, 0.1337, 0.2084, 0.2006, 0.1693, 0.2220], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0096, 0.0112, 0.0091, 0.0122, 0.0094, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 09:36:56,351 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4806, 1.3626, 1.6241, 1.8110, 1.4716, 3.1591, 1.2416, 1.4501], + device='cuda:0'), covar=tensor([0.0947, 0.1853, 0.1271, 0.0957, 0.1697, 0.0271, 0.1620, 0.1824], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0076, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:37:03,251 INFO [finetune.py:976] (0/7) Epoch 8, batch 3650, loss[loss=0.2548, simple_loss=0.3106, pruned_loss=0.0995, over 4904.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2662, pruned_loss=0.07083, over 953353.02 frames. ], batch size: 43, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:37:41,727 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43787.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:37:53,053 INFO [finetune.py:976] (0/7) Epoch 8, batch 3700, loss[loss=0.1967, simple_loss=0.2621, pruned_loss=0.06563, over 4894.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.2701, pruned_loss=0.07223, over 952401.93 frames. ], batch size: 35, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:38:37,141 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.066e+01 1.637e+02 2.055e+02 2.501e+02 4.825e+02, threshold=4.110e+02, percent-clipped=4.0 +2023-03-26 09:38:56,855 INFO [finetune.py:976] (0/7) Epoch 8, batch 3750, loss[loss=0.1968, simple_loss=0.2546, pruned_loss=0.06952, over 4866.00 frames. ], tot_loss[loss=0.2085, simple_loss=0.2718, pruned_loss=0.07255, over 952529.40 frames. ], batch size: 31, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:38:57,609 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0896, 1.9645, 1.4693, 2.0644, 2.0099, 1.7599, 2.3864, 2.0545], + device='cuda:0'), covar=tensor([0.1502, 0.2679, 0.3704, 0.3228, 0.2789, 0.1850, 0.3717, 0.2070], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0188, 0.0233, 0.0253, 0.0235, 0.0193, 0.0210, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:39:00,428 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:39:07,139 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4562, 1.3509, 1.2475, 1.5265, 1.5657, 1.5574, 0.8757, 1.2449], + device='cuda:0'), covar=tensor([0.2269, 0.2204, 0.1940, 0.1659, 0.1797, 0.1265, 0.2828, 0.1928], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0209, 0.0206, 0.0188, 0.0240, 0.0179, 0.0215, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:39:13,813 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43870.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:39:30,692 INFO [finetune.py:976] (0/7) Epoch 8, batch 3800, loss[loss=0.1865, simple_loss=0.2561, pruned_loss=0.05844, over 4822.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.273, pruned_loss=0.07242, over 953860.06 frames. ], batch size: 39, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:39:40,897 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:40:01,520 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.597e+02 1.980e+02 2.453e+02 5.062e+02, threshold=3.959e+02, percent-clipped=3.0 +2023-03-26 09:40:09,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3479, 2.0782, 2.7877, 3.6794, 2.7436, 2.7740, 1.6764, 2.9239], + device='cuda:0'), covar=tensor([0.1320, 0.1139, 0.1013, 0.0551, 0.0653, 0.1303, 0.1500, 0.0562], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0166, 0.0103, 0.0140, 0.0127, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 09:40:16,149 INFO [finetune.py:976] (0/7) Epoch 8, batch 3850, loss[loss=0.2521, simple_loss=0.2983, pruned_loss=0.1029, over 4835.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2694, pruned_loss=0.07079, over 953804.06 frames. ], batch size: 47, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:40:26,724 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43959.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:40:33,704 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:40:35,566 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1131, 2.1866, 2.2261, 1.5079, 2.2448, 2.2904, 2.1171, 1.8558], + device='cuda:0'), covar=tensor([0.0608, 0.0582, 0.0661, 0.0914, 0.0531, 0.0632, 0.0641, 0.1067], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0133, 0.0144, 0.0126, 0.0115, 0.0144, 0.0145, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:40:54,206 INFO [finetune.py:976] (0/7) Epoch 8, batch 3900, loss[loss=0.2295, simple_loss=0.2593, pruned_loss=0.09983, over 4208.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2656, pruned_loss=0.06903, over 953567.32 frames. ], batch size: 18, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:41:02,696 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-44000.pt +2023-03-26 09:41:17,075 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:41:22,369 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.530e+02 1.807e+02 2.225e+02 3.627e+02, threshold=3.614e+02, percent-clipped=0.0 +2023-03-26 09:41:32,537 INFO [finetune.py:976] (0/7) Epoch 8, batch 3950, loss[loss=0.2369, simple_loss=0.2888, pruned_loss=0.09249, over 4900.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2622, pruned_loss=0.06832, over 952138.32 frames. ], batch size: 35, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:42:06,485 INFO [finetune.py:976] (0/7) Epoch 8, batch 4000, loss[loss=0.2184, simple_loss=0.2707, pruned_loss=0.08309, over 4835.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2623, pruned_loss=0.06897, over 950091.05 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:42:37,513 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.793e+02 2.145e+02 2.588e+02 4.712e+02, threshold=4.291e+02, percent-clipped=10.0 +2023-03-26 09:42:56,488 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44143.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:42:57,047 INFO [finetune.py:976] (0/7) Epoch 8, batch 4050, loss[loss=0.2046, simple_loss=0.2789, pruned_loss=0.06515, over 4896.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2646, pruned_loss=0.06965, over 948401.83 frames. ], batch size: 35, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:43:29,437 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-26 09:43:31,085 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44170.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:43:49,918 INFO [finetune.py:976] (0/7) Epoch 8, batch 4100, loss[loss=0.2174, simple_loss=0.2791, pruned_loss=0.07784, over 4812.00 frames. ], tot_loss[loss=0.2038, simple_loss=0.2668, pruned_loss=0.07036, over 949707.79 frames. ], batch size: 39, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:43:55,538 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6077, 1.5555, 2.0485, 1.8617, 1.9642, 4.1125, 1.5320, 1.8984], + device='cuda:0'), covar=tensor([0.1008, 0.1786, 0.1238, 0.0997, 0.1477, 0.0266, 0.1432, 0.1703], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0079, 0.0093, 0.0083, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:44:15,224 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4224, 1.3628, 1.8511, 2.8980, 1.9195, 2.0903, 1.0507, 2.2705], + device='cuda:0'), covar=tensor([0.1898, 0.1552, 0.1357, 0.0678, 0.0892, 0.1425, 0.1777, 0.0715], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0136, 0.0167, 0.0103, 0.0140, 0.0128, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 09:44:15,791 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44218.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:44:22,431 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.696e+02 1.895e+02 2.360e+02 4.949e+02, threshold=3.791e+02, percent-clipped=1.0 +2023-03-26 09:44:24,371 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1204, 2.2867, 2.1835, 1.5006, 2.1982, 2.1203, 2.0837, 1.8081], + device='cuda:0'), covar=tensor([0.0676, 0.0547, 0.0700, 0.0925, 0.0547, 0.0730, 0.0702, 0.1107], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0133, 0.0145, 0.0126, 0.0116, 0.0145, 0.0145, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:44:31,521 INFO [finetune.py:976] (0/7) Epoch 8, batch 4150, loss[loss=0.2033, simple_loss=0.2672, pruned_loss=0.06969, over 4864.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2691, pruned_loss=0.07105, over 953592.92 frames. ], batch size: 44, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:44:46,739 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:44:59,962 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 09:45:07,272 INFO [finetune.py:976] (0/7) Epoch 8, batch 4200, loss[loss=0.1915, simple_loss=0.2678, pruned_loss=0.05763, over 4926.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2692, pruned_loss=0.07067, over 954615.71 frames. ], batch size: 42, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:45:30,599 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:45:37,150 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-26 09:45:40,452 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.746e+02 1.975e+02 2.337e+02 5.106e+02, threshold=3.951e+02, percent-clipped=1.0 +2023-03-26 09:45:54,906 INFO [finetune.py:976] (0/7) Epoch 8, batch 4250, loss[loss=0.1908, simple_loss=0.2399, pruned_loss=0.07091, over 4328.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2665, pruned_loss=0.06964, over 954648.92 frames. ], batch size: 19, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:46:32,566 INFO [finetune.py:976] (0/7) Epoch 8, batch 4300, loss[loss=0.1728, simple_loss=0.236, pruned_loss=0.05478, over 4296.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2635, pruned_loss=0.06843, over 956194.68 frames. ], batch size: 65, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:46:56,828 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.550e+02 1.851e+02 2.365e+02 4.860e+02, threshold=3.701e+02, percent-clipped=2.0 +2023-03-26 09:46:58,154 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7482, 1.6646, 2.1302, 1.4972, 1.8180, 2.0750, 1.6608, 2.2125], + device='cuda:0'), covar=tensor([0.1380, 0.2139, 0.1266, 0.1869, 0.0961, 0.1389, 0.2537, 0.0823], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0204, 0.0196, 0.0195, 0.0180, 0.0220, 0.0219, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:47:05,828 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44443.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:47:06,355 INFO [finetune.py:976] (0/7) Epoch 8, batch 4350, loss[loss=0.171, simple_loss=0.2363, pruned_loss=0.05282, over 4835.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2594, pruned_loss=0.06685, over 956965.14 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:47:25,517 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 09:47:34,061 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8449, 3.8563, 3.7297, 1.9740, 4.0154, 3.1129, 0.6980, 2.6747], + device='cuda:0'), covar=tensor([0.2417, 0.1854, 0.1367, 0.3076, 0.0864, 0.0876, 0.4781, 0.1632], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0171, 0.0160, 0.0129, 0.0155, 0.0122, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:47:37,686 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44491.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:47:39,433 INFO [finetune.py:976] (0/7) Epoch 8, batch 4400, loss[loss=0.2311, simple_loss=0.2863, pruned_loss=0.08792, over 4906.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.262, pruned_loss=0.06822, over 954678.46 frames. ], batch size: 37, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:48:17,063 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:48:18,613 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 1.746e+02 1.995e+02 2.515e+02 6.158e+02, threshold=3.991e+02, percent-clipped=2.0 +2023-03-26 09:48:28,728 INFO [finetune.py:976] (0/7) Epoch 8, batch 4450, loss[loss=0.1895, simple_loss=0.265, pruned_loss=0.05699, over 4824.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2659, pruned_loss=0.06964, over 954516.82 frames. ], batch size: 40, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:48:37,032 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:48:41,297 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7757, 1.6750, 1.6800, 1.7755, 1.4052, 3.6162, 1.5179, 2.0217], + device='cuda:0'), covar=tensor([0.3346, 0.2491, 0.2002, 0.2238, 0.1748, 0.0175, 0.2554, 0.1267], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0115, 0.0120, 0.0123, 0.0116, 0.0098, 0.0101, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:48:51,038 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:19,345 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:28,022 INFO [finetune.py:976] (0/7) Epoch 8, batch 4500, loss[loss=0.2159, simple_loss=0.2953, pruned_loss=0.06827, over 4834.00 frames. ], tot_loss[loss=0.2037, simple_loss=0.2676, pruned_loss=0.06995, over 956807.36 frames. ], batch size: 47, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:49:47,691 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44611.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:48,882 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:50,690 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44615.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:54,940 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 09:50:00,004 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.305e+02 1.720e+02 2.026e+02 2.465e+02 5.780e+02, threshold=4.053e+02, percent-clipped=2.0 +2023-03-26 09:50:10,541 INFO [finetune.py:976] (0/7) Epoch 8, batch 4550, loss[loss=0.1884, simple_loss=0.2632, pruned_loss=0.0568, over 4914.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2686, pruned_loss=0.0701, over 955398.35 frames. ], batch size: 38, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:50:12,588 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 09:50:27,737 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44663.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:50:52,807 INFO [finetune.py:976] (0/7) Epoch 8, batch 4600, loss[loss=0.185, simple_loss=0.2435, pruned_loss=0.06322, over 4756.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2673, pruned_loss=0.06911, over 954983.58 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:51:15,457 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.778e+01 1.566e+02 1.839e+02 2.114e+02 3.234e+02, threshold=3.678e+02, percent-clipped=0.0 +2023-03-26 09:51:25,987 INFO [finetune.py:976] (0/7) Epoch 8, batch 4650, loss[loss=0.24, simple_loss=0.2854, pruned_loss=0.09729, over 4937.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2649, pruned_loss=0.0687, over 954939.49 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:51:36,205 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-26 09:51:59,445 INFO [finetune.py:976] (0/7) Epoch 8, batch 4700, loss[loss=0.1564, simple_loss=0.2204, pruned_loss=0.04619, over 4831.00 frames. ], tot_loss[loss=0.199, simple_loss=0.262, pruned_loss=0.06799, over 953862.67 frames. ], batch size: 25, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:52:03,831 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9905, 1.8262, 1.6242, 1.7937, 1.7825, 1.7217, 1.8446, 2.4843], + device='cuda:0'), covar=tensor([0.5193, 0.6412, 0.4238, 0.5345, 0.5098, 0.3028, 0.4934, 0.2095], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0221, 0.0280, 0.0241, 0.0206, 0.0244, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:52:20,535 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6780, 1.4348, 1.9123, 1.9298, 1.6690, 3.5631, 1.4584, 1.6970], + device='cuda:0'), covar=tensor([0.0903, 0.1844, 0.1081, 0.0921, 0.1557, 0.0260, 0.1487, 0.1748], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 09:52:22,745 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.499e+02 1.877e+02 2.319e+02 4.193e+02, threshold=3.754e+02, percent-clipped=1.0 +2023-03-26 09:52:26,402 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2461, 2.0662, 1.8057, 2.1613, 2.2193, 1.9073, 2.5003, 2.2170], + device='cuda:0'), covar=tensor([0.1581, 0.2736, 0.3531, 0.3073, 0.2853, 0.1871, 0.3241, 0.2118], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0189, 0.0233, 0.0254, 0.0236, 0.0194, 0.0211, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 09:52:32,227 INFO [finetune.py:976] (0/7) Epoch 8, batch 4750, loss[loss=0.3114, simple_loss=0.3406, pruned_loss=0.141, over 4127.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2616, pruned_loss=0.06855, over 954363.28 frames. ], batch size: 65, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:52:44,194 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 09:52:58,206 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44883.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:53:05,273 INFO [finetune.py:976] (0/7) Epoch 8, batch 4800, loss[loss=0.236, simple_loss=0.3035, pruned_loss=0.08423, over 4905.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2652, pruned_loss=0.07033, over 951704.38 frames. ], batch size: 37, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:53:15,867 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:53:41,287 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.589e+02 1.932e+02 2.383e+02 4.430e+02, threshold=3.864e+02, percent-clipped=2.0 +2023-03-26 09:53:53,095 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 09:53:55,404 INFO [finetune.py:976] (0/7) Epoch 8, batch 4850, loss[loss=0.2339, simple_loss=0.3018, pruned_loss=0.08303, over 4764.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2686, pruned_loss=0.07156, over 951750.36 frames. ], batch size: 59, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:54:57,808 INFO [finetune.py:976] (0/7) Epoch 8, batch 4900, loss[loss=0.1439, simple_loss=0.2165, pruned_loss=0.03571, over 4818.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2699, pruned_loss=0.0725, over 952606.34 frames. ], batch size: 25, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:55:25,592 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.694e+02 2.008e+02 2.325e+02 4.035e+02, threshold=4.016e+02, percent-clipped=1.0 +2023-03-26 09:55:44,370 INFO [finetune.py:976] (0/7) Epoch 8, batch 4950, loss[loss=0.2314, simple_loss=0.3001, pruned_loss=0.08137, over 4815.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2711, pruned_loss=0.07274, over 952106.61 frames. ], batch size: 39, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:55:57,050 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:13,217 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:21,541 INFO [finetune.py:976] (0/7) Epoch 8, batch 5000, loss[loss=0.2085, simple_loss=0.2587, pruned_loss=0.0792, over 4881.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2692, pruned_loss=0.0719, over 950045.44 frames. ], batch size: 32, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:56:36,686 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 09:56:37,077 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45117.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:56:45,212 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.621e+02 1.919e+02 2.483e+02 3.797e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-03-26 09:56:45,361 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8704, 1.7534, 1.6440, 1.6874, 2.1144, 2.0938, 1.9654, 1.6154], + device='cuda:0'), covar=tensor([0.0314, 0.0294, 0.0448, 0.0296, 0.0230, 0.0390, 0.0233, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0109, 0.0139, 0.0115, 0.0102, 0.0100, 0.0091, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.9408e-05, 8.5504e-05, 1.1139e-04, 9.0243e-05, 8.0195e-05, 7.4643e-05, + 6.8467e-05, 8.3696e-05], device='cuda:0') +2023-03-26 09:56:52,671 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:52,748 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 09:56:53,771 INFO [finetune.py:976] (0/7) Epoch 8, batch 5050, loss[loss=0.1735, simple_loss=0.2433, pruned_loss=0.05182, over 4897.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2656, pruned_loss=0.07056, over 951685.06 frames. ], batch size: 35, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:56:53,844 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1285, 4.8538, 4.5926, 2.7646, 4.9971, 3.7346, 0.9456, 3.6495], + device='cuda:0'), covar=tensor([0.2068, 0.1248, 0.1264, 0.2814, 0.0573, 0.0841, 0.4568, 0.1236], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0173, 0.0160, 0.0129, 0.0156, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 09:57:19,952 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45183.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:57:26,575 INFO [finetune.py:976] (0/7) Epoch 8, batch 5100, loss[loss=0.1758, simple_loss=0.2477, pruned_loss=0.05198, over 4757.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2623, pruned_loss=0.06864, over 952659.77 frames. ], batch size: 27, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:57:34,991 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:57:52,407 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-26 09:57:55,113 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.630e+02 1.903e+02 2.262e+02 3.588e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 09:57:55,785 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45231.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:58:04,087 INFO [finetune.py:976] (0/7) Epoch 8, batch 5150, loss[loss=0.1988, simple_loss=0.262, pruned_loss=0.06785, over 4899.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.262, pruned_loss=0.06864, over 953763.87 frames. ], batch size: 36, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:58:11,256 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45254.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:58:40,359 INFO [finetune.py:976] (0/7) Epoch 8, batch 5200, loss[loss=0.186, simple_loss=0.2559, pruned_loss=0.05806, over 4822.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.2645, pruned_loss=0.06918, over 953013.54 frames. ], batch size: 25, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:59:09,678 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.659e+02 1.911e+02 2.326e+02 4.760e+02, threshold=3.822e+02, percent-clipped=1.0 +2023-03-26 09:59:18,777 INFO [finetune.py:976] (0/7) Epoch 8, batch 5250, loss[loss=0.2082, simple_loss=0.2745, pruned_loss=0.07097, over 4908.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2672, pruned_loss=0.06971, over 952263.90 frames. ], batch size: 37, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:59:27,668 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45348.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:00:03,242 INFO [finetune.py:976] (0/7) Epoch 8, batch 5300, loss[loss=0.1968, simple_loss=0.2666, pruned_loss=0.06353, over 4217.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2679, pruned_loss=0.06957, over 952013.93 frames. ], batch size: 65, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:00:16,200 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45409.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:00:18,469 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45412.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:00:30,691 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.516e+02 1.858e+02 2.399e+02 4.469e+02, threshold=3.716e+02, percent-clipped=2.0 +2023-03-26 10:00:39,966 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45437.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:00:49,408 INFO [finetune.py:976] (0/7) Epoch 8, batch 5350, loss[loss=0.2046, simple_loss=0.2543, pruned_loss=0.07745, over 4855.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.268, pruned_loss=0.06943, over 952628.91 frames. ], batch size: 31, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:01:54,085 INFO [finetune.py:976] (0/7) Epoch 8, batch 5400, loss[loss=0.2516, simple_loss=0.2979, pruned_loss=0.1027, over 4705.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2658, pruned_loss=0.06914, over 951571.72 frames. ], batch size: 23, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:01:57,914 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 10:02:02,808 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.2651, 4.5573, 4.7814, 5.0631, 4.9476, 4.7185, 5.3851, 1.5845], + device='cuda:0'), covar=tensor([0.0719, 0.0792, 0.0816, 0.0898, 0.1168, 0.1589, 0.0463, 0.5555], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0242, 0.0277, 0.0294, 0.0333, 0.0283, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:02:06,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4200, 1.4050, 1.6046, 2.4774, 1.6946, 2.1890, 0.8857, 1.9903], + device='cuda:0'), covar=tensor([0.1846, 0.1554, 0.1294, 0.0832, 0.0985, 0.1127, 0.1753, 0.0838], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0103, 0.0140, 0.0128, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 10:02:34,313 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.640e+02 2.019e+02 2.439e+02 4.086e+02, threshold=4.037e+02, percent-clipped=1.0 +2023-03-26 10:02:54,208 INFO [finetune.py:976] (0/7) Epoch 8, batch 5450, loss[loss=0.2284, simple_loss=0.2893, pruned_loss=0.08373, over 4819.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2636, pruned_loss=0.06874, over 954524.19 frames. ], batch size: 30, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:03:20,294 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-03-26 10:03:31,604 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-03-26 10:03:54,238 INFO [finetune.py:976] (0/7) Epoch 8, batch 5500, loss[loss=0.169, simple_loss=0.2495, pruned_loss=0.04422, over 4806.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2607, pruned_loss=0.06766, over 955164.83 frames. ], batch size: 45, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:03:54,438 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 10:04:14,825 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9768, 1.4030, 1.9426, 1.8666, 1.6338, 1.6010, 1.7399, 1.7816], + device='cuda:0'), covar=tensor([0.4297, 0.4895, 0.4084, 0.4729, 0.5675, 0.4222, 0.5567, 0.3950], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0240, 0.0253, 0.0255, 0.0246, 0.0223, 0.0272, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:04:18,373 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.550e+02 1.829e+02 2.210e+02 3.729e+02, threshold=3.658e+02, percent-clipped=0.0 +2023-03-26 10:04:28,410 INFO [finetune.py:976] (0/7) Epoch 8, batch 5550, loss[loss=0.1862, simple_loss=0.2487, pruned_loss=0.06184, over 4752.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2627, pruned_loss=0.06877, over 954157.76 frames. ], batch size: 26, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:04:38,175 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7130, 1.4151, 1.9201, 1.2966, 1.7238, 1.9124, 1.4407, 2.0481], + device='cuda:0'), covar=tensor([0.1308, 0.2223, 0.1257, 0.1894, 0.0888, 0.1390, 0.2734, 0.0763], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0206, 0.0197, 0.0196, 0.0182, 0.0220, 0.0219, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:05:19,423 INFO [finetune.py:976] (0/7) Epoch 8, batch 5600, loss[loss=0.2421, simple_loss=0.3061, pruned_loss=0.08902, over 4822.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2665, pruned_loss=0.06968, over 951981.86 frames. ], batch size: 33, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:05:19,642 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-26 10:05:24,101 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45702.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:05:25,256 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45704.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:05:25,847 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2723, 1.2620, 1.6042, 2.4428, 1.6154, 2.1478, 0.7747, 1.9887], + device='cuda:0'), covar=tensor([0.1765, 0.1533, 0.1141, 0.0677, 0.0940, 0.1100, 0.1719, 0.0685], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0119, 0.0136, 0.0168, 0.0104, 0.0141, 0.0128, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 10:05:29,905 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45712.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:05:29,942 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4873, 2.1461, 1.7784, 0.8233, 2.0340, 1.8082, 1.5599, 1.8553], + device='cuda:0'), covar=tensor([0.0815, 0.0999, 0.1803, 0.2314, 0.1750, 0.2550, 0.2708, 0.1311], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0199, 0.0201, 0.0188, 0.0216, 0.0205, 0.0221, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:05:40,353 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.739e+02 1.993e+02 2.505e+02 5.014e+02, threshold=3.987e+02, percent-clipped=3.0 +2023-03-26 10:05:44,505 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:05:48,901 INFO [finetune.py:976] (0/7) Epoch 8, batch 5650, loss[loss=0.2285, simple_loss=0.3125, pruned_loss=0.07225, over 4918.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2706, pruned_loss=0.07086, over 953207.55 frames. ], batch size: 42, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:05:58,592 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:00,381 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:06,756 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:13,161 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45785.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:18,811 INFO [finetune.py:976] (0/7) Epoch 8, batch 5700, loss[loss=0.2206, simple_loss=0.2494, pruned_loss=0.09592, over 4283.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2677, pruned_loss=0.07104, over 936614.27 frames. ], batch size: 18, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:06:35,816 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-8.pt +2023-03-26 10:06:54,872 INFO [finetune.py:976] (0/7) Epoch 9, batch 0, loss[loss=0.2014, simple_loss=0.2674, pruned_loss=0.06766, over 4846.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2674, pruned_loss=0.06766, over 4846.00 frames. ], batch size: 44, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:06:54,874 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 10:07:00,271 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7902, 1.5710, 2.0624, 2.9059, 2.0627, 2.3365, 0.9802, 2.3578], + device='cuda:0'), covar=tensor([0.1753, 0.1514, 0.1141, 0.0653, 0.0922, 0.1187, 0.1843, 0.0720], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0103, 0.0140, 0.0127, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:07:11,029 INFO [finetune.py:1010] (0/7) Epoch 9, validation: loss=0.1616, simple_loss=0.233, pruned_loss=0.04515, over 2265189.00 frames. +2023-03-26 10:07:11,029 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 10:07:11,221 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-26 10:07:17,602 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.132e+01 1.600e+02 1.914e+02 2.307e+02 4.538e+02, threshold=3.829e+02, percent-clipped=2.0 +2023-03-26 10:07:22,765 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45835.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:07:42,389 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6456, 3.2253, 3.3687, 3.3700, 3.2075, 3.0800, 3.7242, 1.2837], + device='cuda:0'), covar=tensor([0.1197, 0.1639, 0.1524, 0.1586, 0.2108, 0.2378, 0.1487, 0.6837], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0243, 0.0279, 0.0295, 0.0334, 0.0283, 0.0304, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:07:46,528 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:07:55,655 INFO [finetune.py:976] (0/7) Epoch 9, batch 50, loss[loss=0.2227, simple_loss=0.2846, pruned_loss=0.08036, over 4808.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2718, pruned_loss=0.07158, over 217051.39 frames. ], batch size: 39, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:08:35,059 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6610, 1.5213, 2.0113, 2.9617, 2.0595, 2.2297, 0.9521, 2.3353], + device='cuda:0'), covar=tensor([0.1630, 0.1361, 0.1133, 0.0541, 0.0803, 0.1228, 0.1748, 0.0593], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0102, 0.0139, 0.0127, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:08:35,698 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:08:36,814 INFO [finetune.py:976] (0/7) Epoch 9, batch 100, loss[loss=0.1593, simple_loss=0.2278, pruned_loss=0.04538, over 4846.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2659, pruned_loss=0.06942, over 380385.22 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:08:42,573 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.760e+02 2.008e+02 2.423e+02 3.807e+02, threshold=4.016e+02, percent-clipped=0.0 +2023-03-26 10:08:42,864 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 10:09:10,374 INFO [finetune.py:976] (0/7) Epoch 9, batch 150, loss[loss=0.1787, simple_loss=0.2465, pruned_loss=0.05541, over 4824.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.259, pruned_loss=0.0672, over 508456.10 frames. ], batch size: 39, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:09:28,073 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-46000.pt +2023-03-26 10:09:32,035 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46004.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:09:49,161 INFO [finetune.py:976] (0/7) Epoch 9, batch 200, loss[loss=0.1363, simple_loss=0.2032, pruned_loss=0.03472, over 4778.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2587, pruned_loss=0.06787, over 608108.79 frames. ], batch size: 28, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:09:58,660 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.678e+02 2.056e+02 2.461e+02 4.455e+02, threshold=4.113e+02, percent-clipped=4.0 +2023-03-26 10:10:04,106 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-03-26 10:10:22,912 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46052.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:10:26,558 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:10:36,486 INFO [finetune.py:976] (0/7) Epoch 9, batch 250, loss[loss=0.1535, simple_loss=0.2326, pruned_loss=0.0372, over 4739.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2638, pruned_loss=0.07012, over 683305.15 frames. ], batch size: 54, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:11:09,155 INFO [finetune.py:976] (0/7) Epoch 9, batch 300, loss[loss=0.2053, simple_loss=0.2708, pruned_loss=0.06985, over 4804.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2681, pruned_loss=0.07055, over 744904.79 frames. ], batch size: 41, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:11:14,967 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.304e+02 1.737e+02 2.021e+02 2.354e+02 3.684e+02, threshold=4.042e+02, percent-clipped=0.0 +2023-03-26 10:11:15,053 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:11:16,896 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3689, 3.7210, 3.9635, 4.1834, 4.1280, 3.9060, 4.4579, 1.3678], + device='cuda:0'), covar=tensor([0.0764, 0.0831, 0.0736, 0.1051, 0.1204, 0.1361, 0.0595, 0.5282], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0243, 0.0278, 0.0294, 0.0333, 0.0282, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:11:39,643 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9375, 1.7347, 2.1809, 1.3818, 2.0108, 2.1998, 1.7653, 2.3746], + device='cuda:0'), covar=tensor([0.1597, 0.2213, 0.1423, 0.2345, 0.1052, 0.1636, 0.2757, 0.1038], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0205, 0.0198, 0.0196, 0.0182, 0.0220, 0.0220, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:11:41,959 INFO [finetune.py:976] (0/7) Epoch 9, batch 350, loss[loss=0.321, simple_loss=0.356, pruned_loss=0.143, over 4926.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2711, pruned_loss=0.07199, over 790618.35 frames. ], batch size: 42, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:12:11,361 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:12:17,566 INFO [finetune.py:976] (0/7) Epoch 9, batch 400, loss[loss=0.1928, simple_loss=0.2567, pruned_loss=0.06441, over 4828.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2711, pruned_loss=0.07152, over 826066.92 frames. ], batch size: 49, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:12:23,440 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.951e+01 1.676e+02 2.053e+02 2.418e+02 4.627e+02, threshold=4.106e+02, percent-clipped=2.0 +2023-03-26 10:12:34,729 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-03-26 10:12:40,492 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5997, 1.4677, 1.6673, 1.7999, 1.5554, 3.4535, 1.3357, 1.5833], + device='cuda:0'), covar=tensor([0.0952, 0.1914, 0.1115, 0.1013, 0.1636, 0.0228, 0.1584, 0.1816], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0079, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 10:13:00,669 INFO [finetune.py:976] (0/7) Epoch 9, batch 450, loss[loss=0.1358, simple_loss=0.2073, pruned_loss=0.03212, over 4929.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2685, pruned_loss=0.06994, over 854243.77 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:13:03,228 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:13:35,969 INFO [finetune.py:976] (0/7) Epoch 9, batch 500, loss[loss=0.1511, simple_loss=0.217, pruned_loss=0.0426, over 4748.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2652, pruned_loss=0.0687, over 878102.10 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:13:45,335 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.586e+02 1.900e+02 2.408e+02 3.619e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-03-26 10:13:54,711 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:14:00,496 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9200, 1.5911, 2.2644, 1.5611, 2.0099, 2.1730, 1.6176, 2.2905], + device='cuda:0'), covar=tensor([0.1003, 0.1656, 0.1069, 0.1505, 0.0665, 0.1021, 0.2375, 0.0625], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0205, 0.0198, 0.0195, 0.0182, 0.0219, 0.0218, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:14:04,129 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9874, 1.3731, 1.0261, 1.8143, 2.3303, 1.5320, 1.7265, 1.7699], + device='cuda:0'), covar=tensor([0.1434, 0.2133, 0.1933, 0.1123, 0.1832, 0.1902, 0.1416, 0.2043], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0092, 0.0123, 0.0095, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:14:08,372 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:14:10,974 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-03-26 10:14:17,260 INFO [finetune.py:976] (0/7) Epoch 9, batch 550, loss[loss=0.2298, simple_loss=0.2902, pruned_loss=0.08468, over 4812.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2629, pruned_loss=0.0683, over 895768.26 frames. ], batch size: 40, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:14:21,048 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2192, 1.9522, 1.7896, 2.1060, 2.0281, 1.9202, 1.8725, 2.9261], + device='cuda:0'), covar=tensor([0.4920, 0.6973, 0.4308, 0.5879, 0.5411, 0.2923, 0.6115, 0.1912], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0221, 0.0279, 0.0242, 0.0207, 0.0244, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:14:38,821 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4473, 1.4337, 1.2918, 1.4158, 1.8586, 1.7191, 1.5687, 1.2766], + device='cuda:0'), covar=tensor([0.0305, 0.0289, 0.0522, 0.0271, 0.0195, 0.0420, 0.0249, 0.0373], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0109, 0.0138, 0.0114, 0.0102, 0.0100, 0.0090, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.9347e-05, 8.5481e-05, 1.1036e-04, 8.9889e-05, 7.9752e-05, 7.4318e-05, + 6.7968e-05, 8.3119e-05], device='cuda:0') +2023-03-26 10:14:39,972 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46406.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:14:50,106 INFO [finetune.py:976] (0/7) Epoch 9, batch 600, loss[loss=0.2492, simple_loss=0.3222, pruned_loss=0.08804, over 4806.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2649, pruned_loss=0.06979, over 908192.17 frames. ], batch size: 45, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:14:54,846 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.702e+02 1.969e+02 2.390e+02 4.680e+02, threshold=3.938e+02, percent-clipped=3.0 +2023-03-26 10:14:54,934 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46430.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:15:36,213 INFO [finetune.py:976] (0/7) Epoch 9, batch 650, loss[loss=0.2412, simple_loss=0.2853, pruned_loss=0.0985, over 4911.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2682, pruned_loss=0.07158, over 919855.27 frames. ], batch size: 37, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:15:40,462 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:16:05,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:16:09,544 INFO [finetune.py:976] (0/7) Epoch 9, batch 700, loss[loss=0.2703, simple_loss=0.3152, pruned_loss=0.1127, over 4922.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2699, pruned_loss=0.07151, over 928560.46 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:16:14,890 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.666e+02 2.010e+02 2.529e+02 4.289e+02, threshold=4.019e+02, percent-clipped=2.0 +2023-03-26 10:16:23,826 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1712, 3.5644, 3.7964, 4.0015, 3.9279, 3.6997, 4.2659, 1.4460], + device='cuda:0'), covar=tensor([0.0802, 0.0828, 0.0741, 0.0921, 0.1146, 0.1311, 0.0601, 0.4843], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0240, 0.0273, 0.0289, 0.0328, 0.0278, 0.0298, 0.0290], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:16:37,471 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:16:38,159 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3106, 2.0859, 1.7569, 2.2914, 2.2212, 1.8794, 2.5662, 2.2105], + device='cuda:0'), covar=tensor([0.1551, 0.2718, 0.3584, 0.3066, 0.2826, 0.2025, 0.3408, 0.2211], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0188, 0.0232, 0.0253, 0.0236, 0.0193, 0.0211, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:16:42,850 INFO [finetune.py:976] (0/7) Epoch 9, batch 750, loss[loss=0.1775, simple_loss=0.2504, pruned_loss=0.05228, over 4813.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2712, pruned_loss=0.07122, over 935887.34 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:16:58,323 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:03,679 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:06,096 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 10:17:15,992 INFO [finetune.py:976] (0/7) Epoch 9, batch 800, loss[loss=0.1497, simple_loss=0.2025, pruned_loss=0.04842, over 4285.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.271, pruned_loss=0.07099, over 940233.69 frames. ], batch size: 18, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:17:19,132 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5304, 1.3432, 1.3741, 0.8000, 1.3975, 1.5431, 1.6121, 1.2859], + device='cuda:0'), covar=tensor([0.0792, 0.0600, 0.0459, 0.0504, 0.0480, 0.0528, 0.0276, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0156, 0.0121, 0.0135, 0.0132, 0.0126, 0.0146, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.6164e-05, 1.1502e-04, 8.7378e-05, 9.8252e-05, 9.4215e-05, 9.2031e-05, + 1.0689e-04, 1.0834e-04], device='cuda:0') +2023-03-26 10:17:20,814 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.216e+01 1.561e+02 1.863e+02 2.153e+02 3.377e+02, threshold=3.726e+02, percent-clipped=0.0 +2023-03-26 10:17:22,502 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46632.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:39,120 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46656.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:17:43,748 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46663.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:49,096 INFO [finetune.py:976] (0/7) Epoch 9, batch 850, loss[loss=0.2244, simple_loss=0.2843, pruned_loss=0.08222, over 4819.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2675, pruned_loss=0.06962, over 943041.83 frames. ], batch size: 41, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:18:34,826 INFO [finetune.py:976] (0/7) Epoch 9, batch 900, loss[loss=0.2062, simple_loss=0.2544, pruned_loss=0.079, over 4850.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2637, pruned_loss=0.06809, over 947284.27 frames. ], batch size: 49, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:18:35,638 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 10:18:39,659 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.365e+01 1.501e+02 1.768e+02 2.130e+02 3.855e+02, threshold=3.537e+02, percent-clipped=1.0 +2023-03-26 10:18:45,679 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6231, 1.5263, 1.5039, 1.6318, 1.1102, 3.3722, 1.3362, 1.8229], + device='cuda:0'), covar=tensor([0.3446, 0.2496, 0.2175, 0.2346, 0.1946, 0.0206, 0.2709, 0.1334], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 10:19:10,162 INFO [finetune.py:976] (0/7) Epoch 9, batch 950, loss[loss=0.2549, simple_loss=0.3119, pruned_loss=0.09899, over 4713.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2629, pruned_loss=0.06789, over 950329.11 frames. ], batch size: 59, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:19:15,858 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 10:19:36,009 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:19:44,264 INFO [finetune.py:976] (0/7) Epoch 9, batch 1000, loss[loss=0.2135, simple_loss=0.2785, pruned_loss=0.07431, over 4895.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2647, pruned_loss=0.06851, over 953161.70 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:19:49,056 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.690e+02 1.992e+02 2.479e+02 5.334e+02, threshold=3.983e+02, percent-clipped=2.0 +2023-03-26 10:20:10,930 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2057, 2.1904, 2.0000, 1.7077, 2.6279, 2.7671, 2.3392, 2.1146], + device='cuda:0'), covar=tensor([0.0368, 0.0343, 0.0487, 0.0379, 0.0233, 0.0465, 0.0379, 0.0368], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0108, 0.0137, 0.0114, 0.0101, 0.0099, 0.0090, 0.0107], + device='cuda:0'), out_proj_covar=tensor([6.9084e-05, 8.4652e-05, 1.0983e-04, 8.9317e-05, 7.9360e-05, 7.3673e-05, + 6.8017e-05, 8.2627e-05], device='cuda:0') +2023-03-26 10:20:22,562 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:20:23,062 INFO [finetune.py:976] (0/7) Epoch 9, batch 1050, loss[loss=0.1953, simple_loss=0.268, pruned_loss=0.06126, over 4869.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2661, pruned_loss=0.06854, over 953695.06 frames. ], batch size: 34, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:20:51,211 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6726, 3.6294, 3.4155, 1.4818, 3.7003, 2.8696, 0.7948, 2.5076], + device='cuda:0'), covar=tensor([0.2440, 0.1622, 0.1483, 0.3360, 0.1043, 0.0942, 0.4119, 0.1296], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0172, 0.0160, 0.0128, 0.0156, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 10:21:05,008 INFO [finetune.py:976] (0/7) Epoch 9, batch 1100, loss[loss=0.2145, simple_loss=0.285, pruned_loss=0.07196, over 4914.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2675, pruned_loss=0.06927, over 953982.80 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:21:10,486 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.654e+02 2.044e+02 2.534e+02 3.814e+02, threshold=4.089e+02, percent-clipped=0.0 +2023-03-26 10:21:11,159 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46932.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:21:23,030 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46951.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 10:21:28,244 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46958.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:21:37,702 INFO [finetune.py:976] (0/7) Epoch 9, batch 1150, loss[loss=0.1954, simple_loss=0.2625, pruned_loss=0.06413, over 4780.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.27, pruned_loss=0.0705, over 953539.06 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:21:42,585 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:21:44,423 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9246, 3.4054, 3.6029, 3.7997, 3.6881, 3.3965, 3.9612, 1.2266], + device='cuda:0'), covar=tensor([0.0863, 0.0802, 0.0838, 0.1017, 0.1324, 0.1613, 0.0776, 0.5158], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0242, 0.0275, 0.0293, 0.0331, 0.0281, 0.0300, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:22:10,619 INFO [finetune.py:976] (0/7) Epoch 9, batch 1200, loss[loss=0.1957, simple_loss=0.2646, pruned_loss=0.06337, over 4789.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2678, pruned_loss=0.06935, over 954118.82 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:22:16,138 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.814e+01 1.625e+02 2.018e+02 2.406e+02 3.989e+02, threshold=4.036e+02, percent-clipped=0.0 +2023-03-26 10:22:43,589 INFO [finetune.py:976] (0/7) Epoch 9, batch 1250, loss[loss=0.2408, simple_loss=0.2909, pruned_loss=0.09533, over 4858.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2653, pruned_loss=0.06918, over 955294.73 frames. ], batch size: 49, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:22:43,846 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 10:22:50,906 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.98 vs. limit=5.0 +2023-03-26 10:23:21,438 INFO [finetune.py:976] (0/7) Epoch 9, batch 1300, loss[loss=0.1719, simple_loss=0.2412, pruned_loss=0.05134, over 4915.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.2617, pruned_loss=0.06792, over 955743.20 frames. ], batch size: 37, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:23:31,787 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.169e+01 1.658e+02 1.956e+02 2.404e+02 5.414e+02, threshold=3.912e+02, percent-clipped=2.0 +2023-03-26 10:23:57,902 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:24:03,031 INFO [finetune.py:976] (0/7) Epoch 9, batch 1350, loss[loss=0.2213, simple_loss=0.291, pruned_loss=0.07577, over 4702.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2639, pruned_loss=0.06937, over 955450.30 frames. ], batch size: 59, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:24:36,401 INFO [finetune.py:976] (0/7) Epoch 9, batch 1400, loss[loss=0.1839, simple_loss=0.2532, pruned_loss=0.05732, over 4765.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2667, pruned_loss=0.07128, over 953504.61 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:24:42,786 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.739e+02 2.002e+02 2.347e+02 4.228e+02, threshold=4.005e+02, percent-clipped=2.0 +2023-03-26 10:24:49,034 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:24:55,130 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:24:59,389 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47258.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:00,755 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 10:25:09,255 INFO [finetune.py:976] (0/7) Epoch 9, batch 1450, loss[loss=0.1907, simple_loss=0.2599, pruned_loss=0.06078, over 4751.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2694, pruned_loss=0.07189, over 954892.99 frames. ], batch size: 26, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:25:27,569 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:28,776 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6884, 3.4528, 3.3526, 1.5927, 3.4891, 2.7254, 0.9521, 2.4357], + device='cuda:0'), covar=tensor([0.2965, 0.1704, 0.1622, 0.3283, 0.1238, 0.1044, 0.4134, 0.1469], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0173, 0.0160, 0.0129, 0.0156, 0.0122, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 10:25:29,400 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:36,197 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:53,873 INFO [finetune.py:976] (0/7) Epoch 9, batch 1500, loss[loss=0.2031, simple_loss=0.2746, pruned_loss=0.06576, over 4886.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.27, pruned_loss=0.07134, over 954386.67 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:25:54,704 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 10:26:00,802 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.648e+02 1.983e+02 2.305e+02 4.092e+02, threshold=3.967e+02, percent-clipped=1.0 +2023-03-26 10:26:26,876 INFO [finetune.py:976] (0/7) Epoch 9, batch 1550, loss[loss=0.1523, simple_loss=0.2182, pruned_loss=0.04316, over 4790.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2691, pruned_loss=0.07049, over 954018.86 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:26:33,187 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:26:33,891 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 10:26:52,406 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9691, 1.7396, 1.4524, 1.5988, 1.7364, 1.6836, 1.6943, 2.4487], + device='cuda:0'), covar=tensor([0.4967, 0.5011, 0.4144, 0.4795, 0.5006, 0.3142, 0.4714, 0.2170], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0221, 0.0278, 0.0241, 0.0207, 0.0243, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:27:00,837 INFO [finetune.py:976] (0/7) Epoch 9, batch 1600, loss[loss=0.1946, simple_loss=0.2504, pruned_loss=0.06944, over 4696.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2656, pruned_loss=0.06908, over 954611.05 frames. ], batch size: 23, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:27:07,814 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.632e+02 1.991e+02 2.321e+02 5.028e+02, threshold=3.982e+02, percent-clipped=1.0 +2023-03-26 10:27:14,443 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:27:22,838 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6990, 1.7636, 1.8198, 1.1232, 1.9398, 1.9092, 1.8469, 1.5744], + device='cuda:0'), covar=tensor([0.0589, 0.0596, 0.0651, 0.0813, 0.0572, 0.0633, 0.0562, 0.0970], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0132, 0.0144, 0.0124, 0.0116, 0.0144, 0.0144, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:27:30,684 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:27:34,229 INFO [finetune.py:976] (0/7) Epoch 9, batch 1650, loss[loss=0.192, simple_loss=0.249, pruned_loss=0.0675, over 4921.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2627, pruned_loss=0.0678, over 954988.93 frames. ], batch size: 37, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:27:39,152 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2931, 2.2021, 1.5563, 2.4441, 2.1962, 1.8522, 3.1765, 2.2866], + device='cuda:0'), covar=tensor([0.1459, 0.2425, 0.3686, 0.3176, 0.2874, 0.1798, 0.2806, 0.2019], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0188, 0.0232, 0.0253, 0.0237, 0.0193, 0.0211, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:27:48,779 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6762, 1.6869, 2.2112, 2.0109, 1.8650, 3.5207, 1.5945, 2.0408], + device='cuda:0'), covar=tensor([0.0908, 0.1534, 0.1347, 0.0875, 0.1290, 0.0298, 0.1228, 0.1362], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0083, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 10:28:02,971 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:28:07,770 INFO [finetune.py:976] (0/7) Epoch 9, batch 1700, loss[loss=0.1488, simple_loss=0.2268, pruned_loss=0.03538, over 4791.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2611, pruned_loss=0.06764, over 954521.00 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:28:13,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.617e+02 1.864e+02 2.209e+02 5.015e+02, threshold=3.728e+02, percent-clipped=2.0 +2023-03-26 10:28:47,729 INFO [finetune.py:976] (0/7) Epoch 9, batch 1750, loss[loss=0.2773, simple_loss=0.3434, pruned_loss=0.1056, over 4843.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2629, pruned_loss=0.06855, over 954632.27 frames. ], batch size: 49, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:29:08,518 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8144, 1.2897, 0.9198, 1.6765, 2.0273, 1.4222, 1.5919, 1.6430], + device='cuda:0'), covar=tensor([0.1504, 0.2185, 0.2098, 0.1232, 0.2115, 0.2061, 0.1495, 0.2125], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0122, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:29:13,706 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47597.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:29:29,376 INFO [finetune.py:976] (0/7) Epoch 9, batch 1800, loss[loss=0.2148, simple_loss=0.2737, pruned_loss=0.07795, over 4793.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2652, pruned_loss=0.06856, over 953833.84 frames. ], batch size: 51, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:29:33,763 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3851, 1.4644, 1.7285, 1.7991, 1.5715, 3.3038, 1.3602, 1.6591], + device='cuda:0'), covar=tensor([0.1018, 0.1835, 0.1343, 0.1020, 0.1565, 0.0277, 0.1517, 0.1720], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 10:29:34,863 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.736e+02 1.994e+02 2.573e+02 6.193e+02, threshold=3.989e+02, percent-clipped=4.0 +2023-03-26 10:29:55,594 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 10:30:02,647 INFO [finetune.py:976] (0/7) Epoch 9, batch 1850, loss[loss=0.2353, simple_loss=0.2907, pruned_loss=0.09, over 4795.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2666, pruned_loss=0.06904, over 953855.44 frames. ], batch size: 51, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:30:06,412 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2128, 1.7840, 1.9922, 0.7275, 2.2013, 2.5673, 1.9346, 1.8759], + device='cuda:0'), covar=tensor([0.1331, 0.1260, 0.0874, 0.0984, 0.0712, 0.0667, 0.0652, 0.0913], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0158, 0.0123, 0.0136, 0.0133, 0.0127, 0.0147, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.6973e-05, 1.1639e-04, 8.8668e-05, 9.9142e-05, 9.4993e-05, 9.2878e-05, + 1.0797e-04, 1.0971e-04], device='cuda:0') +2023-03-26 10:30:35,896 INFO [finetune.py:976] (0/7) Epoch 9, batch 1900, loss[loss=0.2633, simple_loss=0.3329, pruned_loss=0.09685, over 4890.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2682, pruned_loss=0.06947, over 952749.23 frames. ], batch size: 43, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:30:46,261 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.608e+02 1.835e+02 2.236e+02 3.803e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-03-26 10:30:52,997 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:30:54,849 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:21,753 INFO [finetune.py:976] (0/7) Epoch 9, batch 1950, loss[loss=0.1872, simple_loss=0.2621, pruned_loss=0.05609, over 4802.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2668, pruned_loss=0.06852, over 954755.02 frames. ], batch size: 41, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:31:31,533 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:39,143 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:46,691 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47809.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:31:53,946 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1689, 1.9459, 1.6542, 1.8892, 2.0490, 1.7903, 2.3673, 2.1756], + device='cuda:0'), covar=tensor([0.1469, 0.2562, 0.3680, 0.3143, 0.2999, 0.1862, 0.3625, 0.2034], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0187, 0.0231, 0.0252, 0.0236, 0.0193, 0.0210, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:31:55,010 INFO [finetune.py:976] (0/7) Epoch 9, batch 2000, loss[loss=0.2098, simple_loss=0.2583, pruned_loss=0.08063, over 4849.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2631, pruned_loss=0.06756, over 953800.01 frames. ], batch size: 44, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:32:00,456 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.527e+02 1.824e+02 2.186e+02 3.277e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-26 10:32:11,434 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:32:19,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5926, 1.6490, 1.6360, 0.9331, 1.6897, 1.8545, 1.8830, 1.4274], + device='cuda:0'), covar=tensor([0.0909, 0.0579, 0.0388, 0.0588, 0.0413, 0.0521, 0.0319, 0.0629], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0158, 0.0123, 0.0136, 0.0133, 0.0127, 0.0147, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.6767e-05, 1.1602e-04, 8.8707e-05, 9.8883e-05, 9.5187e-05, 9.2750e-05, + 1.0783e-04, 1.0943e-04], device='cuda:0') +2023-03-26 10:32:31,539 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47870.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:32:36,818 INFO [finetune.py:976] (0/7) Epoch 9, batch 2050, loss[loss=0.1826, simple_loss=0.25, pruned_loss=0.05759, over 4755.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2601, pruned_loss=0.06649, over 954931.87 frames. ], batch size: 28, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:32:57,820 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:33:15,823 INFO [finetune.py:976] (0/7) Epoch 9, batch 2100, loss[loss=0.2108, simple_loss=0.2818, pruned_loss=0.0699, over 4909.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2598, pruned_loss=0.06643, over 954310.73 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:33:21,286 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.604e+02 1.917e+02 2.370e+02 5.169e+02, threshold=3.834e+02, percent-clipped=3.0 +2023-03-26 10:33:29,783 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:33:30,021 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 10:33:54,984 INFO [finetune.py:976] (0/7) Epoch 9, batch 2150, loss[loss=0.2975, simple_loss=0.3407, pruned_loss=0.1272, over 4909.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2634, pruned_loss=0.06786, over 954577.87 frames. ], batch size: 35, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:34:27,579 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-48000.pt +2023-03-26 10:35:00,855 INFO [finetune.py:976] (0/7) Epoch 9, batch 2200, loss[loss=0.1936, simple_loss=0.2688, pruned_loss=0.05916, over 4815.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.266, pruned_loss=0.06841, over 952990.27 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:35:08,952 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9542, 1.6350, 2.3046, 1.5171, 2.2075, 2.0785, 1.5804, 2.3259], + device='cuda:0'), covar=tensor([0.1296, 0.1885, 0.1343, 0.1994, 0.0734, 0.1494, 0.2491, 0.0807], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0205, 0.0196, 0.0194, 0.0180, 0.0219, 0.0218, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:35:11,848 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.338e+02 1.754e+02 2.052e+02 2.528e+02 4.321e+02, threshold=4.105e+02, percent-clipped=1.0 +2023-03-26 10:35:18,706 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48035.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:35:52,831 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 10:36:02,992 INFO [finetune.py:976] (0/7) Epoch 9, batch 2250, loss[loss=0.2421, simple_loss=0.2935, pruned_loss=0.09534, over 4912.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.269, pruned_loss=0.07021, over 953511.06 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:36:03,114 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:36:20,637 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:36:32,281 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48094.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:05,451 INFO [finetune.py:976] (0/7) Epoch 9, batch 2300, loss[loss=0.181, simple_loss=0.2523, pruned_loss=0.05485, over 4895.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.269, pruned_loss=0.06992, over 953892.34 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:37:16,654 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.508e+01 1.652e+02 1.874e+02 2.360e+02 5.580e+02, threshold=3.748e+02, percent-clipped=1.0 +2023-03-26 10:37:18,018 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48133.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:34,217 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:54,352 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48165.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:37:57,301 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 10:38:01,014 INFO [finetune.py:976] (0/7) Epoch 9, batch 2350, loss[loss=0.1894, simple_loss=0.244, pruned_loss=0.06736, over 4071.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.266, pruned_loss=0.06913, over 953932.33 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:38:12,397 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7946, 1.5487, 1.4223, 1.2991, 1.5570, 1.5932, 1.5371, 2.1235], + device='cuda:0'), covar=tensor([0.4934, 0.4691, 0.3859, 0.4441, 0.4071, 0.2730, 0.4314, 0.1999], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0221, 0.0279, 0.0242, 0.0208, 0.0244, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:38:15,369 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4752, 1.0683, 0.7816, 1.3589, 2.0286, 0.7449, 1.2804, 1.3997], + device='cuda:0'), covar=tensor([0.1560, 0.2201, 0.1808, 0.1236, 0.1900, 0.2061, 0.1549, 0.2015], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:38:35,577 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48212.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:38:37,423 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3517, 1.5554, 1.3107, 1.4943, 1.7480, 1.7038, 1.4418, 1.3313], + device='cuda:0'), covar=tensor([0.0324, 0.0265, 0.0498, 0.0288, 0.0248, 0.0447, 0.0330, 0.0370], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0110, 0.0140, 0.0115, 0.0103, 0.0101, 0.0092, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.0347e-05, 8.5840e-05, 1.1176e-04, 9.0030e-05, 8.0900e-05, 7.5320e-05, + 6.9535e-05, 8.3321e-05], device='cuda:0') +2023-03-26 10:38:43,033 INFO [finetune.py:976] (0/7) Epoch 9, batch 2400, loss[loss=0.1811, simple_loss=0.2483, pruned_loss=0.05696, over 4907.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2626, pruned_loss=0.06821, over 955475.18 frames. ], batch size: 46, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:38:43,255 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 10:38:49,468 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.602e+02 2.015e+02 2.421e+02 3.465e+02, threshold=4.031e+02, percent-clipped=0.0 +2023-03-26 10:39:18,905 INFO [finetune.py:976] (0/7) Epoch 9, batch 2450, loss[loss=0.2206, simple_loss=0.2934, pruned_loss=0.07396, over 4808.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2614, pruned_loss=0.06789, over 954455.07 frames. ], batch size: 45, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:39:19,241 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 10:39:19,625 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48273.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:40:01,691 INFO [finetune.py:976] (0/7) Epoch 9, batch 2500, loss[loss=0.2759, simple_loss=0.338, pruned_loss=0.1069, over 4842.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.263, pruned_loss=0.06843, over 952714.94 frames. ], batch size: 47, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:40:03,382 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:09,017 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.681e+02 1.962e+02 2.420e+02 5.026e+02, threshold=3.923e+02, percent-clipped=2.0 +2023-03-26 10:40:35,356 INFO [finetune.py:976] (0/7) Epoch 9, batch 2550, loss[loss=0.1791, simple_loss=0.2477, pruned_loss=0.05521, over 4826.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2662, pruned_loss=0.06934, over 951862.61 frames. ], batch size: 51, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:40:45,878 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:51,851 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:53,637 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:08,895 INFO [finetune.py:976] (0/7) Epoch 9, batch 2600, loss[loss=0.2393, simple_loss=0.2942, pruned_loss=0.09215, over 4907.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2675, pruned_loss=0.06959, over 953104.93 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:41:09,600 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:12,597 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:15,191 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.754e+02 2.041e+02 2.553e+02 4.015e+02, threshold=4.083e+02, percent-clipped=1.0 +2023-03-26 10:41:23,746 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48442.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:25,437 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:33,998 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:38,220 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48465.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:41:42,379 INFO [finetune.py:976] (0/7) Epoch 9, batch 2650, loss[loss=0.2022, simple_loss=0.2694, pruned_loss=0.06749, over 4881.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2698, pruned_loss=0.07063, over 955186.09 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:41:56,006 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48484.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:42:06,174 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:42:19,471 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48513.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:42:24,896 INFO [finetune.py:976] (0/7) Epoch 9, batch 2700, loss[loss=0.2017, simple_loss=0.2667, pruned_loss=0.06831, over 4859.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2687, pruned_loss=0.07002, over 955255.45 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:42:30,333 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.590e+02 1.873e+02 2.487e+02 4.580e+02, threshold=3.745e+02, percent-clipped=2.0 +2023-03-26 10:43:07,959 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48568.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:43:10,318 INFO [finetune.py:976] (0/7) Epoch 9, batch 2750, loss[loss=0.2255, simple_loss=0.2691, pruned_loss=0.09098, over 4847.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2656, pruned_loss=0.06921, over 956296.66 frames. ], batch size: 44, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:43:15,400 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 10:43:24,184 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 10:43:45,668 INFO [finetune.py:976] (0/7) Epoch 9, batch 2800, loss[loss=0.192, simple_loss=0.2565, pruned_loss=0.06372, over 4913.00 frames. ], tot_loss[loss=0.199, simple_loss=0.262, pruned_loss=0.06804, over 954565.76 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:43:48,261 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-26 10:43:51,108 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.496e+02 1.797e+02 2.211e+02 4.995e+02, threshold=3.593e+02, percent-clipped=1.0 +2023-03-26 10:44:19,114 INFO [finetune.py:976] (0/7) Epoch 9, batch 2850, loss[loss=0.2278, simple_loss=0.2954, pruned_loss=0.08009, over 4800.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2609, pruned_loss=0.06776, over 954075.98 frames. ], batch size: 51, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:44:24,042 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:44:29,598 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9417, 1.0814, 1.8508, 1.8125, 1.7008, 1.6109, 1.6548, 1.7300], + device='cuda:0'), covar=tensor([0.3922, 0.4942, 0.4170, 0.4432, 0.5476, 0.4253, 0.5298, 0.3897], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0240, 0.0254, 0.0255, 0.0249, 0.0224, 0.0274, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:44:39,826 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9077, 1.3227, 0.7232, 1.8735, 2.3148, 1.6853, 1.5567, 1.9241], + device='cuda:0'), covar=tensor([0.1433, 0.2041, 0.2293, 0.1126, 0.1894, 0.2140, 0.1427, 0.1837], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0091, 0.0121, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:44:57,264 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8974, 1.7322, 1.6604, 1.9953, 2.3210, 1.9245, 1.6687, 1.6391], + device='cuda:0'), covar=tensor([0.1926, 0.2063, 0.1732, 0.1560, 0.1787, 0.1122, 0.2461, 0.1625], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0208, 0.0205, 0.0187, 0.0240, 0.0179, 0.0214, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:45:04,573 INFO [finetune.py:976] (0/7) Epoch 9, batch 2900, loss[loss=0.2802, simple_loss=0.331, pruned_loss=0.1147, over 4161.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2648, pruned_loss=0.069, over 954013.02 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:45:08,333 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48728.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:10,029 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.660e+02 1.926e+02 2.355e+02 4.281e+02, threshold=3.853e+02, percent-clipped=2.0 +2023-03-26 10:45:25,492 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:30,844 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6211, 0.6439, 1.6333, 1.5175, 1.4242, 1.3744, 1.4189, 1.5086], + device='cuda:0'), covar=tensor([0.4122, 0.4875, 0.4087, 0.4077, 0.5267, 0.3956, 0.4803, 0.3749], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0239, 0.0253, 0.0255, 0.0249, 0.0224, 0.0273, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:45:38,483 INFO [finetune.py:976] (0/7) Epoch 9, batch 2950, loss[loss=0.1768, simple_loss=0.2534, pruned_loss=0.05009, over 4903.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2668, pruned_loss=0.0695, over 952520.16 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:45:40,982 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:42,826 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48779.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:46:11,307 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:46:11,795 INFO [finetune.py:976] (0/7) Epoch 9, batch 3000, loss[loss=0.2085, simple_loss=0.2612, pruned_loss=0.07786, over 4168.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2674, pruned_loss=0.06973, over 953662.13 frames. ], batch size: 18, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:46:11,796 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 10:46:22,397 INFO [finetune.py:1010] (0/7) Epoch 9, validation: loss=0.159, simple_loss=0.2302, pruned_loss=0.04393, over 2265189.00 frames. +2023-03-26 10:46:22,397 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 10:46:27,908 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.631e+02 1.894e+02 2.277e+02 3.777e+02, threshold=3.789e+02, percent-clipped=0.0 +2023-03-26 10:46:32,266 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48838.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:46:51,704 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48868.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:46:54,487 INFO [finetune.py:976] (0/7) Epoch 9, batch 3050, loss[loss=0.1471, simple_loss=0.2183, pruned_loss=0.03795, over 4748.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2682, pruned_loss=0.06982, over 954484.74 frames. ], batch size: 27, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:47:03,419 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48882.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:47:13,619 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:47:24,761 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48916.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:47:25,839 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8468, 3.4116, 3.5352, 3.7497, 3.5880, 3.3034, 3.9168, 1.2203], + device='cuda:0'), covar=tensor([0.1013, 0.0871, 0.0982, 0.1145, 0.1496, 0.1874, 0.0906, 0.5259], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0291, 0.0330, 0.0281, 0.0300, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:47:29,279 INFO [finetune.py:976] (0/7) Epoch 9, batch 3100, loss[loss=0.172, simple_loss=0.2413, pruned_loss=0.05134, over 4745.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2668, pruned_loss=0.06939, over 956076.75 frames. ], batch size: 27, lr: 3.79e-03, grad_scale: 32.0 +2023-03-26 10:47:36,135 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.624e+02 1.916e+02 2.206e+02 4.881e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-26 10:48:07,637 INFO [finetune.py:976] (0/7) Epoch 9, batch 3150, loss[loss=0.1955, simple_loss=0.2643, pruned_loss=0.06333, over 4827.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2631, pruned_loss=0.06806, over 956496.12 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:48:16,688 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:48:51,010 INFO [finetune.py:976] (0/7) Epoch 9, batch 3200, loss[loss=0.1883, simple_loss=0.2534, pruned_loss=0.06157, over 4907.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2595, pruned_loss=0.06689, over 957080.09 frames. ], batch size: 35, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:48:55,657 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:48:57,863 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.665e+02 1.973e+02 2.326e+02 6.022e+02, threshold=3.945e+02, percent-clipped=4.0 +2023-03-26 10:48:58,699 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 10:49:12,707 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49053.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:49:30,123 INFO [finetune.py:976] (0/7) Epoch 9, batch 3250, loss[loss=0.1609, simple_loss=0.2297, pruned_loss=0.04603, over 4727.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2601, pruned_loss=0.06687, over 956506.23 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:49:40,158 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49079.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:05,715 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:21,993 INFO [finetune.py:976] (0/7) Epoch 9, batch 3300, loss[loss=0.2619, simple_loss=0.3122, pruned_loss=0.1058, over 4792.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2649, pruned_loss=0.06881, over 957007.19 frames. ], batch size: 51, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:50:26,598 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49127.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:28,940 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.707e+02 1.914e+02 2.346e+02 3.542e+02, threshold=3.827e+02, percent-clipped=0.0 +2023-03-26 10:50:56,004 INFO [finetune.py:976] (0/7) Epoch 9, batch 3350, loss[loss=0.2473, simple_loss=0.3053, pruned_loss=0.09462, over 4755.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2674, pruned_loss=0.07013, over 953112.62 frames. ], batch size: 54, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:50:57,463 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 10:50:59,108 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49177.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:51:09,617 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 10:51:10,665 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9030, 1.2797, 1.8076, 1.8240, 1.6754, 1.5904, 1.7668, 1.6701], + device='cuda:0'), covar=tensor([0.4578, 0.5219, 0.4402, 0.4614, 0.5932, 0.4457, 0.5606, 0.4283], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0240, 0.0253, 0.0256, 0.0248, 0.0225, 0.0273, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:51:17,566 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:51:30,249 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9235, 4.6253, 4.3470, 2.6187, 4.6615, 3.5386, 0.8103, 3.2899], + device='cuda:0'), covar=tensor([0.2330, 0.1241, 0.1354, 0.2550, 0.0784, 0.0793, 0.4389, 0.1083], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0173, 0.0159, 0.0127, 0.0155, 0.0121, 0.0146, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 10:51:49,067 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6483, 1.5690, 1.5326, 1.6229, 1.2039, 3.4206, 1.4724, 2.0344], + device='cuda:0'), covar=tensor([0.3223, 0.2355, 0.2011, 0.2234, 0.1733, 0.0172, 0.2802, 0.1233], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 10:51:50,171 INFO [finetune.py:976] (0/7) Epoch 9, batch 3400, loss[loss=0.2149, simple_loss=0.2837, pruned_loss=0.07303, over 4895.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2672, pruned_loss=0.06987, over 952776.52 frames. ], batch size: 43, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:51:59,484 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-26 10:51:59,825 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.353e+01 1.606e+02 1.878e+02 2.295e+02 4.525e+02, threshold=3.756e+02, percent-clipped=2.0 +2023-03-26 10:52:10,013 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:52:55,266 INFO [finetune.py:976] (0/7) Epoch 9, batch 3450, loss[loss=0.2163, simple_loss=0.2685, pruned_loss=0.0821, over 4825.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2668, pruned_loss=0.06931, over 951754.31 frames. ], batch size: 30, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:52:57,285 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8788, 2.1105, 2.0293, 1.3258, 2.1128, 2.1997, 1.9547, 1.7777], + device='cuda:0'), covar=tensor([0.0711, 0.0683, 0.0837, 0.1018, 0.0605, 0.0785, 0.0767, 0.1104], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0130, 0.0142, 0.0123, 0.0116, 0.0142, 0.0142, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:53:28,002 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:53:53,350 INFO [finetune.py:976] (0/7) Epoch 9, batch 3500, loss[loss=0.2152, simple_loss=0.2765, pruned_loss=0.07699, over 4821.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2634, pruned_loss=0.06819, over 952184.53 frames. ], batch size: 38, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:53:58,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.331e+01 1.641e+02 1.916e+02 2.289e+02 6.335e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-26 10:54:13,544 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 10:54:34,214 INFO [finetune.py:976] (0/7) Epoch 9, batch 3550, loss[loss=0.2324, simple_loss=0.2865, pruned_loss=0.08915, over 4939.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.261, pruned_loss=0.06787, over 953163.40 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:09,814 INFO [finetune.py:976] (0/7) Epoch 9, batch 3600, loss[loss=0.1781, simple_loss=0.2421, pruned_loss=0.05708, over 4762.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2584, pruned_loss=0.06632, over 955000.83 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:09,916 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:14,804 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1958, 2.0843, 1.7906, 2.1609, 2.0582, 2.0344, 1.9927, 2.9801], + device='cuda:0'), covar=tensor([0.4769, 0.6078, 0.4027, 0.5643, 0.5424, 0.2950, 0.5519, 0.1832], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0222, 0.0281, 0.0243, 0.0209, 0.0246, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:55:15,220 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.662e+02 2.002e+02 2.382e+02 4.044e+02, threshold=4.004e+02, percent-clipped=1.0 +2023-03-26 10:55:15,961 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49432.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:55:43,178 INFO [finetune.py:976] (0/7) Epoch 9, batch 3650, loss[loss=0.1776, simple_loss=0.2476, pruned_loss=0.05384, over 4725.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2598, pruned_loss=0.06691, over 955208.94 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:46,385 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:50,108 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:55,203 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 10:55:56,194 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49493.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 10:55:56,761 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49494.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:11,257 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-26 10:56:14,866 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-26 10:56:17,055 INFO [finetune.py:976] (0/7) Epoch 9, batch 3700, loss[loss=0.2292, simple_loss=0.2884, pruned_loss=0.08497, over 4153.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2646, pruned_loss=0.06856, over 951755.43 frames. ], batch size: 65, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:56:18,949 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:22,513 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.782e+02 2.076e+02 2.384e+02 4.659e+02, threshold=4.152e+02, percent-clipped=5.0 +2023-03-26 10:56:26,819 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8470, 1.2461, 0.8543, 1.6394, 2.1095, 1.3647, 1.5401, 1.7004], + device='cuda:0'), covar=tensor([0.1563, 0.2358, 0.2364, 0.1431, 0.2035, 0.2240, 0.1644, 0.2122], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0123, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 10:56:29,208 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49542.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:47,842 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 10:56:50,562 INFO [finetune.py:976] (0/7) Epoch 9, batch 3750, loss[loss=0.175, simple_loss=0.2144, pruned_loss=0.06775, over 4080.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2646, pruned_loss=0.068, over 951080.08 frames. ], batch size: 17, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:57:02,677 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:57:28,147 INFO [finetune.py:976] (0/7) Epoch 9, batch 3800, loss[loss=0.2017, simple_loss=0.2655, pruned_loss=0.069, over 4724.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2645, pruned_loss=0.06755, over 951451.69 frames. ], batch size: 23, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:57:39,046 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.635e+02 1.863e+02 2.259e+02 4.048e+02, threshold=3.725e+02, percent-clipped=0.0 +2023-03-26 10:58:06,700 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8063, 1.5588, 1.4258, 1.1403, 1.5822, 1.5213, 1.5308, 2.1740], + device='cuda:0'), covar=tensor([0.4669, 0.4420, 0.3705, 0.4378, 0.4178, 0.2619, 0.4059, 0.1909], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0221, 0.0281, 0.0243, 0.0209, 0.0246, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 10:58:10,830 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.95 vs. limit=5.0 +2023-03-26 10:58:12,406 INFO [finetune.py:976] (0/7) Epoch 9, batch 3850, loss[loss=0.2622, simple_loss=0.3057, pruned_loss=0.1094, over 4848.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.265, pruned_loss=0.06819, over 953519.85 frames. ], batch size: 47, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:58:48,063 INFO [finetune.py:976] (0/7) Epoch 9, batch 3900, loss[loss=0.1858, simple_loss=0.2462, pruned_loss=0.06276, over 4764.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2623, pruned_loss=0.06743, over 953912.73 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:58:58,258 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.638e+02 1.913e+02 2.415e+02 4.821e+02, threshold=3.825e+02, percent-clipped=2.0 +2023-03-26 10:59:10,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9934, 1.8298, 1.7604, 1.9752, 1.3579, 4.5329, 1.7146, 2.3401], + device='cuda:0'), covar=tensor([0.3313, 0.2366, 0.2034, 0.2259, 0.1754, 0.0117, 0.2438, 0.1220], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 10:59:31,481 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:59:36,464 INFO [finetune.py:976] (0/7) Epoch 9, batch 3950, loss[loss=0.1583, simple_loss=0.2286, pruned_loss=0.04402, over 4741.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2582, pruned_loss=0.06531, over 955144.83 frames. ], batch size: 27, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:59:40,657 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:59:47,184 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49788.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:00:09,537 INFO [finetune.py:976] (0/7) Epoch 9, batch 4000, loss[loss=0.2687, simple_loss=0.3305, pruned_loss=0.1034, over 4820.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2589, pruned_loss=0.06642, over 955066.18 frames. ], batch size: 40, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:00:12,479 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5242, 2.1903, 1.7599, 0.7535, 1.8798, 1.9973, 1.7383, 1.9173], + device='cuda:0'), covar=tensor([0.0660, 0.0793, 0.1391, 0.1933, 0.1253, 0.1842, 0.2033, 0.0827], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0198, 0.0199, 0.0186, 0.0214, 0.0205, 0.0221, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:00:12,539 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-26 11:00:13,714 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49827.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:00:16,513 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.645e+02 2.017e+02 2.376e+02 6.319e+02, threshold=4.034e+02, percent-clipped=3.0 +2023-03-26 11:00:42,824 INFO [finetune.py:976] (0/7) Epoch 9, batch 4050, loss[loss=0.2489, simple_loss=0.3023, pruned_loss=0.09773, over 4824.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2642, pruned_loss=0.06885, over 953854.25 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:00:56,953 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:00:57,003 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0029, 1.9704, 1.5602, 1.9937, 1.9496, 1.6672, 2.3085, 2.0314], + device='cuda:0'), covar=tensor([0.1308, 0.2205, 0.3066, 0.2595, 0.2626, 0.1686, 0.2918, 0.1728], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0188, 0.0232, 0.0253, 0.0237, 0.0195, 0.0211, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:00:58,210 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3893, 1.4877, 1.5419, 1.7374, 1.5373, 3.2027, 1.3108, 1.5534], + device='cuda:0'), covar=tensor([0.1001, 0.1758, 0.1231, 0.0965, 0.1585, 0.0242, 0.1573, 0.1777], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0075, 0.0078, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-26 11:01:05,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1577, 2.0939, 1.5395, 2.1897, 2.0192, 1.7766, 2.4866, 2.1493], + device='cuda:0'), covar=tensor([0.1317, 0.2423, 0.3397, 0.2765, 0.2703, 0.1754, 0.3409, 0.1878], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0188, 0.0233, 0.0253, 0.0238, 0.0195, 0.0212, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:01:15,999 INFO [finetune.py:976] (0/7) Epoch 9, batch 4100, loss[loss=0.2291, simple_loss=0.3017, pruned_loss=0.07831, over 4912.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2662, pruned_loss=0.06879, over 954824.61 frames. ], batch size: 36, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:01:22,949 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.254e+02 1.718e+02 2.083e+02 2.512e+02 3.689e+02, threshold=4.166e+02, percent-clipped=0.0 +2023-03-26 11:01:28,937 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:01:48,770 INFO [finetune.py:976] (0/7) Epoch 9, batch 4150, loss[loss=0.1528, simple_loss=0.2278, pruned_loss=0.03888, over 4778.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2667, pruned_loss=0.0684, over 955476.75 frames. ], batch size: 29, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:02:09,108 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-50000.pt +2023-03-26 11:02:23,457 INFO [finetune.py:976] (0/7) Epoch 9, batch 4200, loss[loss=0.1975, simple_loss=0.2556, pruned_loss=0.06972, over 4931.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2681, pruned_loss=0.0691, over 956597.58 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:02:31,376 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.706e+02 2.002e+02 2.506e+02 6.230e+02, threshold=4.003e+02, percent-clipped=2.0 +2023-03-26 11:02:35,547 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1125, 1.6710, 1.7909, 0.7733, 2.0208, 2.3156, 1.7686, 1.7262], + device='cuda:0'), covar=tensor([0.1240, 0.1211, 0.0685, 0.0916, 0.0699, 0.0671, 0.0898, 0.0936], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0157, 0.0122, 0.0136, 0.0133, 0.0127, 0.0148, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.6949e-05, 1.1560e-04, 8.8293e-05, 9.8841e-05, 9.5275e-05, 9.3194e-05, + 1.0840e-04, 1.0960e-04], device='cuda:0') +2023-03-26 11:02:57,222 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 11:03:01,000 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7848, 3.2929, 3.4669, 3.6569, 3.5757, 3.3591, 3.8509, 1.3001], + device='cuda:0'), covar=tensor([0.0791, 0.0824, 0.0787, 0.0856, 0.1178, 0.1340, 0.0755, 0.4941], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0276, 0.0292, 0.0329, 0.0281, 0.0300, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:03:12,378 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-03-26 11:03:15,797 INFO [finetune.py:976] (0/7) Epoch 9, batch 4250, loss[loss=0.2762, simple_loss=0.3155, pruned_loss=0.1185, over 4895.00 frames. ], tot_loss[loss=0.2026, simple_loss=0.2669, pruned_loss=0.0691, over 956804.83 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:03:24,149 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:32,329 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50088.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:03:45,045 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 11:03:53,928 INFO [finetune.py:976] (0/7) Epoch 9, batch 4300, loss[loss=0.1752, simple_loss=0.2414, pruned_loss=0.05455, over 4913.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2634, pruned_loss=0.06779, over 958389.18 frames. ], batch size: 43, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:03:53,994 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:56,365 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:57,626 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50128.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:04:00,430 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.608e+02 1.917e+02 2.228e+02 4.011e+02, threshold=3.835e+02, percent-clipped=1.0 +2023-03-26 11:04:03,873 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50136.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:04:49,372 INFO [finetune.py:976] (0/7) Epoch 9, batch 4350, loss[loss=0.2004, simple_loss=0.2691, pruned_loss=0.06579, over 4924.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2603, pruned_loss=0.06675, over 958277.77 frames. ], batch size: 37, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:05:17,875 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50189.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:05:30,439 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50197.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:05:41,553 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6416, 1.6849, 1.3421, 1.5413, 1.9545, 1.8688, 1.6085, 1.4264], + device='cuda:0'), covar=tensor([0.0261, 0.0279, 0.0593, 0.0333, 0.0210, 0.0421, 0.0325, 0.0414], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0108, 0.0138, 0.0114, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.9587e-05, 8.4884e-05, 1.1026e-04, 8.9278e-05, 7.9656e-05, 7.5223e-05, + 6.8646e-05, 8.2996e-05], device='cuda:0') +2023-03-26 11:06:02,455 INFO [finetune.py:976] (0/7) Epoch 9, batch 4400, loss[loss=0.2307, simple_loss=0.3007, pruned_loss=0.08034, over 4907.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2626, pruned_loss=0.06796, over 957008.68 frames. ], batch size: 37, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:06:02,581 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5664, 1.9069, 1.5258, 1.5360, 2.0262, 1.9936, 1.7615, 1.7555], + device='cuda:0'), covar=tensor([0.0407, 0.0298, 0.0533, 0.0340, 0.0298, 0.0563, 0.0317, 0.0378], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0109, 0.0139, 0.0114, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:0'), out_proj_covar=tensor([6.9767e-05, 8.5019e-05, 1.1048e-04, 8.9436e-05, 7.9808e-05, 7.5305e-05, + 6.8755e-05, 8.3104e-05], device='cuda:0') +2023-03-26 11:06:14,078 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.827e+01 1.714e+02 1.989e+02 2.480e+02 5.028e+02, threshold=3.977e+02, percent-clipped=2.0 +2023-03-26 11:06:48,127 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50258.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:06:58,869 INFO [finetune.py:976] (0/7) Epoch 9, batch 4450, loss[loss=0.1662, simple_loss=0.2493, pruned_loss=0.04151, over 4871.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.266, pruned_loss=0.0689, over 952933.38 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:07:17,307 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 11:07:32,435 INFO [finetune.py:976] (0/7) Epoch 9, batch 4500, loss[loss=0.1963, simple_loss=0.2545, pruned_loss=0.06902, over 4705.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2677, pruned_loss=0.06919, over 955110.45 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:07:38,451 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.523e+02 1.896e+02 2.480e+02 4.445e+02, threshold=3.793e+02, percent-clipped=2.0 +2023-03-26 11:07:53,278 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-26 11:08:05,986 INFO [finetune.py:976] (0/7) Epoch 9, batch 4550, loss[loss=0.2182, simple_loss=0.287, pruned_loss=0.07465, over 4858.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2689, pruned_loss=0.0698, over 953853.10 frames. ], batch size: 44, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:08:24,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9208, 1.9176, 1.9064, 1.3755, 2.0367, 2.0888, 1.9244, 1.5948], + device='cuda:0'), covar=tensor([0.0563, 0.0608, 0.0693, 0.0824, 0.0583, 0.0584, 0.0603, 0.1154], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0123, 0.0117, 0.0143, 0.0143, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:08:58,491 INFO [finetune.py:976] (0/7) Epoch 9, batch 4600, loss[loss=0.174, simple_loss=0.2421, pruned_loss=0.05289, over 4764.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2674, pruned_loss=0.06839, over 954703.23 frames. ], batch size: 28, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:08:58,582 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:09:06,260 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.671e+02 1.983e+02 2.416e+02 3.848e+02, threshold=3.965e+02, percent-clipped=1.0 +2023-03-26 11:09:14,551 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 11:09:39,359 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:09:40,499 INFO [finetune.py:976] (0/7) Epoch 9, batch 4650, loss[loss=0.1851, simple_loss=0.2538, pruned_loss=0.05818, over 4723.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2642, pruned_loss=0.0671, over 954020.41 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:09:50,495 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50484.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:10:10,609 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6728, 1.4267, 2.2303, 3.5308, 2.3259, 2.4434, 0.8151, 2.7088], + device='cuda:0'), covar=tensor([0.1782, 0.1612, 0.1337, 0.0586, 0.0811, 0.1444, 0.2065, 0.0567], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0134, 0.0165, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:10:22,782 INFO [finetune.py:976] (0/7) Epoch 9, batch 4700, loss[loss=0.177, simple_loss=0.2453, pruned_loss=0.05439, over 4797.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2616, pruned_loss=0.06631, over 954162.74 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:10:29,352 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.527e+02 1.850e+02 2.224e+02 3.838e+02, threshold=3.699e+02, percent-clipped=0.0 +2023-03-26 11:10:35,625 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8518, 1.7339, 1.6559, 1.9878, 2.1339, 2.0217, 1.3806, 1.5695], + device='cuda:0'), covar=tensor([0.2069, 0.2034, 0.1761, 0.1555, 0.1652, 0.1096, 0.2573, 0.1838], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0210, 0.0208, 0.0190, 0.0242, 0.0181, 0.0215, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:10:36,868 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8412, 1.1561, 1.8065, 1.7295, 1.5632, 1.5103, 1.6403, 1.6276], + device='cuda:0'), covar=tensor([0.4200, 0.4633, 0.3613, 0.4094, 0.4977, 0.4042, 0.4702, 0.3569], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0241, 0.0254, 0.0257, 0.0251, 0.0226, 0.0274, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:10:38,010 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8989, 3.6615, 3.5711, 1.9035, 3.7157, 2.9372, 1.3937, 2.6443], + device='cuda:0'), covar=tensor([0.2260, 0.1916, 0.1697, 0.3195, 0.1183, 0.0929, 0.4115, 0.1541], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0175, 0.0161, 0.0129, 0.0157, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:10:38,065 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50546.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:10:42,729 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50553.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:10:56,119 INFO [finetune.py:976] (0/7) Epoch 9, batch 4750, loss[loss=0.1941, simple_loss=0.2576, pruned_loss=0.06531, over 4911.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2586, pruned_loss=0.06509, over 953241.02 frames. ], batch size: 36, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:11:02,844 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1103, 1.8522, 1.7089, 1.8004, 1.8638, 1.8287, 1.8048, 2.5872], + device='cuda:0'), covar=tensor([0.5145, 0.5891, 0.4092, 0.5143, 0.4952, 0.3077, 0.5027, 0.2055], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0221, 0.0280, 0.0243, 0.0207, 0.0246, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:11:03,490 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 11:11:18,575 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:11:27,725 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6982, 1.5064, 2.2929, 3.4083, 2.2759, 2.3108, 0.8778, 2.6340], + device='cuda:0'), covar=tensor([0.1785, 0.1524, 0.1231, 0.0573, 0.0822, 0.1631, 0.2061, 0.0629], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0134, 0.0164, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:11:29,494 INFO [finetune.py:976] (0/7) Epoch 9, batch 4800, loss[loss=0.2622, simple_loss=0.3211, pruned_loss=0.1016, over 4832.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2625, pruned_loss=0.06666, over 954212.43 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:11:36,104 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.695e+01 1.587e+02 1.907e+02 2.189e+02 3.978e+02, threshold=3.813e+02, percent-clipped=2.0 +2023-03-26 11:11:44,283 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-26 11:12:03,073 INFO [finetune.py:976] (0/7) Epoch 9, batch 4850, loss[loss=0.2641, simple_loss=0.325, pruned_loss=0.1016, over 4817.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.266, pruned_loss=0.06792, over 954417.18 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:12:36,219 INFO [finetune.py:976] (0/7) Epoch 9, batch 4900, loss[loss=0.2306, simple_loss=0.289, pruned_loss=0.08607, over 4204.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.267, pruned_loss=0.06864, over 952666.28 frames. ], batch size: 65, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:12:42,295 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.610e+02 1.915e+02 2.289e+02 4.400e+02, threshold=3.830e+02, percent-clipped=2.0 +2023-03-26 11:12:50,062 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4216, 2.1273, 1.6027, 0.7914, 1.8510, 1.8361, 1.5716, 1.7819], + device='cuda:0'), covar=tensor([0.0749, 0.0888, 0.1558, 0.2194, 0.1441, 0.2336, 0.2568, 0.0970], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0200, 0.0200, 0.0188, 0.0215, 0.0207, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:12:59,141 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 11:13:08,690 INFO [finetune.py:976] (0/7) Epoch 9, batch 4950, loss[loss=0.2219, simple_loss=0.282, pruned_loss=0.08085, over 4760.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2691, pruned_loss=0.06933, over 954566.96 frames. ], batch size: 54, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:13:15,363 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50782.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:13:16,531 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:13:33,483 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50808.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:13:51,977 INFO [finetune.py:976] (0/7) Epoch 9, batch 5000, loss[loss=0.2475, simple_loss=0.2947, pruned_loss=0.1002, over 4841.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2677, pruned_loss=0.06896, over 954565.30 frames. ], batch size: 49, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:13:54,622 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 11:14:03,066 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.212e+02 1.752e+02 2.044e+02 2.444e+02 6.074e+02, threshold=4.089e+02, percent-clipped=4.0 +2023-03-26 11:14:03,136 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:14:03,216 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7838, 1.6073, 1.4743, 1.8411, 2.0390, 1.8670, 1.1288, 1.4639], + device='cuda:0'), covar=tensor([0.2241, 0.2168, 0.2020, 0.1793, 0.1754, 0.1154, 0.2915, 0.2064], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0209, 0.0206, 0.0188, 0.0240, 0.0180, 0.0214, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:14:13,393 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50843.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:14:24,237 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50853.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:14:36,968 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-03-26 11:14:37,384 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50869.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:14:40,057 INFO [finetune.py:976] (0/7) Epoch 9, batch 5050, loss[loss=0.1827, simple_loss=0.2531, pruned_loss=0.05614, over 4877.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2643, pruned_loss=0.06808, over 956336.93 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:14:52,015 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 11:14:53,256 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-26 11:15:00,237 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50901.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:15:00,840 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:15:21,255 INFO [finetune.py:976] (0/7) Epoch 9, batch 5100, loss[loss=0.1561, simple_loss=0.2181, pruned_loss=0.04704, over 4773.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2604, pruned_loss=0.06666, over 955105.56 frames. ], batch size: 26, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:15:26,170 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-26 11:15:29,782 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.571e+02 1.989e+02 2.366e+02 5.072e+02, threshold=3.977e+02, percent-clipped=2.0 +2023-03-26 11:15:38,839 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50946.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:15:52,162 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7402, 1.1155, 0.9114, 1.6695, 2.1287, 1.3716, 1.5145, 1.6197], + device='cuda:0'), covar=tensor([0.1559, 0.2266, 0.2142, 0.1288, 0.1928, 0.2120, 0.1480, 0.2098], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0121, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:15:55,121 INFO [finetune.py:976] (0/7) Epoch 9, batch 5150, loss[loss=0.1769, simple_loss=0.2306, pruned_loss=0.06155, over 4213.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2596, pruned_loss=0.06649, over 955086.43 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:15:59,394 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1587, 1.9744, 2.1230, 1.1002, 2.3041, 2.5237, 2.0666, 1.9086], + device='cuda:0'), covar=tensor([0.1006, 0.0756, 0.0457, 0.0704, 0.0564, 0.0555, 0.0703, 0.0715], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0156, 0.0122, 0.0135, 0.0132, 0.0127, 0.0147, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.7070e-05, 1.1503e-04, 8.8286e-05, 9.8046e-05, 9.4515e-05, 9.2549e-05, + 1.0759e-04, 1.0894e-04], device='cuda:0') +2023-03-26 11:16:16,063 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9263, 1.6445, 2.1764, 1.4716, 1.9283, 2.0671, 1.5896, 2.2462], + device='cuda:0'), covar=tensor([0.1243, 0.2277, 0.1344, 0.1964, 0.0991, 0.1424, 0.2841, 0.0875], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0204, 0.0194, 0.0190, 0.0178, 0.0217, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:16:19,653 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:16:29,111 INFO [finetune.py:976] (0/7) Epoch 9, batch 5200, loss[loss=0.2195, simple_loss=0.2953, pruned_loss=0.07181, over 4815.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2635, pruned_loss=0.06763, over 956274.08 frames. ], batch size: 51, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:16:37,434 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.691e+02 2.095e+02 2.506e+02 4.401e+02, threshold=4.191e+02, percent-clipped=2.0 +2023-03-26 11:16:38,739 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-26 11:17:12,595 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51062.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:17:18,569 INFO [finetune.py:976] (0/7) Epoch 9, batch 5250, loss[loss=0.1885, simple_loss=0.2654, pruned_loss=0.05579, over 4805.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2654, pruned_loss=0.06815, over 956373.28 frames. ], batch size: 45, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:17:50,172 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 11:17:51,199 INFO [finetune.py:976] (0/7) Epoch 9, batch 5300, loss[loss=0.1863, simple_loss=0.2551, pruned_loss=0.05874, over 4756.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2672, pruned_loss=0.06879, over 954804.22 frames. ], batch size: 28, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:17:51,951 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51123.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:17:57,266 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.689e+02 2.023e+02 2.414e+02 5.734e+02, threshold=4.045e+02, percent-clipped=1.0 +2023-03-26 11:18:01,385 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51138.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:18:17,926 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2347, 1.3657, 1.3634, 0.7054, 1.2437, 1.4973, 1.5662, 1.2325], + device='cuda:0'), covar=tensor([0.0853, 0.0480, 0.0433, 0.0520, 0.0432, 0.0508, 0.0263, 0.0671], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0156, 0.0122, 0.0135, 0.0132, 0.0127, 0.0147, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.6913e-05, 1.1463e-04, 8.8135e-05, 9.8181e-05, 9.4688e-05, 9.2556e-05, + 1.0753e-04, 1.0881e-04], device='cuda:0') +2023-03-26 11:18:19,583 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51164.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 11:18:24,357 INFO [finetune.py:976] (0/7) Epoch 9, batch 5350, loss[loss=0.2017, simple_loss=0.2665, pruned_loss=0.06846, over 4908.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2675, pruned_loss=0.06878, over 954128.14 frames. ], batch size: 37, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:18:24,467 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51172.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:18:25,058 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0115, 1.7506, 2.5687, 1.4990, 2.1521, 2.1609, 1.5987, 2.4541], + device='cuda:0'), covar=tensor([0.1563, 0.2071, 0.1206, 0.2082, 0.1199, 0.1745, 0.2892, 0.1140], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0203, 0.0194, 0.0191, 0.0179, 0.0217, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:18:56,198 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:19:18,525 INFO [finetune.py:976] (0/7) Epoch 9, batch 5400, loss[loss=0.1775, simple_loss=0.2363, pruned_loss=0.05932, over 4337.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2652, pruned_loss=0.06835, over 954769.01 frames. ], batch size: 65, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:19:26,734 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.913e+01 1.608e+02 1.826e+02 2.251e+02 3.272e+02, threshold=3.651e+02, percent-clipped=0.0 +2023-03-26 11:19:32,706 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51233.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:19:44,430 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8024, 1.2862, 0.8465, 1.6573, 2.0825, 1.0599, 1.5461, 1.6729], + device='cuda:0'), covar=tensor([0.1300, 0.2006, 0.1966, 0.1107, 0.1816, 0.1874, 0.1357, 0.1898], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0122, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:19:48,893 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:20:03,230 INFO [finetune.py:976] (0/7) Epoch 9, batch 5450, loss[loss=0.1555, simple_loss=0.2273, pruned_loss=0.04184, over 4909.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2608, pruned_loss=0.06602, over 955998.01 frames. ], batch size: 43, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:20:21,400 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6312, 2.4493, 2.0128, 2.6973, 2.5736, 2.1823, 3.0683, 2.5781], + device='cuda:0'), covar=tensor([0.1355, 0.2766, 0.3519, 0.3099, 0.2829, 0.1800, 0.3272, 0.2092], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0188, 0.0233, 0.0255, 0.0239, 0.0196, 0.0212, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:20:31,733 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:20:50,272 INFO [finetune.py:976] (0/7) Epoch 9, batch 5500, loss[loss=0.187, simple_loss=0.2595, pruned_loss=0.05723, over 4839.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2583, pruned_loss=0.06538, over 955866.62 frames. ], batch size: 47, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:20:56,825 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.581e+02 1.869e+02 2.249e+02 3.902e+02, threshold=3.738e+02, percent-clipped=2.0 +2023-03-26 11:21:48,954 INFO [finetune.py:976] (0/7) Epoch 9, batch 5550, loss[loss=0.2001, simple_loss=0.2764, pruned_loss=0.06187, over 4792.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2627, pruned_loss=0.06799, over 955502.12 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:21:59,522 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51382.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:07,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51392.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:24,399 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:27,580 INFO [finetune.py:976] (0/7) Epoch 9, batch 5600, loss[loss=0.1917, simple_loss=0.2613, pruned_loss=0.06107, over 4779.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2659, pruned_loss=0.06848, over 953615.02 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:22:33,289 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.626e+02 2.005e+02 2.362e+02 4.096e+02, threshold=4.011e+02, percent-clipped=2.0 +2023-03-26 11:22:36,851 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51438.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:39,806 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51443.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:46,019 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51453.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:52,442 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51464.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:22:57,071 INFO [finetune.py:976] (0/7) Epoch 9, batch 5650, loss[loss=0.223, simple_loss=0.3027, pruned_loss=0.07162, over 4811.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.27, pruned_loss=0.07013, over 952660.79 frames. ], batch size: 51, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:23:05,304 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:23:10,702 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:23:19,663 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2326, 1.3898, 1.4460, 0.6952, 1.3393, 1.5480, 1.5764, 1.2408], + device='cuda:0'), covar=tensor([0.0734, 0.0421, 0.0452, 0.0450, 0.0449, 0.0496, 0.0331, 0.0570], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0155, 0.0122, 0.0135, 0.0132, 0.0126, 0.0146, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.6140e-05, 1.1414e-04, 8.7826e-05, 9.7761e-05, 9.4654e-05, 9.2113e-05, + 1.0691e-04, 1.0846e-04], device='cuda:0') +2023-03-26 11:23:20,775 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51512.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:23:20,825 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3793, 1.3711, 1.4433, 0.8061, 1.4161, 1.4920, 1.3725, 1.2903], + device='cuda:0'), covar=tensor([0.0495, 0.0581, 0.0559, 0.0783, 0.0704, 0.0539, 0.0548, 0.1008], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0133, 0.0144, 0.0124, 0.0117, 0.0144, 0.0144, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:23:26,711 INFO [finetune.py:976] (0/7) Epoch 9, batch 5700, loss[loss=0.1936, simple_loss=0.2407, pruned_loss=0.07318, over 4286.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.265, pruned_loss=0.06957, over 933980.37 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:23:30,373 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51528.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:23:32,893 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.987e+01 1.624e+02 1.963e+02 2.341e+02 6.572e+02, threshold=3.927e+02, percent-clipped=1.0 +2023-03-26 11:23:33,599 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4612, 1.3497, 1.4122, 1.3365, 0.7774, 2.1570, 0.7831, 1.2809], + device='cuda:0'), covar=tensor([0.2969, 0.2127, 0.1927, 0.2193, 0.1840, 0.0405, 0.2432, 0.1157], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0116, 0.0098, 0.0099, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 11:23:43,281 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-9.pt +2023-03-26 11:23:57,292 INFO [finetune.py:976] (0/7) Epoch 10, batch 0, loss[loss=0.2215, simple_loss=0.2836, pruned_loss=0.07968, over 4894.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2836, pruned_loss=0.07968, over 4894.00 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:23:57,293 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 11:24:16,170 INFO [finetune.py:1010] (0/7) Epoch 10, validation: loss=0.1604, simple_loss=0.2317, pruned_loss=0.04451, over 2265189.00 frames. +2023-03-26 11:24:16,171 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 11:24:22,514 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:24:58,294 INFO [finetune.py:976] (0/7) Epoch 10, batch 50, loss[loss=0.1861, simple_loss=0.2482, pruned_loss=0.06198, over 4852.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2686, pruned_loss=0.0708, over 216255.22 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:24:58,424 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3638, 1.2139, 1.2296, 1.3399, 1.6333, 1.4684, 1.3655, 1.1772], + device='cuda:0'), covar=tensor([0.0325, 0.0294, 0.0584, 0.0288, 0.0208, 0.0438, 0.0251, 0.0333], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0108, 0.0139, 0.0114, 0.0102, 0.0101, 0.0091, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.0019e-05, 8.4819e-05, 1.1058e-04, 8.9774e-05, 7.9891e-05, 7.4985e-05, + 6.8657e-05, 8.2436e-05], device='cuda:0') +2023-03-26 11:25:01,128 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:25:06,927 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3611, 2.6477, 2.2808, 1.8451, 2.3199, 2.7897, 2.5290, 2.3560], + device='cuda:0'), covar=tensor([0.0609, 0.0558, 0.0859, 0.0923, 0.0770, 0.0703, 0.0677, 0.0905], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0134, 0.0145, 0.0124, 0.0118, 0.0144, 0.0144, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:25:20,208 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.735e+02 2.131e+02 2.642e+02 7.480e+02, threshold=4.262e+02, percent-clipped=4.0 +2023-03-26 11:25:31,993 INFO [finetune.py:976] (0/7) Epoch 10, batch 100, loss[loss=0.1633, simple_loss=0.2352, pruned_loss=0.04568, over 4890.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2617, pruned_loss=0.06787, over 378521.74 frames. ], batch size: 32, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:25:32,636 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:04,771 INFO [finetune.py:976] (0/7) Epoch 10, batch 150, loss[loss=0.2013, simple_loss=0.2598, pruned_loss=0.0714, over 4895.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2563, pruned_loss=0.06549, over 506763.44 frames. ], batch size: 43, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:26:18,718 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:33,393 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.594e+02 1.858e+02 2.240e+02 3.308e+02, threshold=3.716e+02, percent-clipped=0.0 +2023-03-26 11:26:37,114 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:47,600 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:48,176 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:52,002 INFO [finetune.py:976] (0/7) Epoch 10, batch 200, loss[loss=0.1872, simple_loss=0.2536, pruned_loss=0.06034, over 4921.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.257, pruned_loss=0.06659, over 607462.04 frames. ], batch size: 38, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:27:04,853 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:27:07,375 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3572, 2.3338, 1.9690, 1.1157, 2.0304, 1.8415, 1.6389, 2.0212], + device='cuda:0'), covar=tensor([0.0976, 0.0752, 0.1504, 0.1972, 0.1705, 0.2008, 0.2196, 0.1114], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0200, 0.0202, 0.0189, 0.0217, 0.0208, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:27:25,424 INFO [finetune.py:976] (0/7) Epoch 10, batch 250, loss[loss=0.2506, simple_loss=0.3097, pruned_loss=0.09571, over 4834.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2593, pruned_loss=0.06607, over 683735.62 frames. ], batch size: 45, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:27:33,013 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51808.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:27:45,689 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51828.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:27:48,003 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.605e+02 1.930e+02 2.331e+02 5.576e+02, threshold=3.861e+02, percent-clipped=5.0 +2023-03-26 11:27:51,184 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8326, 2.5600, 2.3635, 1.4639, 2.4873, 2.0384, 1.7979, 2.2943], + device='cuda:0'), covar=tensor([0.1052, 0.0750, 0.1566, 0.1975, 0.1716, 0.2200, 0.2195, 0.1150], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0202, 0.0203, 0.0190, 0.0218, 0.0209, 0.0226, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:27:58,881 INFO [finetune.py:976] (0/7) Epoch 10, batch 300, loss[loss=0.2276, simple_loss=0.2937, pruned_loss=0.08074, over 4840.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2624, pruned_loss=0.06659, over 745098.51 frames. ], batch size: 47, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:28:00,156 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:28:17,687 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51876.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:28:18,274 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3596, 3.7682, 3.9307, 4.1228, 4.1048, 3.8846, 4.4644, 1.4491], + device='cuda:0'), covar=tensor([0.0827, 0.0873, 0.0823, 0.1113, 0.1313, 0.1525, 0.0656, 0.5314], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0277, 0.0293, 0.0330, 0.0280, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:28:31,953 INFO [finetune.py:976] (0/7) Epoch 10, batch 350, loss[loss=0.2403, simple_loss=0.2834, pruned_loss=0.09861, over 4768.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.264, pruned_loss=0.06711, over 790922.42 frames. ], batch size: 26, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:28:48,506 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5929, 1.3829, 2.0154, 3.0466, 2.1131, 2.3326, 0.8191, 2.3863], + device='cuda:0'), covar=tensor([0.1710, 0.1451, 0.1220, 0.0619, 0.0837, 0.1467, 0.1862, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0118, 0.0135, 0.0165, 0.0102, 0.0140, 0.0127, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:28:54,278 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.678e+02 2.044e+02 2.443e+02 3.814e+02, threshold=4.089e+02, percent-clipped=0.0 +2023-03-26 11:29:04,642 INFO [finetune.py:976] (0/7) Epoch 10, batch 400, loss[loss=0.2664, simple_loss=0.3164, pruned_loss=0.1082, over 4822.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2654, pruned_loss=0.06765, over 826918.05 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:29:05,360 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4165, 2.1214, 2.8773, 1.8215, 2.5498, 2.4272, 2.1922, 2.7727], + device='cuda:0'), covar=tensor([0.1396, 0.2021, 0.1493, 0.2470, 0.0966, 0.1951, 0.2601, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0205, 0.0195, 0.0192, 0.0180, 0.0216, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:29:56,981 INFO [finetune.py:976] (0/7) Epoch 10, batch 450, loss[loss=0.2026, simple_loss=0.2754, pruned_loss=0.06492, over 4895.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2648, pruned_loss=0.0672, over 856520.23 frames. ], batch size: 35, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:29:57,730 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-52000.pt +2023-03-26 11:30:21,153 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.724e+02 2.025e+02 2.574e+02 4.346e+02, threshold=4.050e+02, percent-clipped=1.0 +2023-03-26 11:30:24,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52038.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:26,192 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:30,939 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:31,463 INFO [finetune.py:976] (0/7) Epoch 10, batch 500, loss[loss=0.1724, simple_loss=0.2377, pruned_loss=0.05356, over 4856.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.2635, pruned_loss=0.06711, over 879519.15 frames. ], batch size: 31, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:30:56,804 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52086.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:02,788 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:04,604 INFO [finetune.py:976] (0/7) Epoch 10, batch 550, loss[loss=0.1655, simple_loss=0.2334, pruned_loss=0.04878, over 4715.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2605, pruned_loss=0.06637, over 896836.46 frames. ], batch size: 23, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:31:05,931 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1799, 2.8505, 2.6513, 1.4262, 2.6166, 2.3280, 2.2036, 2.4719], + device='cuda:0'), covar=tensor([0.0981, 0.0902, 0.1670, 0.2189, 0.2046, 0.2087, 0.1993, 0.1140], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0203, 0.0204, 0.0191, 0.0218, 0.0210, 0.0226, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:31:05,943 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:07,054 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:17,450 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 11:31:27,062 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.591e+02 1.822e+02 2.163e+02 6.487e+02, threshold=3.643e+02, percent-clipped=1.0 +2023-03-26 11:31:32,618 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1100, 1.0699, 1.0409, 0.4762, 0.9749, 1.2140, 1.2785, 0.9880], + device='cuda:0'), covar=tensor([0.0779, 0.0484, 0.0463, 0.0504, 0.0517, 0.0504, 0.0391, 0.0587], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0154, 0.0121, 0.0134, 0.0131, 0.0124, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5626e-05, 1.1353e-04, 8.7010e-05, 9.7241e-05, 9.3711e-05, 9.0884e-05, + 1.0594e-04, 1.0754e-04], device='cuda:0') +2023-03-26 11:31:37,961 INFO [finetune.py:976] (0/7) Epoch 10, batch 600, loss[loss=0.1924, simple_loss=0.2565, pruned_loss=0.06421, over 4913.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2616, pruned_loss=0.06716, over 909380.95 frames. ], batch size: 36, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:31:39,240 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52151.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:15,960 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52193.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:19,560 INFO [finetune.py:976] (0/7) Epoch 10, batch 650, loss[loss=0.2434, simple_loss=0.2908, pruned_loss=0.098, over 4837.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2655, pruned_loss=0.06838, over 921105.49 frames. ], batch size: 30, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:32:19,619 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:39,850 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8581, 4.8200, 4.5632, 2.7354, 4.8165, 3.7884, 0.7447, 3.2307], + device='cuda:0'), covar=tensor([0.2414, 0.1496, 0.1215, 0.2757, 0.0772, 0.0808, 0.4642, 0.1388], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0175, 0.0160, 0.0128, 0.0157, 0.0123, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:32:42,612 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.681e+02 1.969e+02 2.336e+02 3.855e+02, threshold=3.938e+02, percent-clipped=2.0 +2023-03-26 11:32:53,490 INFO [finetune.py:976] (0/7) Epoch 10, batch 700, loss[loss=0.2364, simple_loss=0.2952, pruned_loss=0.08878, over 4805.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.2666, pruned_loss=0.06817, over 931228.04 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:32:56,655 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52254.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:33:26,709 INFO [finetune.py:976] (0/7) Epoch 10, batch 750, loss[loss=0.2183, simple_loss=0.2781, pruned_loss=0.07924, over 4880.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2671, pruned_loss=0.06825, over 937475.16 frames. ], batch size: 32, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:33:45,068 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:02,791 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.612e+02 1.864e+02 2.364e+02 4.342e+02, threshold=3.728e+02, percent-clipped=1.0 +2023-03-26 11:34:15,210 INFO [finetune.py:976] (0/7) Epoch 10, batch 800, loss[loss=0.1757, simple_loss=0.2432, pruned_loss=0.05413, over 4851.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2666, pruned_loss=0.06766, over 942183.10 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:34:30,566 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52373.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:49,105 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:50,910 INFO [finetune.py:976] (0/7) Epoch 10, batch 850, loss[loss=0.1972, simple_loss=0.2526, pruned_loss=0.07093, over 4851.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2637, pruned_loss=0.06675, over 947310.96 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:34:54,117 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52403.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:35:06,045 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1385, 1.7985, 2.0931, 1.9767, 1.7954, 1.8140, 1.9718, 1.9370], + device='cuda:0'), covar=tensor([0.4075, 0.4830, 0.3561, 0.4497, 0.5358, 0.4363, 0.5926, 0.3602], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0240, 0.0254, 0.0258, 0.0251, 0.0226, 0.0274, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:35:09,090 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5273, 1.7409, 1.8072, 0.9969, 1.8884, 2.0486, 1.8740, 1.5953], + device='cuda:0'), covar=tensor([0.1143, 0.0679, 0.0501, 0.0582, 0.0411, 0.0533, 0.0457, 0.0761], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0154, 0.0121, 0.0134, 0.0131, 0.0125, 0.0145, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5771e-05, 1.1345e-04, 8.7415e-05, 9.6927e-05, 9.3792e-05, 9.1096e-05, + 1.0609e-04, 1.0758e-04], device='cuda:0') +2023-03-26 11:35:14,904 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.556e+02 1.848e+02 2.239e+02 3.627e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 11:35:36,901 INFO [finetune.py:976] (0/7) Epoch 10, batch 900, loss[loss=0.1449, simple_loss=0.2224, pruned_loss=0.03369, over 4855.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2609, pruned_loss=0.06614, over 949195.40 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:35:38,209 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52451.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:36:25,701 INFO [finetune.py:976] (0/7) Epoch 10, batch 950, loss[loss=0.2021, simple_loss=0.265, pruned_loss=0.06962, over 4927.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2603, pruned_loss=0.06613, over 951496.91 frames. ], batch size: 38, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:36:45,734 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 11:36:46,733 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.551e+02 1.918e+02 2.238e+02 5.409e+02, threshold=3.837e+02, percent-clipped=4.0 +2023-03-26 11:36:47,006 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.93 vs. limit=5.0 +2023-03-26 11:37:01,195 INFO [finetune.py:976] (0/7) Epoch 10, batch 1000, loss[loss=0.1869, simple_loss=0.2683, pruned_loss=0.05273, over 4914.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2632, pruned_loss=0.06703, over 953729.75 frames. ], batch size: 36, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:37:01,272 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52549.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:37:03,111 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5979, 3.3594, 3.1865, 1.5295, 3.4808, 2.5556, 0.7659, 2.2318], + device='cuda:0'), covar=tensor([0.2504, 0.1805, 0.1739, 0.3435, 0.1191, 0.1122, 0.4678, 0.1595], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0173, 0.0160, 0.0128, 0.0156, 0.0123, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:37:04,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52555.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:38:00,467 INFO [finetune.py:976] (0/7) Epoch 10, batch 1050, loss[loss=0.2141, simple_loss=0.2817, pruned_loss=0.07324, over 4912.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2653, pruned_loss=0.0676, over 953678.57 frames. ], batch size: 37, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:38:12,706 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8866, 1.6296, 2.1817, 1.5054, 2.0944, 2.1163, 1.5231, 2.3701], + device='cuda:0'), covar=tensor([0.1473, 0.2079, 0.1498, 0.2169, 0.0981, 0.1631, 0.2828, 0.0879], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0204, 0.0194, 0.0191, 0.0178, 0.0215, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:38:21,494 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52616.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:38:31,486 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.591e+02 1.928e+02 2.293e+02 3.930e+02, threshold=3.855e+02, percent-clipped=1.0 +2023-03-26 11:38:44,944 INFO [finetune.py:976] (0/7) Epoch 10, batch 1100, loss[loss=0.2343, simple_loss=0.3003, pruned_loss=0.08419, over 4891.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2658, pruned_loss=0.06766, over 955401.59 frames. ], batch size: 43, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:38:59,650 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:17,267 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:19,452 INFO [finetune.py:976] (0/7) Epoch 10, batch 1150, loss[loss=0.1804, simple_loss=0.2496, pruned_loss=0.05561, over 4881.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2674, pruned_loss=0.06859, over 954673.53 frames. ], batch size: 35, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:39:40,840 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.654e+02 1.930e+02 2.314e+02 4.484e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 11:39:48,707 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:52,581 INFO [finetune.py:976] (0/7) Epoch 10, batch 1200, loss[loss=0.1888, simple_loss=0.2504, pruned_loss=0.06363, over 4744.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2659, pruned_loss=0.06812, over 956110.35 frames. ], batch size: 54, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:39:56,824 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52752.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:40:20,820 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3055, 2.1389, 1.9145, 2.2398, 2.2275, 1.9560, 2.5943, 2.2485], + device='cuda:0'), covar=tensor([0.1362, 0.2540, 0.3086, 0.2770, 0.2682, 0.1747, 0.3515, 0.2002], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0187, 0.0231, 0.0251, 0.0237, 0.0195, 0.0211, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:40:35,768 INFO [finetune.py:976] (0/7) Epoch 10, batch 1250, loss[loss=0.1787, simple_loss=0.2446, pruned_loss=0.05633, over 4786.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2634, pruned_loss=0.0674, over 956348.98 frames. ], batch size: 29, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:40:46,542 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 11:40:47,083 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52813.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:41:05,426 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.514e+02 1.794e+02 2.223e+02 4.744e+02, threshold=3.588e+02, percent-clipped=2.0 +2023-03-26 11:41:14,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3592, 1.6066, 1.0006, 2.3115, 2.6799, 1.8128, 2.0259, 2.1457], + device='cuda:0'), covar=tensor([0.1282, 0.1965, 0.2049, 0.0995, 0.1660, 0.1843, 0.1339, 0.1924], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0113, 0.0092, 0.0121, 0.0095, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:41:18,384 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8359, 1.8162, 1.5862, 2.0293, 2.3678, 2.0273, 1.6753, 1.4733], + device='cuda:0'), covar=tensor([0.2233, 0.2113, 0.1938, 0.1622, 0.1894, 0.1202, 0.2483, 0.1975], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0208, 0.0207, 0.0187, 0.0240, 0.0179, 0.0213, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:41:19,446 INFO [finetune.py:976] (0/7) Epoch 10, batch 1300, loss[loss=0.1987, simple_loss=0.2562, pruned_loss=0.07061, over 4801.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2586, pruned_loss=0.06486, over 957174.80 frames. ], batch size: 45, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:41:19,532 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:41:42,069 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-26 11:41:51,924 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:41:53,102 INFO [finetune.py:976] (0/7) Epoch 10, batch 1350, loss[loss=0.2169, simple_loss=0.2901, pruned_loss=0.0718, over 4825.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2594, pruned_loss=0.06515, over 957433.32 frames. ], batch size: 40, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:42:01,946 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52911.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:42:12,036 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 11:42:15,925 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.950e+01 1.660e+02 2.003e+02 2.564e+02 3.985e+02, threshold=4.006e+02, percent-clipped=2.0 +2023-03-26 11:42:20,749 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6508, 3.7862, 3.5882, 1.9053, 3.8721, 2.9132, 0.7515, 2.6227], + device='cuda:0'), covar=tensor([0.2502, 0.2212, 0.1464, 0.3160, 0.0994, 0.0997, 0.4471, 0.1546], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0175, 0.0161, 0.0129, 0.0157, 0.0123, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:42:30,883 INFO [finetune.py:976] (0/7) Epoch 10, batch 1400, loss[loss=0.1379, simple_loss=0.2094, pruned_loss=0.03322, over 4747.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2623, pruned_loss=0.06644, over 956821.91 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:42:43,960 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6778, 1.6238, 1.6286, 1.6366, 1.5151, 3.2164, 1.6257, 2.1274], + device='cuda:0'), covar=tensor([0.3061, 0.2156, 0.1913, 0.2045, 0.1496, 0.0231, 0.2771, 0.1136], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0099, 0.0099, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 11:42:48,557 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:42:53,596 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-03-26 11:42:54,092 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7369, 1.7988, 1.8742, 1.2312, 1.9797, 2.1594, 2.0427, 1.5486], + device='cuda:0'), covar=tensor([0.0962, 0.0621, 0.0456, 0.0562, 0.0414, 0.0579, 0.0398, 0.0859], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0155, 0.0121, 0.0135, 0.0131, 0.0125, 0.0145, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5760e-05, 1.1388e-04, 8.7329e-05, 9.7603e-05, 9.3766e-05, 9.1395e-05, + 1.0631e-04, 1.0755e-04], device='cuda:0') +2023-03-26 11:43:14,420 INFO [finetune.py:976] (0/7) Epoch 10, batch 1450, loss[loss=0.2257, simple_loss=0.2896, pruned_loss=0.08089, over 4818.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2659, pruned_loss=0.06764, over 955905.57 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:43:35,021 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:43:45,115 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.605e+02 1.913e+02 2.318e+02 4.347e+02, threshold=3.826e+02, percent-clipped=3.0 +2023-03-26 11:43:50,005 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8395, 1.6364, 1.4681, 1.3762, 1.8134, 1.5130, 1.8076, 1.7664], + device='cuda:0'), covar=tensor([0.1455, 0.2466, 0.3484, 0.2927, 0.2897, 0.1922, 0.3201, 0.2094], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0188, 0.0233, 0.0253, 0.0239, 0.0196, 0.0211, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:43:55,916 INFO [finetune.py:976] (0/7) Epoch 10, batch 1500, loss[loss=0.175, simple_loss=0.2461, pruned_loss=0.05194, over 4754.00 frames. ], tot_loss[loss=0.2021, simple_loss=0.2675, pruned_loss=0.06832, over 955743.89 frames. ], batch size: 26, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:44:00,674 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:29,472 INFO [finetune.py:976] (0/7) Epoch 10, batch 1550, loss[loss=0.22, simple_loss=0.2919, pruned_loss=0.07403, over 4891.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2668, pruned_loss=0.06783, over 956118.22 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:44:35,496 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:41,524 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:52,482 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.652e+02 2.005e+02 2.543e+02 4.651e+02, threshold=4.009e+02, percent-clipped=4.0 +2023-03-26 11:45:03,285 INFO [finetune.py:976] (0/7) Epoch 10, batch 1600, loss[loss=0.1775, simple_loss=0.2414, pruned_loss=0.05681, over 4849.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2636, pruned_loss=0.06648, over 957570.98 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:45:48,099 INFO [finetune.py:976] (0/7) Epoch 10, batch 1650, loss[loss=0.2007, simple_loss=0.2674, pruned_loss=0.06698, over 4755.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2605, pruned_loss=0.0655, over 955766.11 frames. ], batch size: 27, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:45:56,054 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53211.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:46:10,716 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.364e+01 1.592e+02 1.774e+02 2.189e+02 3.836e+02, threshold=3.549e+02, percent-clipped=0.0 +2023-03-26 11:46:16,358 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:46:23,558 INFO [finetune.py:976] (0/7) Epoch 10, batch 1700, loss[loss=0.1483, simple_loss=0.217, pruned_loss=0.03981, over 4834.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2578, pruned_loss=0.06434, over 955242.27 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:46:29,717 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53259.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:46:32,762 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7145, 1.2786, 0.8525, 1.6370, 2.1146, 1.4130, 1.5929, 1.6989], + device='cuda:0'), covar=tensor([0.1535, 0.2187, 0.2192, 0.1217, 0.2057, 0.2109, 0.1402, 0.1956], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0112, 0.0093, 0.0121, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:46:42,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3785, 2.8343, 2.4271, 1.8935, 2.6376, 2.9401, 2.8537, 2.4360], + device='cuda:0'), covar=tensor([0.0665, 0.0468, 0.0735, 0.0908, 0.0529, 0.0670, 0.0584, 0.0892], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0143, 0.0123, 0.0118, 0.0141, 0.0142, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:46:49,384 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 11:46:56,425 INFO [finetune.py:976] (0/7) Epoch 10, batch 1750, loss[loss=0.2171, simple_loss=0.2944, pruned_loss=0.06997, over 4841.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2607, pruned_loss=0.06534, over 955889.00 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:46:58,838 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:00,081 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:07,265 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:18,948 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.605e+02 1.832e+02 2.176e+02 4.638e+02, threshold=3.664e+02, percent-clipped=2.0 +2023-03-26 11:47:28,908 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 11:47:29,900 INFO [finetune.py:976] (0/7) Epoch 10, batch 1800, loss[loss=0.168, simple_loss=0.2429, pruned_loss=0.0465, over 4926.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.2645, pruned_loss=0.06625, over 957432.24 frames. ], batch size: 42, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:47:45,416 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53365.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:46,668 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53367.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:55,971 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:05,448 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 11:48:18,378 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-03-26 11:48:26,360 INFO [finetune.py:976] (0/7) Epoch 10, batch 1850, loss[loss=0.2043, simple_loss=0.2577, pruned_loss=0.07548, over 4885.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2666, pruned_loss=0.06771, over 955552.56 frames. ], batch size: 32, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:48:32,690 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:35,084 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53412.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:51,321 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53428.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:48:58,642 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.752e+02 2.111e+02 2.637e+02 7.323e+02, threshold=4.222e+02, percent-clipped=6.0 +2023-03-26 11:49:10,462 INFO [finetune.py:976] (0/7) Epoch 10, batch 1900, loss[loss=0.2135, simple_loss=0.2816, pruned_loss=0.07269, over 4825.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2666, pruned_loss=0.0671, over 955932.75 frames. ], batch size: 39, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:49:11,362 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 11:49:14,806 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:49:43,866 INFO [finetune.py:976] (0/7) Epoch 10, batch 1950, loss[loss=0.163, simple_loss=0.2373, pruned_loss=0.04434, over 4806.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2646, pruned_loss=0.06561, over 956586.61 frames. ], batch size: 40, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:49:48,247 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 11:50:09,883 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.535e+02 1.778e+02 2.101e+02 3.650e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-03-26 11:50:21,546 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2995, 2.1558, 1.8640, 1.1076, 2.1043, 1.9091, 1.6517, 2.1128], + device='cuda:0'), covar=tensor([0.0741, 0.0660, 0.1161, 0.1670, 0.0960, 0.1617, 0.1767, 0.0663], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0201, 0.0202, 0.0188, 0.0216, 0.0208, 0.0223, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:50:29,329 INFO [finetune.py:976] (0/7) Epoch 10, batch 2000, loss[loss=0.1788, simple_loss=0.2508, pruned_loss=0.05343, over 4825.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2618, pruned_loss=0.06518, over 955098.68 frames. ], batch size: 25, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:51:10,600 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2708, 2.9672, 2.8616, 1.1615, 3.0862, 2.2030, 0.9068, 1.9193], + device='cuda:0'), covar=tensor([0.2319, 0.2219, 0.1729, 0.3609, 0.1227, 0.1151, 0.3882, 0.1629], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0174, 0.0161, 0.0128, 0.0156, 0.0123, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:51:12,698 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 11:51:21,830 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53597.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:51:23,506 INFO [finetune.py:976] (0/7) Epoch 10, batch 2050, loss[loss=0.2122, simple_loss=0.2701, pruned_loss=0.07713, over 4829.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2589, pruned_loss=0.0643, over 955132.81 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:51:35,867 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9291, 2.1638, 1.1901, 2.8225, 3.1365, 2.3935, 2.6006, 2.5559], + device='cuda:0'), covar=tensor([0.1144, 0.1811, 0.2162, 0.0922, 0.1456, 0.1617, 0.1192, 0.1852], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:51:44,832 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.628e+02 1.908e+02 2.249e+02 5.707e+02, threshold=3.816e+02, percent-clipped=1.0 +2023-03-26 11:51:56,177 INFO [finetune.py:976] (0/7) Epoch 10, batch 2100, loss[loss=0.2323, simple_loss=0.2854, pruned_loss=0.0896, over 4914.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2582, pruned_loss=0.06458, over 956502.94 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:52:03,967 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53660.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:52:10,634 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53671.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:52:25,028 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0837, 4.8810, 4.6754, 2.7345, 5.0236, 3.6248, 0.8017, 3.3514], + device='cuda:0'), covar=tensor([0.2170, 0.1811, 0.1117, 0.2874, 0.0683, 0.0891, 0.4767, 0.1353], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0172, 0.0158, 0.0127, 0.0154, 0.0121, 0.0144, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 11:52:37,545 INFO [finetune.py:976] (0/7) Epoch 10, batch 2150, loss[loss=0.2517, simple_loss=0.3101, pruned_loss=0.0966, over 4876.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2624, pruned_loss=0.06631, over 958083.60 frames. ], batch size: 31, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:52:52,709 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53712.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:53:03,959 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53723.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:53:19,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.764e+02 2.057e+02 2.459e+02 5.535e+02, threshold=4.114e+02, percent-clipped=2.0 +2023-03-26 11:53:28,532 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2309, 1.5428, 0.7785, 2.1748, 2.6084, 1.8990, 1.9747, 2.1117], + device='cuda:0'), covar=tensor([0.1362, 0.1924, 0.2195, 0.1067, 0.1774, 0.1988, 0.1322, 0.1937], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0094, 0.0112, 0.0092, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:53:34,458 INFO [finetune.py:976] (0/7) Epoch 10, batch 2200, loss[loss=0.1959, simple_loss=0.2654, pruned_loss=0.06319, over 4748.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2652, pruned_loss=0.06732, over 958258.54 frames. ], batch size: 54, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:53:43,148 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:54:07,986 INFO [finetune.py:976] (0/7) Epoch 10, batch 2250, loss[loss=0.1708, simple_loss=0.2507, pruned_loss=0.04545, over 4861.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2657, pruned_loss=0.0672, over 958053.22 frames. ], batch size: 31, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:54:16,446 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 11:54:21,756 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1018, 1.8808, 1.6734, 1.7962, 1.7956, 1.7671, 1.7745, 2.5582], + device='cuda:0'), covar=tensor([0.4683, 0.5539, 0.3898, 0.5147, 0.5004, 0.2810, 0.5179, 0.1968], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0222, 0.0279, 0.0243, 0.0209, 0.0244, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:54:30,209 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.658e+02 1.958e+02 2.430e+02 3.560e+02, threshold=3.915e+02, percent-clipped=0.0 +2023-03-26 11:54:41,561 INFO [finetune.py:976] (0/7) Epoch 10, batch 2300, loss[loss=0.1906, simple_loss=0.2596, pruned_loss=0.06082, over 4910.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2655, pruned_loss=0.06626, over 958881.18 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:55:15,979 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:55:17,109 INFO [finetune.py:976] (0/7) Epoch 10, batch 2350, loss[loss=0.1714, simple_loss=0.2365, pruned_loss=0.05315, over 4932.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2637, pruned_loss=0.06598, over 957096.86 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:55:23,475 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6942, 1.4942, 0.9530, 0.2978, 1.2189, 1.4572, 1.4216, 1.4193], + device='cuda:0'), covar=tensor([0.0828, 0.0774, 0.1413, 0.1898, 0.1480, 0.2121, 0.2188, 0.0844], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0201, 0.0202, 0.0188, 0.0217, 0.0208, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:55:47,268 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.627e+02 1.969e+02 2.442e+02 4.599e+02, threshold=3.938e+02, percent-clipped=2.0 +2023-03-26 11:55:58,279 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:05,534 INFO [finetune.py:976] (0/7) Epoch 10, batch 2400, loss[loss=0.1793, simple_loss=0.2457, pruned_loss=0.05648, over 4901.00 frames. ], tot_loss[loss=0.196, simple_loss=0.261, pruned_loss=0.06545, over 955560.81 frames. ], batch size: 32, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 11:56:15,980 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:24,608 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53971.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:41,931 INFO [finetune.py:976] (0/7) Epoch 10, batch 2450, loss[loss=0.2035, simple_loss=0.2669, pruned_loss=0.0701, over 4826.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2578, pruned_loss=0.06464, over 956074.50 frames. ], batch size: 39, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:56:42,663 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-54000.pt +2023-03-26 11:56:49,145 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54008.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:56,916 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54019.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:59,865 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=54023.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:57:00,493 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5547, 1.3729, 1.2834, 1.5359, 1.5650, 1.5212, 1.0210, 1.3136], + device='cuda:0'), covar=tensor([0.1937, 0.1923, 0.1771, 0.1498, 0.1672, 0.1171, 0.2461, 0.1796], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0207, 0.0207, 0.0188, 0.0240, 0.0180, 0.0213, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:57:05,184 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.615e+02 1.942e+02 2.268e+02 4.833e+02, threshold=3.884e+02, percent-clipped=2.0 +2023-03-26 11:57:16,019 INFO [finetune.py:976] (0/7) Epoch 10, batch 2500, loss[loss=0.2392, simple_loss=0.2972, pruned_loss=0.09063, over 4871.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2605, pruned_loss=0.06643, over 954610.27 frames. ], batch size: 34, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:57:25,777 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-26 11:57:42,136 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54071.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:57:45,729 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7113, 1.5218, 2.1761, 3.5294, 2.3440, 2.3561, 1.1120, 2.7258], + device='cuda:0'), covar=tensor([0.1730, 0.1465, 0.1315, 0.0513, 0.0737, 0.1335, 0.1731, 0.0583], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0164, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 11:58:00,120 INFO [finetune.py:976] (0/7) Epoch 10, batch 2550, loss[loss=0.1649, simple_loss=0.2244, pruned_loss=0.05271, over 4146.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2653, pruned_loss=0.0679, over 955392.25 frames. ], batch size: 18, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:58:01,522 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8674, 1.3160, 1.8191, 1.7868, 1.5929, 1.5935, 1.7487, 1.6498], + device='cuda:0'), covar=tensor([0.4381, 0.4817, 0.3946, 0.4610, 0.5435, 0.4264, 0.5421, 0.3831], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0239, 0.0252, 0.0257, 0.0251, 0.0227, 0.0273, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 11:58:12,204 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 11:58:35,815 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.671e+02 2.051e+02 2.356e+02 3.900e+02, threshold=4.103e+02, percent-clipped=1.0 +2023-03-26 11:58:46,742 INFO [finetune.py:976] (0/7) Epoch 10, batch 2600, loss[loss=0.1948, simple_loss=0.2641, pruned_loss=0.06272, over 4886.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2676, pruned_loss=0.06843, over 955003.42 frames. ], batch size: 32, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:59:19,472 INFO [finetune.py:976] (0/7) Epoch 10, batch 2650, loss[loss=0.1723, simple_loss=0.2484, pruned_loss=0.04806, over 4892.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2672, pruned_loss=0.06779, over 955235.39 frames. ], batch size: 43, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:59:43,742 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.585e+01 1.566e+02 1.779e+02 2.159e+02 3.883e+02, threshold=3.557e+02, percent-clipped=0.0 +2023-03-26 11:59:53,473 INFO [finetune.py:976] (0/7) Epoch 10, batch 2700, loss[loss=0.2061, simple_loss=0.2627, pruned_loss=0.07473, over 4759.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2666, pruned_loss=0.06758, over 956743.71 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:00:03,697 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8525, 1.5582, 2.4562, 3.5747, 2.5138, 2.5230, 1.5178, 2.8353], + device='cuda:0'), covar=tensor([0.1681, 0.1495, 0.1155, 0.0560, 0.0712, 0.1422, 0.1587, 0.0557], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0118, 0.0135, 0.0165, 0.0102, 0.0138, 0.0127, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:00:26,576 INFO [finetune.py:976] (0/7) Epoch 10, batch 2750, loss[loss=0.2044, simple_loss=0.2566, pruned_loss=0.07616, over 4833.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2645, pruned_loss=0.06749, over 956537.43 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:00:50,902 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.545e+02 1.929e+02 2.415e+02 3.548e+02, threshold=3.859e+02, percent-clipped=0.0 +2023-03-26 12:01:01,566 INFO [finetune.py:976] (0/7) Epoch 10, batch 2800, loss[loss=0.18, simple_loss=0.2491, pruned_loss=0.05545, over 4749.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2612, pruned_loss=0.06598, over 956860.04 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:01:02,293 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7599, 1.3047, 0.8545, 1.6635, 2.1302, 1.2920, 1.5947, 1.5786], + device='cuda:0'), covar=tensor([0.1461, 0.2116, 0.2115, 0.1237, 0.1932, 0.2041, 0.1444, 0.2104], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:01:08,308 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 12:01:14,862 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 12:01:48,148 INFO [finetune.py:976] (0/7) Epoch 10, batch 2850, loss[loss=0.2321, simple_loss=0.2788, pruned_loss=0.09274, over 4133.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2595, pruned_loss=0.06537, over 954844.18 frames. ], batch size: 65, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:01:59,336 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-03-26 12:02:10,455 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.958e+01 1.628e+02 1.894e+02 2.190e+02 3.699e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-03-26 12:02:18,130 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 12:02:22,197 INFO [finetune.py:976] (0/7) Epoch 10, batch 2900, loss[loss=0.2368, simple_loss=0.3116, pruned_loss=0.08101, over 4836.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2628, pruned_loss=0.06663, over 954845.14 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:02:26,703 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-26 12:02:39,094 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4239, 1.4170, 1.3936, 0.7654, 1.3766, 1.6368, 1.6325, 1.2467], + device='cuda:0'), covar=tensor([0.0745, 0.0443, 0.0474, 0.0434, 0.0424, 0.0423, 0.0227, 0.0532], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0120, 0.0132, 0.0130, 0.0123, 0.0143, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.4363e-05, 1.1228e-04, 8.6333e-05, 9.5490e-05, 9.2851e-05, 8.9762e-05, + 1.0433e-04, 1.0675e-04], device='cuda:0') +2023-03-26 12:02:57,320 INFO [finetune.py:976] (0/7) Epoch 10, batch 2950, loss[loss=0.1779, simple_loss=0.2442, pruned_loss=0.0558, over 4778.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2658, pruned_loss=0.0678, over 955167.85 frames. ], batch size: 26, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:02:59,909 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6719, 1.8071, 1.8373, 1.1500, 1.8769, 2.1505, 2.0102, 1.5537], + device='cuda:0'), covar=tensor([0.0899, 0.0618, 0.0383, 0.0539, 0.0359, 0.0570, 0.0308, 0.0661], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0120, 0.0132, 0.0130, 0.0123, 0.0143, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.4349e-05, 1.1222e-04, 8.6424e-05, 9.5444e-05, 9.2886e-05, 8.9774e-05, + 1.0449e-04, 1.0697e-04], device='cuda:0') +2023-03-26 12:03:05,313 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 12:03:18,744 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.694e+02 2.010e+02 2.318e+02 4.609e+02, threshold=4.019e+02, percent-clipped=2.0 +2023-03-26 12:03:27,585 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9689, 1.7399, 1.4945, 1.7203, 1.6844, 1.6631, 1.7209, 2.4342], + device='cuda:0'), covar=tensor([0.4642, 0.5101, 0.4025, 0.4596, 0.4810, 0.2688, 0.4589, 0.1968], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0259, 0.0222, 0.0279, 0.0243, 0.0208, 0.0246, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:03:40,002 INFO [finetune.py:976] (0/7) Epoch 10, batch 3000, loss[loss=0.1977, simple_loss=0.2678, pruned_loss=0.06385, over 4895.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2661, pruned_loss=0.06749, over 953842.68 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:03:40,004 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 12:03:42,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8448, 1.0423, 1.9161, 1.7120, 1.5865, 1.4798, 1.5610, 1.7104], + device='cuda:0'), covar=tensor([0.4196, 0.4602, 0.4036, 0.4402, 0.5482, 0.4179, 0.5285, 0.3744], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0238, 0.0252, 0.0256, 0.0251, 0.0227, 0.0274, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:03:47,947 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8606, 1.6680, 1.5937, 1.8839, 2.1608, 1.7996, 1.4320, 1.6110], + device='cuda:0'), covar=tensor([0.1794, 0.2088, 0.1679, 0.1540, 0.1387, 0.1134, 0.2323, 0.1699], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0208, 0.0208, 0.0189, 0.0241, 0.0181, 0.0213, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:03:56,630 INFO [finetune.py:1010] (0/7) Epoch 10, validation: loss=0.1584, simple_loss=0.2295, pruned_loss=0.04366, over 2265189.00 frames. +2023-03-26 12:03:56,631 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 12:04:13,642 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-26 12:04:29,066 INFO [finetune.py:976] (0/7) Epoch 10, batch 3050, loss[loss=0.175, simple_loss=0.246, pruned_loss=0.05197, over 4899.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.265, pruned_loss=0.06637, over 954342.41 frames. ], batch size: 35, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:04:43,221 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2413, 1.6394, 2.0416, 2.0100, 1.7886, 1.8062, 1.9396, 1.9134], + device='cuda:0'), covar=tensor([0.5267, 0.6109, 0.4705, 0.5958, 0.7141, 0.5018, 0.7069, 0.4611], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0240, 0.0253, 0.0257, 0.0252, 0.0228, 0.0275, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:04:43,247 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 12:04:52,092 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.606e+02 1.839e+02 2.259e+02 4.011e+02, threshold=3.679e+02, percent-clipped=0.0 +2023-03-26 12:05:02,811 INFO [finetune.py:976] (0/7) Epoch 10, batch 3100, loss[loss=0.2333, simple_loss=0.2858, pruned_loss=0.09034, over 4780.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2634, pruned_loss=0.06577, over 951921.80 frames. ], batch size: 51, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:05:14,422 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54664.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:05:36,461 INFO [finetune.py:976] (0/7) Epoch 10, batch 3150, loss[loss=0.1647, simple_loss=0.2363, pruned_loss=0.04656, over 4935.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2602, pruned_loss=0.06469, over 952254.86 frames. ], batch size: 33, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:05:54,652 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54725.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:05:56,471 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2390, 2.0158, 2.2019, 0.9296, 2.4179, 2.6402, 2.1954, 1.9894], + device='cuda:0'), covar=tensor([0.0902, 0.0776, 0.0459, 0.0666, 0.0514, 0.0511, 0.0374, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0120, 0.0132, 0.0130, 0.0124, 0.0144, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.4257e-05, 1.1253e-04, 8.6705e-05, 9.5480e-05, 9.3294e-05, 9.0191e-05, + 1.0526e-04, 1.0733e-04], device='cuda:0') +2023-03-26 12:05:59,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.653e+02 1.993e+02 2.298e+02 5.311e+02, threshold=3.986e+02, percent-clipped=2.0 +2023-03-26 12:06:10,112 INFO [finetune.py:976] (0/7) Epoch 10, batch 3200, loss[loss=0.1606, simple_loss=0.2245, pruned_loss=0.04835, over 4714.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2578, pruned_loss=0.06411, over 953611.22 frames. ], batch size: 23, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:06:10,940 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 12:06:11,480 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.17 vs. limit=5.0 +2023-03-26 12:06:53,307 INFO [finetune.py:976] (0/7) Epoch 10, batch 3250, loss[loss=0.2132, simple_loss=0.2635, pruned_loss=0.08146, over 4739.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2579, pruned_loss=0.06507, over 952038.00 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:07:26,172 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.711e+02 2.094e+02 2.546e+02 5.601e+02, threshold=4.189e+02, percent-clipped=2.0 +2023-03-26 12:07:46,365 INFO [finetune.py:976] (0/7) Epoch 10, batch 3300, loss[loss=0.2265, simple_loss=0.2905, pruned_loss=0.08118, over 4817.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.262, pruned_loss=0.06721, over 952089.98 frames. ], batch size: 40, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:07:55,811 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54862.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:08:10,030 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 12:08:20,117 INFO [finetune.py:976] (0/7) Epoch 10, batch 3350, loss[loss=0.2294, simple_loss=0.2814, pruned_loss=0.08872, over 4787.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.264, pruned_loss=0.06724, over 952215.82 frames. ], batch size: 51, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:08:40,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9097, 3.3808, 3.6175, 3.8223, 3.6661, 3.4812, 3.9655, 1.3342], + device='cuda:0'), covar=tensor([0.0866, 0.0869, 0.0933, 0.0893, 0.1283, 0.1575, 0.0774, 0.5042], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0244, 0.0276, 0.0289, 0.0329, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:08:47,739 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54923.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:08:57,774 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.693e+02 1.965e+02 2.442e+02 4.084e+02, threshold=3.930e+02, percent-clipped=0.0 +2023-03-26 12:08:59,113 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9395, 1.8915, 1.9832, 1.3395, 1.8904, 2.0154, 2.0008, 1.5958], + device='cuda:0'), covar=tensor([0.0516, 0.0549, 0.0592, 0.0850, 0.0661, 0.0576, 0.0504, 0.1023], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0133, 0.0142, 0.0123, 0.0119, 0.0141, 0.0141, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:09:07,543 INFO [finetune.py:976] (0/7) Epoch 10, batch 3400, loss[loss=0.2126, simple_loss=0.2785, pruned_loss=0.07328, over 4900.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.265, pruned_loss=0.06744, over 953288.60 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:09:44,220 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 12:09:44,611 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1725, 1.7100, 2.3888, 3.9890, 2.8528, 2.7375, 0.8433, 3.3245], + device='cuda:0'), covar=tensor([0.1898, 0.2117, 0.1729, 0.1014, 0.0918, 0.1725, 0.2388, 0.0652], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0164, 0.0100, 0.0138, 0.0126, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:09:56,944 INFO [finetune.py:976] (0/7) Epoch 10, batch 3450, loss[loss=0.1952, simple_loss=0.2631, pruned_loss=0.0636, over 4907.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.264, pruned_loss=0.06646, over 952776.93 frames. ], batch size: 37, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:09:58,434 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9243, 1.1379, 1.8560, 1.8220, 1.6403, 1.5598, 1.6973, 1.7262], + device='cuda:0'), covar=tensor([0.3864, 0.4285, 0.3476, 0.3706, 0.4328, 0.3682, 0.4518, 0.3396], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0238, 0.0252, 0.0256, 0.0250, 0.0227, 0.0273, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:10:00,312 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 12:10:16,045 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55020.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:10:30,755 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.349e+01 1.534e+02 1.957e+02 2.350e+02 5.428e+02, threshold=3.914e+02, percent-clipped=3.0 +2023-03-26 12:10:39,045 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1205, 1.9177, 1.4568, 0.5982, 1.6318, 1.7534, 1.5891, 1.8519], + device='cuda:0'), covar=tensor([0.0844, 0.0762, 0.1292, 0.1786, 0.1287, 0.2113, 0.2124, 0.0736], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0201, 0.0201, 0.0186, 0.0215, 0.0206, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:10:51,432 INFO [finetune.py:976] (0/7) Epoch 10, batch 3500, loss[loss=0.1919, simple_loss=0.2525, pruned_loss=0.06561, over 4771.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2618, pruned_loss=0.06589, over 953563.84 frames. ], batch size: 28, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:10:52,707 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6620, 3.3548, 3.1531, 1.5988, 3.2755, 2.5945, 0.7858, 2.2484], + device='cuda:0'), covar=tensor([0.2581, 0.2052, 0.1821, 0.3275, 0.1504, 0.1066, 0.4337, 0.1616], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0173, 0.0158, 0.0127, 0.0154, 0.0122, 0.0144, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:11:36,223 INFO [finetune.py:976] (0/7) Epoch 10, batch 3550, loss[loss=0.2176, simple_loss=0.2576, pruned_loss=0.08877, over 4797.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2585, pruned_loss=0.06506, over 955064.39 frames. ], batch size: 25, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:11:37,526 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5947, 1.9820, 1.5727, 1.3950, 2.1762, 2.2010, 1.9008, 1.8683], + device='cuda:0'), covar=tensor([0.0477, 0.0346, 0.0626, 0.0486, 0.0320, 0.0483, 0.0381, 0.0384], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0109, 0.0139, 0.0114, 0.0101, 0.0103, 0.0092, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.0927e-05, 8.5410e-05, 1.1063e-04, 8.9626e-05, 7.9314e-05, 7.6471e-05, + 6.9713e-05, 8.2996e-05], device='cuda:0') +2023-03-26 12:11:58,137 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0344, 1.7752, 1.5954, 1.6026, 1.8166, 1.7483, 1.7563, 2.4753], + device='cuda:0'), covar=tensor([0.4110, 0.4067, 0.3399, 0.4033, 0.4062, 0.2465, 0.3786, 0.1728], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0222, 0.0278, 0.0242, 0.0209, 0.0244, 0.0212], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:11:58,574 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.369e+01 1.530e+02 1.896e+02 2.280e+02 4.793e+02, threshold=3.791e+02, percent-clipped=5.0 +2023-03-26 12:12:09,212 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:09,742 INFO [finetune.py:976] (0/7) Epoch 10, batch 3600, loss[loss=0.1573, simple_loss=0.2263, pruned_loss=0.04416, over 4762.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2563, pruned_loss=0.06442, over 956351.75 frames. ], batch size: 26, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:12:43,417 INFO [finetune.py:976] (0/7) Epoch 10, batch 3650, loss[loss=0.2508, simple_loss=0.3036, pruned_loss=0.09897, over 4915.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2607, pruned_loss=0.06686, over 954727.72 frames. ], batch size: 36, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:12:49,690 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:50,914 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:56,771 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55218.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:13:05,133 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7182, 1.6786, 1.5230, 1.5368, 2.0695, 1.9268, 1.7976, 1.4866], + device='cuda:0'), covar=tensor([0.0296, 0.0293, 0.0561, 0.0330, 0.0216, 0.0441, 0.0294, 0.0439], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0109, 0.0138, 0.0114, 0.0101, 0.0102, 0.0091, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.0504e-05, 8.4845e-05, 1.0998e-04, 8.9048e-05, 7.8876e-05, 7.5876e-05, + 6.9269e-05, 8.2553e-05], device='cuda:0') +2023-03-26 12:13:14,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.616e+02 1.938e+02 2.270e+02 4.700e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-26 12:13:26,509 INFO [finetune.py:976] (0/7) Epoch 10, batch 3700, loss[loss=0.2343, simple_loss=0.3008, pruned_loss=0.08389, over 4881.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2628, pruned_loss=0.06679, over 954577.64 frames. ], batch size: 32, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:13:40,330 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:14:00,007 INFO [finetune.py:976] (0/7) Epoch 10, batch 3750, loss[loss=0.257, simple_loss=0.3091, pruned_loss=0.1024, over 4732.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2644, pruned_loss=0.06696, over 954944.83 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:14:16,780 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55320.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:14:33,804 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.583e+02 1.835e+02 2.150e+02 3.880e+02, threshold=3.669e+02, percent-clipped=1.0 +2023-03-26 12:14:45,562 INFO [finetune.py:976] (0/7) Epoch 10, batch 3800, loss[loss=0.1899, simple_loss=0.2567, pruned_loss=0.06154, over 4914.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2663, pruned_loss=0.06813, over 956116.40 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:14:57,816 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55368.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:15:27,044 INFO [finetune.py:976] (0/7) Epoch 10, batch 3850, loss[loss=0.1699, simple_loss=0.2259, pruned_loss=0.057, over 4814.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2636, pruned_loss=0.06633, over 956077.55 frames. ], batch size: 25, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:15:38,105 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6344, 1.6446, 1.6374, 1.0035, 1.7786, 1.9955, 1.8430, 1.4764], + device='cuda:0'), covar=tensor([0.0998, 0.0695, 0.0506, 0.0589, 0.0456, 0.0485, 0.0357, 0.0700], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0156, 0.0123, 0.0134, 0.0132, 0.0125, 0.0146, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5531e-05, 1.1425e-04, 8.8418e-05, 9.7162e-05, 9.4680e-05, 9.0995e-05, + 1.0681e-04, 1.0779e-04], device='cuda:0') +2023-03-26 12:15:49,874 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.578e+02 1.920e+02 2.344e+02 4.809e+02, threshold=3.839e+02, percent-clipped=3.0 +2023-03-26 12:16:01,517 INFO [finetune.py:976] (0/7) Epoch 10, batch 3900, loss[loss=0.2221, simple_loss=0.2864, pruned_loss=0.07896, over 4828.00 frames. ], tot_loss[loss=0.1954, simple_loss=0.2604, pruned_loss=0.06521, over 953739.89 frames. ], batch size: 40, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:16:14,903 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1846, 1.6640, 2.0839, 1.9890, 1.7697, 1.7689, 1.9540, 1.9006], + device='cuda:0'), covar=tensor([0.4826, 0.5385, 0.4012, 0.5356, 0.5862, 0.4719, 0.6400, 0.4128], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0239, 0.0253, 0.0257, 0.0252, 0.0229, 0.0274, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:16:15,490 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1411, 2.1542, 2.1455, 1.4831, 2.2592, 2.2974, 2.2121, 1.8457], + device='cuda:0'), covar=tensor([0.0633, 0.0591, 0.0677, 0.0863, 0.0584, 0.0680, 0.0606, 0.0952], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0133, 0.0143, 0.0124, 0.0120, 0.0142, 0.0143, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:16:44,729 INFO [finetune.py:976] (0/7) Epoch 10, batch 3950, loss[loss=0.1537, simple_loss=0.2199, pruned_loss=0.04375, over 4703.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2567, pruned_loss=0.06362, over 955186.26 frames. ], batch size: 23, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:16:48,766 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:16:58,372 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55518.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:17:13,738 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.580e+02 1.857e+02 2.217e+02 3.906e+02, threshold=3.714e+02, percent-clipped=1.0 +2023-03-26 12:17:35,771 INFO [finetune.py:976] (0/7) Epoch 10, batch 4000, loss[loss=0.1626, simple_loss=0.2389, pruned_loss=0.04319, over 4791.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2563, pruned_loss=0.06375, over 955819.17 frames. ], batch size: 29, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:17:48,291 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55566.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:17:48,907 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:18:02,214 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8997, 3.7903, 3.6643, 2.1892, 3.9418, 3.1079, 1.3434, 2.7677], + device='cuda:0'), covar=tensor([0.3140, 0.1774, 0.1294, 0.2822, 0.0859, 0.0828, 0.3772, 0.1405], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0173, 0.0159, 0.0128, 0.0155, 0.0122, 0.0146, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:18:09,109 INFO [finetune.py:976] (0/7) Epoch 10, batch 4050, loss[loss=0.1974, simple_loss=0.2727, pruned_loss=0.06102, over 4821.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2596, pruned_loss=0.06535, over 955085.15 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:18:34,628 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.736e+02 2.138e+02 2.510e+02 4.140e+02, threshold=4.276e+02, percent-clipped=4.0 +2023-03-26 12:18:44,820 INFO [finetune.py:976] (0/7) Epoch 10, batch 4100, loss[loss=0.232, simple_loss=0.3066, pruned_loss=0.07873, over 4832.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2637, pruned_loss=0.0663, over 954327.78 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:19:12,862 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8806, 1.6895, 1.5061, 1.1763, 1.6768, 1.7030, 1.6343, 2.2129], + device='cuda:0'), covar=tensor([0.4707, 0.4423, 0.3633, 0.4378, 0.4188, 0.2657, 0.4055, 0.1997], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0222, 0.0278, 0.0242, 0.0209, 0.0245, 0.0212], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:19:17,499 INFO [finetune.py:976] (0/7) Epoch 10, batch 4150, loss[loss=0.1786, simple_loss=0.2595, pruned_loss=0.04886, over 4899.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2649, pruned_loss=0.06718, over 950478.14 frames. ], batch size: 36, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:19:49,995 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.688e+02 2.035e+02 2.462e+02 3.895e+02, threshold=4.069e+02, percent-clipped=0.0 +2023-03-26 12:19:59,097 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:19:59,615 INFO [finetune.py:976] (0/7) Epoch 10, batch 4200, loss[loss=0.1886, simple_loss=0.2665, pruned_loss=0.05536, over 4817.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.265, pruned_loss=0.0671, over 949821.20 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:20:53,374 INFO [finetune.py:976] (0/7) Epoch 10, batch 4250, loss[loss=0.1687, simple_loss=0.2373, pruned_loss=0.05007, over 4832.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.2633, pruned_loss=0.06717, over 951917.08 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:20:57,633 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:06,134 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55809.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:38,837 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.929e+01 1.631e+02 1.908e+02 2.201e+02 4.056e+02, threshold=3.816e+02, percent-clipped=0.0 +2023-03-26 12:21:48,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:48,566 INFO [finetune.py:976] (0/7) Epoch 10, batch 4300, loss[loss=0.2117, simple_loss=0.2749, pruned_loss=0.07424, over 4713.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.261, pruned_loss=0.0663, over 954540.57 frames. ], batch size: 23, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:21:50,369 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:02,822 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1673, 1.3785, 1.3806, 0.8219, 1.2143, 1.5834, 1.6124, 1.2587], + device='cuda:0'), covar=tensor([0.0882, 0.0465, 0.0430, 0.0473, 0.0424, 0.0513, 0.0265, 0.0628], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0155, 0.0122, 0.0133, 0.0132, 0.0125, 0.0144, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.5333e-05, 1.1366e-04, 8.7660e-05, 9.6616e-05, 9.4188e-05, 9.1103e-05, + 1.0582e-04, 1.0714e-04], device='cuda:0') +2023-03-26 12:22:10,849 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55867.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:36,974 INFO [finetune.py:976] (0/7) Epoch 10, batch 4350, loss[loss=0.2084, simple_loss=0.2587, pruned_loss=0.079, over 4222.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2574, pruned_loss=0.06496, over 954877.32 frames. ], batch size: 65, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:22:48,829 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:58,948 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:23:23,255 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.622e+02 1.928e+02 2.410e+02 3.855e+02, threshold=3.856e+02, percent-clipped=1.0 +2023-03-26 12:23:35,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3909, 3.8564, 3.9957, 4.2201, 4.1279, 3.8977, 4.4843, 1.4233], + device='cuda:0'), covar=tensor([0.0836, 0.0827, 0.0826, 0.0991, 0.1312, 0.1794, 0.0753, 0.5538], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0246, 0.0278, 0.0293, 0.0332, 0.0286, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:23:37,719 INFO [finetune.py:976] (0/7) Epoch 10, batch 4400, loss[loss=0.2127, simple_loss=0.2815, pruned_loss=0.07193, over 4922.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2572, pruned_loss=0.06421, over 954059.31 frames. ], batch size: 38, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:24:11,803 INFO [finetune.py:976] (0/7) Epoch 10, batch 4450, loss[loss=0.2316, simple_loss=0.2803, pruned_loss=0.09146, over 4715.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2623, pruned_loss=0.06573, over 954986.73 frames. ], batch size: 23, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:24:12,593 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-56000.pt +2023-03-26 12:24:16,190 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1379, 1.7935, 2.3724, 3.4497, 2.4850, 2.6428, 1.1979, 2.7854], + device='cuda:0'), covar=tensor([0.1404, 0.1247, 0.1095, 0.0580, 0.0648, 0.2167, 0.1600, 0.0554], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:24:18,763 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 12:24:36,679 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.562e+02 1.840e+02 2.258e+02 4.729e+02, threshold=3.681e+02, percent-clipped=2.0 +2023-03-26 12:24:46,912 INFO [finetune.py:976] (0/7) Epoch 10, batch 4500, loss[loss=0.2114, simple_loss=0.2728, pruned_loss=0.07498, over 4829.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2644, pruned_loss=0.06636, over 953860.03 frames. ], batch size: 30, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:25:31,191 INFO [finetune.py:976] (0/7) Epoch 10, batch 4550, loss[loss=0.1922, simple_loss=0.2553, pruned_loss=0.06452, over 4902.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2658, pruned_loss=0.06714, over 954508.06 frames. ], batch size: 32, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:25:34,314 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56104.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:25:40,426 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2219, 1.9258, 2.6865, 1.6455, 2.3460, 2.4498, 1.8257, 2.5816], + device='cuda:0'), covar=tensor([0.1261, 0.1752, 0.1339, 0.2182, 0.0824, 0.1453, 0.2334, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0204, 0.0193, 0.0190, 0.0177, 0.0215, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:25:53,264 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.686e+02 1.941e+02 2.447e+02 3.858e+02, threshold=3.882e+02, percent-clipped=3.0 +2023-03-26 12:26:02,517 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56145.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:26:04,873 INFO [finetune.py:976] (0/7) Epoch 10, batch 4600, loss[loss=0.1542, simple_loss=0.2109, pruned_loss=0.04876, over 4052.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.265, pruned_loss=0.06682, over 952198.27 frames. ], batch size: 17, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:26:04,994 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1016, 2.0062, 1.8983, 2.0851, 1.5516, 3.7771, 1.7693, 2.4808], + device='cuda:0'), covar=tensor([0.2696, 0.2122, 0.1846, 0.1906, 0.1493, 0.0179, 0.2190, 0.1003], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0098, 0.0099, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 12:26:07,564 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 12:26:40,466 INFO [finetune.py:976] (0/7) Epoch 10, batch 4650, loss[loss=0.2186, simple_loss=0.2729, pruned_loss=0.08217, over 4884.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2623, pruned_loss=0.06632, over 953324.73 frames. ], batch size: 32, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:26:43,716 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:26:44,995 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:26:51,091 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6771, 3.2778, 3.3066, 1.6615, 3.4124, 2.5903, 1.0766, 2.3545], + device='cuda:0'), covar=tensor([0.2850, 0.2154, 0.1459, 0.3033, 0.1150, 0.0999, 0.3807, 0.1541], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0175, 0.0159, 0.0129, 0.0156, 0.0122, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:26:57,342 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-03-26 12:27:11,826 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.561e+02 1.851e+02 2.355e+02 3.865e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 12:27:23,135 INFO [finetune.py:976] (0/7) Epoch 10, batch 4700, loss[loss=0.1987, simple_loss=0.2523, pruned_loss=0.07257, over 4733.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2577, pruned_loss=0.06454, over 952624.80 frames. ], batch size: 59, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:27:25,092 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7106, 1.6240, 1.5793, 1.6889, 1.1296, 3.4453, 1.3768, 1.8851], + device='cuda:0'), covar=tensor([0.3283, 0.2342, 0.1997, 0.2251, 0.1842, 0.0186, 0.2673, 0.1245], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0098, 0.0099, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 12:28:03,980 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1560, 2.2433, 1.8524, 2.5961, 2.2704, 1.9315, 2.7702, 2.3378], + device='cuda:0'), covar=tensor([0.1576, 0.2529, 0.3061, 0.2424, 0.2242, 0.1687, 0.2956, 0.1835], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0187, 0.0232, 0.0252, 0.0239, 0.0196, 0.0211, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:28:09,015 INFO [finetune.py:976] (0/7) Epoch 10, batch 4750, loss[loss=0.1363, simple_loss=0.1992, pruned_loss=0.03669, over 4826.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2549, pruned_loss=0.06386, over 951565.08 frames. ], batch size: 25, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:28:15,114 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5503, 1.4299, 2.0060, 2.7601, 1.9169, 2.1966, 1.1576, 2.2392], + device='cuda:0'), covar=tensor([0.1645, 0.1405, 0.1110, 0.0670, 0.0811, 0.1574, 0.1573, 0.0604], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:28:16,370 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5699, 1.5355, 1.5328, 1.5825, 1.2052, 2.7621, 1.2644, 1.6545], + device='cuda:0'), covar=tensor([0.3026, 0.2116, 0.1873, 0.2102, 0.1739, 0.0332, 0.2999, 0.1285], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0098, 0.0099, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 12:28:28,574 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0785, 0.8336, 0.8474, 0.9241, 1.2173, 1.1715, 1.0507, 0.9249], + device='cuda:0'), covar=tensor([0.0320, 0.0308, 0.0729, 0.0327, 0.0262, 0.0330, 0.0274, 0.0380], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0110, 0.0140, 0.0114, 0.0101, 0.0103, 0.0092, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.0935e-05, 8.5775e-05, 1.1108e-04, 8.9383e-05, 7.8821e-05, 7.6383e-05, + 6.9585e-05, 8.2696e-05], device='cuda:0') +2023-03-26 12:28:30,218 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.456e+02 1.800e+02 2.273e+02 6.888e+02, threshold=3.601e+02, percent-clipped=2.0 +2023-03-26 12:28:42,335 INFO [finetune.py:976] (0/7) Epoch 10, batch 4800, loss[loss=0.1919, simple_loss=0.2602, pruned_loss=0.06175, over 4927.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2587, pruned_loss=0.0654, over 952221.97 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:28:57,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4980, 1.5126, 1.3416, 1.5045, 1.8618, 1.6304, 1.5721, 1.2818], + device='cuda:0'), covar=tensor([0.0334, 0.0297, 0.0531, 0.0280, 0.0190, 0.0428, 0.0277, 0.0407], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0110, 0.0139, 0.0114, 0.0100, 0.0103, 0.0092, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.0844e-05, 8.5598e-05, 1.1062e-04, 8.9343e-05, 7.8499e-05, 7.6383e-05, + 6.9361e-05, 8.2579e-05], device='cuda:0') +2023-03-26 12:29:14,949 INFO [finetune.py:976] (0/7) Epoch 10, batch 4850, loss[loss=0.1891, simple_loss=0.2545, pruned_loss=0.06179, over 4843.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2632, pruned_loss=0.06709, over 952807.04 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:29:19,212 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56404.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:29:22,136 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1877, 2.1156, 2.3326, 1.6067, 2.2770, 2.3702, 2.2589, 1.9077], + device='cuda:0'), covar=tensor([0.0599, 0.0708, 0.0650, 0.0849, 0.0621, 0.0694, 0.0657, 0.1071], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0134, 0.0145, 0.0125, 0.0121, 0.0144, 0.0144, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:29:37,028 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.738e+02 2.004e+02 2.451e+02 5.164e+02, threshold=4.009e+02, percent-clipped=2.0 +2023-03-26 12:29:48,224 INFO [finetune.py:976] (0/7) Epoch 10, batch 4900, loss[loss=0.1761, simple_loss=0.2366, pruned_loss=0.05778, over 4770.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2668, pruned_loss=0.06912, over 954167.82 frames. ], batch size: 26, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:29:48,294 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1189, 4.6038, 4.4469, 2.6473, 4.6973, 3.5573, 0.7068, 3.2618], + device='cuda:0'), covar=tensor([0.2254, 0.1492, 0.1216, 0.2883, 0.0758, 0.0801, 0.4966, 0.1282], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0175, 0.0160, 0.0130, 0.0156, 0.0122, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:29:50,492 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56452.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:05,951 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 12:30:26,420 INFO [finetune.py:976] (0/7) Epoch 10, batch 4950, loss[loss=0.2371, simple_loss=0.2964, pruned_loss=0.08892, over 4930.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.267, pruned_loss=0.06902, over 955375.76 frames. ], batch size: 41, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:30:32,438 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:34,355 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:49,789 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56523.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 12:30:55,719 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.612e+02 1.965e+02 2.275e+02 4.231e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-26 12:31:06,813 INFO [finetune.py:976] (0/7) Epoch 10, batch 5000, loss[loss=0.188, simple_loss=0.2588, pruned_loss=0.0586, over 4818.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.265, pruned_loss=0.06776, over 956255.22 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:31:08,672 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:31:12,876 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6374, 1.7390, 1.7681, 1.0303, 1.8570, 2.0223, 2.0729, 1.5529], + device='cuda:0'), covar=tensor([0.1197, 0.0734, 0.0534, 0.0573, 0.0582, 0.0517, 0.0330, 0.0870], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0153, 0.0121, 0.0132, 0.0131, 0.0124, 0.0144, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.5044e-05, 1.1261e-04, 8.7368e-05, 9.5806e-05, 9.3861e-05, 9.0568e-05, + 1.0545e-04, 1.0702e-04], device='cuda:0') +2023-03-26 12:31:21,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2285, 3.7035, 3.8284, 4.0511, 3.9746, 3.6761, 4.3083, 1.3135], + device='cuda:0'), covar=tensor([0.0776, 0.0744, 0.0762, 0.0932, 0.1223, 0.1747, 0.0718, 0.5396], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0247, 0.0279, 0.0293, 0.0335, 0.0288, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:31:29,649 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56584.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:31:39,223 INFO [finetune.py:976] (0/7) Epoch 10, batch 5050, loss[loss=0.1882, simple_loss=0.2546, pruned_loss=0.0609, over 4825.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2609, pruned_loss=0.06549, over 958003.79 frames. ], batch size: 40, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:32:04,806 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.580e+02 1.790e+02 2.049e+02 5.062e+02, threshold=3.579e+02, percent-clipped=1.0 +2023-03-26 12:32:14,688 INFO [finetune.py:976] (0/7) Epoch 10, batch 5100, loss[loss=0.1937, simple_loss=0.2544, pruned_loss=0.06651, over 4128.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2565, pruned_loss=0.06379, over 955943.04 frames. ], batch size: 18, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:32:55,111 INFO [finetune.py:976] (0/7) Epoch 10, batch 5150, loss[loss=0.1777, simple_loss=0.2485, pruned_loss=0.05347, over 4778.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2564, pruned_loss=0.06365, over 956067.93 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:33:27,187 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.919e+01 1.632e+02 1.974e+02 2.331e+02 5.610e+02, threshold=3.948e+02, percent-clipped=3.0 +2023-03-26 12:33:35,154 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9702, 1.8203, 2.2224, 2.3506, 2.0122, 3.8825, 1.8710, 1.9660], + device='cuda:0'), covar=tensor([0.0874, 0.1563, 0.0957, 0.0817, 0.1367, 0.0266, 0.1265, 0.1532], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0075, 0.0078, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 12:33:36,882 INFO [finetune.py:976] (0/7) Epoch 10, batch 5200, loss[loss=0.204, simple_loss=0.2814, pruned_loss=0.06325, over 4808.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2618, pruned_loss=0.06558, over 953793.33 frames. ], batch size: 39, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:10,223 INFO [finetune.py:976] (0/7) Epoch 10, batch 5250, loss[loss=0.1616, simple_loss=0.2385, pruned_loss=0.04237, over 4765.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2628, pruned_loss=0.06584, over 952608.71 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:11,646 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:12,866 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56803.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:34,253 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.706e+02 2.047e+02 2.503e+02 5.084e+02, threshold=4.093e+02, percent-clipped=2.0 +2023-03-26 12:34:43,553 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 12:34:43,965 INFO [finetune.py:976] (0/7) Epoch 10, batch 5300, loss[loss=0.1744, simple_loss=0.2587, pruned_loss=0.04507, over 4842.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2645, pruned_loss=0.06692, over 950524.58 frames. ], batch size: 49, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:44,028 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:53,659 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56864.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:59,481 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0487, 1.6386, 2.3003, 1.5759, 2.1656, 2.1505, 1.5656, 2.3797], + device='cuda:0'), covar=tensor([0.1387, 0.2408, 0.1438, 0.2095, 0.0938, 0.1591, 0.2984, 0.0866], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0207, 0.0196, 0.0191, 0.0179, 0.0216, 0.0220, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:35:04,616 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56879.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:35:17,628 INFO [finetune.py:976] (0/7) Epoch 10, batch 5350, loss[loss=0.1896, simple_loss=0.2552, pruned_loss=0.062, over 4859.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2644, pruned_loss=0.06637, over 949981.67 frames. ], batch size: 47, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:35:23,293 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56908.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:35:49,118 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.924e+01 1.584e+02 1.842e+02 2.192e+02 3.665e+02, threshold=3.684e+02, percent-clipped=0.0 +2023-03-26 12:36:02,275 INFO [finetune.py:976] (0/7) Epoch 10, batch 5400, loss[loss=0.1814, simple_loss=0.26, pruned_loss=0.05141, over 4868.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2612, pruned_loss=0.06515, over 950014.54 frames. ], batch size: 34, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:36:15,010 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:32,492 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56993.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:33,660 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:35,992 INFO [finetune.py:976] (0/7) Epoch 10, batch 5450, loss[loss=0.1963, simple_loss=0.2467, pruned_loss=0.07295, over 4808.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2583, pruned_loss=0.06421, over 950843.79 frames. ], batch size: 25, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:36:57,713 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.491e+02 1.807e+02 2.344e+02 4.842e+02, threshold=3.613e+02, percent-clipped=5.0 +2023-03-26 12:37:07,817 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8825, 1.0913, 1.8390, 1.7587, 1.6191, 1.6004, 1.6645, 1.7172], + device='cuda:0'), covar=tensor([0.3915, 0.4462, 0.3890, 0.3970, 0.5066, 0.3867, 0.4732, 0.3766], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0240, 0.0254, 0.0258, 0.0253, 0.0230, 0.0275, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:37:09,488 INFO [finetune.py:976] (0/7) Epoch 10, batch 5500, loss[loss=0.2169, simple_loss=0.2764, pruned_loss=0.07868, over 4778.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2554, pruned_loss=0.06335, over 952304.68 frames. ], batch size: 59, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:37:12,655 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57054.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:37:13,839 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:37:22,481 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 12:37:26,612 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1803, 1.5059, 0.6675, 2.1614, 2.4943, 1.7594, 1.5920, 2.0211], + device='cuda:0'), covar=tensor([0.1449, 0.2087, 0.2322, 0.1156, 0.1931, 0.2147, 0.1580, 0.2022], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0112, 0.0092, 0.0120, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:37:43,350 INFO [finetune.py:976] (0/7) Epoch 10, batch 5550, loss[loss=0.2693, simple_loss=0.3239, pruned_loss=0.1073, over 4203.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2593, pruned_loss=0.06561, over 952651.95 frames. ], batch size: 65, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:38:06,740 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.869e+01 1.587e+02 1.788e+02 2.090e+02 3.209e+02, threshold=3.576e+02, percent-clipped=0.0 +2023-03-26 12:38:25,602 INFO [finetune.py:976] (0/7) Epoch 10, batch 5600, loss[loss=0.232, simple_loss=0.3041, pruned_loss=0.07993, over 4841.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2633, pruned_loss=0.06659, over 952091.96 frames. ], batch size: 44, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:38:33,393 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 12:38:35,033 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57159.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:46,650 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57179.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:38:50,752 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:53,040 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9567, 1.5245, 2.2379, 1.5292, 1.9946, 2.0380, 1.5672, 2.2930], + device='cuda:0'), covar=tensor([0.1254, 0.2322, 0.1080, 0.1804, 0.0939, 0.1511, 0.2832, 0.0913], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0207, 0.0195, 0.0192, 0.0178, 0.0217, 0.0219, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:38:57,554 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57197.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:58,646 INFO [finetune.py:976] (0/7) Epoch 10, batch 5650, loss[loss=0.1749, simple_loss=0.2603, pruned_loss=0.04477, over 4802.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2667, pruned_loss=0.06738, over 954436.46 frames. ], batch size: 41, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:39:00,796 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 12:39:15,284 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57227.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:39:15,950 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3550, 2.1349, 1.9000, 2.2389, 2.0556, 2.0535, 2.0637, 2.8042], + device='cuda:0'), covar=tensor([0.4444, 0.4955, 0.3761, 0.4024, 0.4051, 0.2710, 0.4067, 0.1750], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0258, 0.0223, 0.0278, 0.0243, 0.0209, 0.0245, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:39:19,289 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.551e+02 1.804e+02 2.162e+02 3.713e+02, threshold=3.608e+02, percent-clipped=1.0 +2023-03-26 12:39:25,341 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:39:27,113 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:39:28,225 INFO [finetune.py:976] (0/7) Epoch 10, batch 5700, loss[loss=0.1845, simple_loss=0.2415, pruned_loss=0.06375, over 4473.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2635, pruned_loss=0.0677, over 934206.09 frames. ], batch size: 19, lr: 3.72e-03, grad_scale: 32.0 +2023-03-26 12:39:31,704 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.55 vs. limit=5.0 +2023-03-26 12:39:34,008 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57258.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:39:37,767 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57264.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:39:44,256 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3850, 1.0638, 0.7531, 1.2801, 1.7971, 0.7025, 1.2022, 1.3197], + device='cuda:0'), covar=tensor([0.1310, 0.1930, 0.1568, 0.1059, 0.1696, 0.1769, 0.1338, 0.1820], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:39:45,613 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-10.pt +2023-03-26 12:40:00,694 INFO [finetune.py:976] (0/7) Epoch 11, batch 0, loss[loss=0.2161, simple_loss=0.2822, pruned_loss=0.07493, over 4814.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2822, pruned_loss=0.07493, over 4814.00 frames. ], batch size: 33, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:40:00,695 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 12:40:16,056 INFO [finetune.py:1010] (0/7) Epoch 11, validation: loss=0.1597, simple_loss=0.2306, pruned_loss=0.04438, over 2265189.00 frames. +2023-03-26 12:40:16,056 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 12:40:37,176 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57305.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:40:50,176 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 12:40:59,547 INFO [finetune.py:976] (0/7) Epoch 11, batch 50, loss[loss=0.2065, simple_loss=0.2774, pruned_loss=0.06779, over 4776.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2685, pruned_loss=0.06828, over 217279.03 frames. ], batch size: 29, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:41:07,736 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3596, 1.4087, 1.2788, 1.3940, 1.7441, 1.5178, 1.4332, 1.2492], + device='cuda:0'), covar=tensor([0.0354, 0.0261, 0.0532, 0.0269, 0.0178, 0.0525, 0.0315, 0.0355], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0108, 0.0137, 0.0112, 0.0099, 0.0101, 0.0091, 0.0106], + device='cuda:0'), out_proj_covar=tensor([7.0171e-05, 8.4021e-05, 1.0866e-04, 8.7867e-05, 7.7660e-05, 7.5156e-05, + 6.8920e-05, 8.1553e-05], device='cuda:0') +2023-03-26 12:41:10,028 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.580e+02 1.868e+02 2.535e+02 4.204e+02, threshold=3.735e+02, percent-clipped=3.0 +2023-03-26 12:41:18,560 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57349.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:41:19,785 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57351.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:41:38,106 INFO [finetune.py:976] (0/7) Epoch 11, batch 100, loss[loss=0.1656, simple_loss=0.2319, pruned_loss=0.04963, over 4872.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2592, pruned_loss=0.06543, over 381037.33 frames. ], batch size: 34, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:41:50,234 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8018, 1.7538, 1.9130, 1.2551, 1.9638, 2.0030, 1.8389, 1.5546], + device='cuda:0'), covar=tensor([0.0518, 0.0648, 0.0621, 0.0820, 0.0600, 0.0590, 0.0596, 0.1061], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0125, 0.0120, 0.0144, 0.0145, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:41:51,419 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57398.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:11,570 INFO [finetune.py:976] (0/7) Epoch 11, batch 150, loss[loss=0.2325, simple_loss=0.2821, pruned_loss=0.09144, over 4902.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2554, pruned_loss=0.06394, over 508091.27 frames. ], batch size: 43, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:42:16,966 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.728e+02 2.070e+02 2.489e+02 4.280e+02, threshold=4.140e+02, percent-clipped=3.0 +2023-03-26 12:42:31,647 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:31,673 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:33,508 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57462.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:44,007 INFO [finetune.py:976] (0/7) Epoch 11, batch 200, loss[loss=0.2143, simple_loss=0.2815, pruned_loss=0.07355, over 4855.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2536, pruned_loss=0.06369, over 608001.89 frames. ], batch size: 44, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:42:56,411 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:03,705 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:14,469 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57523.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:17,310 INFO [finetune.py:976] (0/7) Epoch 11, batch 250, loss[loss=0.2196, simple_loss=0.2811, pruned_loss=0.07907, over 4822.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2589, pruned_loss=0.06583, over 685356.97 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:43:19,742 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9797, 1.6791, 1.9580, 1.9523, 1.7040, 1.7167, 1.8972, 1.7967], + device='cuda:0'), covar=tensor([0.4366, 0.4786, 0.3792, 0.4537, 0.5730, 0.4072, 0.5816, 0.3732], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0238, 0.0253, 0.0256, 0.0252, 0.0229, 0.0273, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:43:22,637 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.287e+02 1.584e+02 1.966e+02 2.356e+02 4.681e+02, threshold=3.932e+02, percent-clipped=1.0 +2023-03-26 12:43:27,966 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57542.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:43,764 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57553.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:43:45,612 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:55,216 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57564.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:08,372 INFO [finetune.py:976] (0/7) Epoch 11, batch 300, loss[loss=0.1371, simple_loss=0.2156, pruned_loss=0.02929, over 4680.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2618, pruned_loss=0.06609, over 746142.09 frames. ], batch size: 23, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:44:12,555 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0136, 4.1277, 3.8278, 2.0621, 4.2383, 3.1472, 0.8304, 2.7454], + device='cuda:0'), covar=tensor([0.2137, 0.2224, 0.1434, 0.3326, 0.0976, 0.1003, 0.4800, 0.1562], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0159, 0.0128, 0.0155, 0.0121, 0.0144, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:44:24,433 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57600.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:31,732 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:40,866 INFO [finetune.py:976] (0/7) Epoch 11, batch 350, loss[loss=0.1756, simple_loss=0.2517, pruned_loss=0.04979, over 4907.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2625, pruned_loss=0.06602, over 793140.35 frames. ], batch size: 37, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:44:46,741 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.573e+02 1.819e+02 2.403e+02 4.156e+02, threshold=3.639e+02, percent-clipped=1.0 +2023-03-26 12:44:56,805 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:58,452 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:14,036 INFO [finetune.py:976] (0/7) Epoch 11, batch 400, loss[loss=0.1738, simple_loss=0.241, pruned_loss=0.05328, over 4805.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2628, pruned_loss=0.06576, over 829210.89 frames. ], batch size: 55, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:45:30,904 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57697.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:32,138 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:35,178 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 12:45:49,609 INFO [finetune.py:976] (0/7) Epoch 11, batch 450, loss[loss=0.2051, simple_loss=0.2676, pruned_loss=0.07135, over 4896.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2624, pruned_loss=0.06595, over 856753.90 frames. ], batch size: 46, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:45:53,389 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57733.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:55,473 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.602e+02 1.902e+02 2.220e+02 3.989e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-26 12:46:15,574 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57754.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:46:17,392 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6683, 3.3746, 3.2274, 1.7305, 3.5037, 2.5384, 1.1485, 2.3181], + device='cuda:0'), covar=tensor([0.2080, 0.1929, 0.1606, 0.3403, 0.1237, 0.1125, 0.4239, 0.1586], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0172, 0.0158, 0.0127, 0.0155, 0.0121, 0.0144, 0.0120], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:46:32,834 INFO [finetune.py:976] (0/7) Epoch 11, batch 500, loss[loss=0.2167, simple_loss=0.2624, pruned_loss=0.0855, over 4855.00 frames. ], tot_loss[loss=0.1954, simple_loss=0.2603, pruned_loss=0.06527, over 879885.74 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:46:45,684 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57794.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:01,348 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57818.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:06,699 INFO [finetune.py:976] (0/7) Epoch 11, batch 550, loss[loss=0.1697, simple_loss=0.2382, pruned_loss=0.05055, over 4844.00 frames. ], tot_loss[loss=0.192, simple_loss=0.256, pruned_loss=0.06398, over 897356.31 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:47:11,533 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.635e+02 1.936e+02 2.160e+02 3.511e+02, threshold=3.871e+02, percent-clipped=0.0 +2023-03-26 12:47:16,809 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:23,737 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:24,968 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57853.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:47:27,446 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0010, 2.0872, 2.0535, 1.4179, 2.2195, 2.3040, 2.1142, 1.8163], + device='cuda:0'), covar=tensor([0.0694, 0.0640, 0.0828, 0.1021, 0.0594, 0.0699, 0.0698, 0.1082], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0125, 0.0120, 0.0144, 0.0145, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:47:30,536 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 12:47:40,101 INFO [finetune.py:976] (0/7) Epoch 11, batch 600, loss[loss=0.1392, simple_loss=0.2112, pruned_loss=0.03363, over 4753.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2564, pruned_loss=0.06389, over 911350.67 frames. ], batch size: 27, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:47:47,993 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57890.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:56,178 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:56,756 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57901.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:48:13,626 INFO [finetune.py:976] (0/7) Epoch 11, batch 650, loss[loss=0.2302, simple_loss=0.2847, pruned_loss=0.08783, over 4812.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2594, pruned_loss=0.06445, over 922094.10 frames. ], batch size: 25, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:48:18,497 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.568e+02 1.897e+02 2.360e+02 4.682e+02, threshold=3.793e+02, percent-clipped=3.0 +2023-03-26 12:48:27,444 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:48:38,226 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 12:48:38,618 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0996, 1.2042, 1.0399, 1.2916, 1.2827, 2.5106, 1.0328, 1.2540], + device='cuda:0'), covar=tensor([0.1314, 0.2513, 0.1528, 0.1255, 0.2137, 0.0483, 0.2172, 0.2547], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 12:48:48,746 INFO [finetune.py:976] (0/7) Epoch 11, batch 700, loss[loss=0.1894, simple_loss=0.2616, pruned_loss=0.05854, over 4894.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.262, pruned_loss=0.06556, over 928606.04 frames. ], batch size: 35, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:48:58,287 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6181, 1.1676, 0.9040, 1.4991, 2.0399, 1.0273, 1.3792, 1.5967], + device='cuda:0'), covar=tensor([0.1488, 0.2170, 0.2022, 0.1283, 0.1955, 0.2062, 0.1564, 0.1892], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0096, 0.0112, 0.0093, 0.0120, 0.0094, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:49:15,781 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-58000.pt +2023-03-26 12:49:24,676 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0104, 1.9279, 1.8223, 2.1496, 2.5752, 2.1770, 1.8516, 1.5296], + device='cuda:0'), covar=tensor([0.2298, 0.2203, 0.1939, 0.1650, 0.2029, 0.1234, 0.2364, 0.2102], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0207, 0.0208, 0.0188, 0.0242, 0.0180, 0.0213, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:49:36,920 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6835, 3.6380, 3.5436, 1.8786, 3.7991, 2.7709, 0.7565, 2.5305], + device='cuda:0'), covar=tensor([0.2355, 0.1918, 0.1361, 0.2983, 0.1001, 0.0969, 0.4198, 0.1398], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0171, 0.0156, 0.0126, 0.0154, 0.0120, 0.0143, 0.0120], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 12:49:43,403 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5579, 1.2065, 0.9482, 1.4923, 1.9234, 1.0729, 1.3959, 1.5545], + device='cuda:0'), covar=tensor([0.1514, 0.2045, 0.1848, 0.1217, 0.2005, 0.1973, 0.1424, 0.1908], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0097, 0.0113, 0.0093, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:49:44,567 INFO [finetune.py:976] (0/7) Epoch 11, batch 750, loss[loss=0.1869, simple_loss=0.2629, pruned_loss=0.05544, over 4893.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2633, pruned_loss=0.06602, over 934497.56 frames. ], batch size: 43, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:49:49,411 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.116e+01 1.579e+02 1.894e+02 2.321e+02 4.436e+02, threshold=3.789e+02, percent-clipped=3.0 +2023-03-26 12:50:02,173 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58054.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:18,110 INFO [finetune.py:976] (0/7) Epoch 11, batch 800, loss[loss=0.1972, simple_loss=0.2665, pruned_loss=0.06391, over 4812.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2625, pruned_loss=0.06467, over 940139.77 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:50:18,826 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3718, 1.4260, 1.2382, 1.4658, 1.7928, 1.6632, 1.4670, 1.2722], + device='cuda:0'), covar=tensor([0.0407, 0.0299, 0.0654, 0.0288, 0.0190, 0.0405, 0.0320, 0.0412], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0109, 0.0140, 0.0115, 0.0101, 0.0104, 0.0093, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.1688e-05, 8.5241e-05, 1.1173e-04, 8.9721e-05, 7.9342e-05, 7.7096e-05, + 7.0622e-05, 8.3390e-05], device='cuda:0') +2023-03-26 12:50:25,466 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:33,870 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:45,544 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:51,415 INFO [finetune.py:976] (0/7) Epoch 11, batch 850, loss[loss=0.1858, simple_loss=0.2602, pruned_loss=0.05572, over 4788.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2606, pruned_loss=0.06453, over 942623.76 frames. ], batch size: 29, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:50:56,226 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.394e+01 1.505e+02 1.749e+02 2.082e+02 4.545e+02, threshold=3.498e+02, percent-clipped=2.0 +2023-03-26 12:50:59,938 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58141.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:01,757 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:05,917 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58151.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:12,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1158, 2.1858, 1.6053, 2.3111, 2.0684, 1.8039, 2.9534, 2.1636], + device='cuda:0'), covar=tensor([0.1403, 0.2356, 0.3288, 0.2967, 0.2737, 0.1733, 0.2216, 0.1952], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0187, 0.0232, 0.0254, 0.0239, 0.0196, 0.0212, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:51:22,648 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:35,958 INFO [finetune.py:976] (0/7) Epoch 11, batch 900, loss[loss=0.1989, simple_loss=0.2587, pruned_loss=0.06955, over 4836.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2576, pruned_loss=0.06353, over 947153.82 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:51:57,408 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:59,451 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:52:01,283 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58205.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:52:05,942 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 12:52:17,503 INFO [finetune.py:976] (0/7) Epoch 11, batch 950, loss[loss=0.2409, simple_loss=0.3058, pruned_loss=0.08796, over 4923.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2565, pruned_loss=0.06346, over 948627.65 frames. ], batch size: 42, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:52:22,879 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.516e+02 1.975e+02 2.310e+02 4.008e+02, threshold=3.950e+02, percent-clipped=1.0 +2023-03-26 12:52:27,418 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-03-26 12:52:47,929 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0802, 2.0003, 1.5871, 1.9104, 1.9576, 1.6666, 2.2640, 2.0932], + device='cuda:0'), covar=tensor([0.1282, 0.2064, 0.3068, 0.2673, 0.2669, 0.1700, 0.3851, 0.1793], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0187, 0.0232, 0.0253, 0.0238, 0.0195, 0.0211, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:52:50,911 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0660, 1.4133, 0.7979, 2.0290, 2.1960, 1.7419, 1.6428, 1.9191], + device='cuda:0'), covar=tensor([0.1386, 0.2000, 0.1969, 0.1147, 0.2005, 0.1820, 0.1342, 0.1900], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0097, 0.0113, 0.0093, 0.0121, 0.0095, 0.0099, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:52:51,452 INFO [finetune.py:976] (0/7) Epoch 11, batch 1000, loss[loss=0.2196, simple_loss=0.2874, pruned_loss=0.07591, over 4820.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2608, pruned_loss=0.0651, over 952472.88 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:53:31,213 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4650, 2.1786, 2.0269, 2.3143, 2.0939, 2.1062, 2.1339, 2.8472], + device='cuda:0'), covar=tensor([0.4547, 0.4930, 0.3743, 0.4548, 0.4570, 0.2842, 0.4521, 0.1937], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0258, 0.0222, 0.0277, 0.0241, 0.0209, 0.0245, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:53:46,405 INFO [finetune.py:976] (0/7) Epoch 11, batch 1050, loss[loss=0.1506, simple_loss=0.2194, pruned_loss=0.04095, over 4764.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2629, pruned_loss=0.06546, over 953507.13 frames. ], batch size: 28, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:53:51,319 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.197e+02 1.617e+02 2.003e+02 2.375e+02 3.670e+02, threshold=4.006e+02, percent-clipped=0.0 +2023-03-26 12:54:42,548 INFO [finetune.py:976] (0/7) Epoch 11, batch 1100, loss[loss=0.2349, simple_loss=0.2916, pruned_loss=0.08908, over 4921.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2639, pruned_loss=0.06584, over 953763.77 frames. ], batch size: 33, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:54:55,679 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:55:35,377 INFO [finetune.py:976] (0/7) Epoch 11, batch 1150, loss[loss=0.1756, simple_loss=0.2597, pruned_loss=0.04577, over 4828.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2655, pruned_loss=0.06642, over 955265.34 frames. ], batch size: 49, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:55:38,967 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0809, 1.9141, 1.6351, 1.8307, 1.7915, 1.7868, 1.8476, 2.6055], + device='cuda:0'), covar=tensor([0.4079, 0.4735, 0.3536, 0.4633, 0.4785, 0.2526, 0.4135, 0.1822], + device='cuda:0'), in_proj_covar=tensor([0.0282, 0.0258, 0.0221, 0.0276, 0.0241, 0.0208, 0.0244, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:55:40,653 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.672e+02 1.870e+02 2.321e+02 4.403e+02, threshold=3.740e+02, percent-clipped=1.0 +2023-03-26 12:55:41,941 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58437.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:56:08,442 INFO [finetune.py:976] (0/7) Epoch 11, batch 1200, loss[loss=0.2056, simple_loss=0.2743, pruned_loss=0.06839, over 4905.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2632, pruned_loss=0.06558, over 954741.04 frames. ], batch size: 36, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:56:21,455 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58497.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:56:23,262 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58500.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:56:40,039 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1449, 1.8974, 1.6722, 1.7381, 1.7961, 1.8032, 1.8534, 2.6010], + device='cuda:0'), covar=tensor([0.4278, 0.5144, 0.3774, 0.4570, 0.4401, 0.2581, 0.4150, 0.1774], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0223, 0.0278, 0.0242, 0.0209, 0.0246, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:56:40,496 INFO [finetune.py:976] (0/7) Epoch 11, batch 1250, loss[loss=0.1783, simple_loss=0.2451, pruned_loss=0.05571, over 4772.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2606, pruned_loss=0.06521, over 953613.15 frames. ], batch size: 28, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:56:46,790 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.332e+01 1.581e+02 1.822e+02 2.261e+02 4.369e+02, threshold=3.644e+02, percent-clipped=3.0 +2023-03-26 12:57:15,450 INFO [finetune.py:976] (0/7) Epoch 11, batch 1300, loss[loss=0.1871, simple_loss=0.245, pruned_loss=0.06461, over 4874.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2577, pruned_loss=0.06394, over 955705.68 frames. ], batch size: 31, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:57:48,895 INFO [finetune.py:976] (0/7) Epoch 11, batch 1350, loss[loss=0.1946, simple_loss=0.2658, pruned_loss=0.0617, over 4855.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2579, pruned_loss=0.06436, over 955974.52 frames. ], batch size: 49, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:57:54,728 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.484e+01 1.581e+02 1.914e+02 2.266e+02 4.857e+02, threshold=3.829e+02, percent-clipped=2.0 +2023-03-26 12:58:01,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1960, 2.1367, 1.6369, 2.1890, 2.0094, 1.7853, 2.3380, 2.2003], + device='cuda:0'), covar=tensor([0.1445, 0.2426, 0.3171, 0.2946, 0.2941, 0.1753, 0.5007, 0.1923], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0188, 0.0233, 0.0255, 0.0241, 0.0196, 0.0212, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 12:58:23,950 INFO [finetune.py:976] (0/7) Epoch 11, batch 1400, loss[loss=0.2015, simple_loss=0.2875, pruned_loss=0.05774, over 4807.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2617, pruned_loss=0.06554, over 955227.52 frames. ], batch size: 45, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:58:26,990 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58681.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:58:56,046 INFO [finetune.py:976] (0/7) Epoch 11, batch 1450, loss[loss=0.2172, simple_loss=0.2771, pruned_loss=0.07864, over 4824.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2638, pruned_loss=0.06641, over 954148.67 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:59:01,959 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.669e+02 2.009e+02 2.318e+02 4.324e+02, threshold=4.017e+02, percent-clipped=1.0 +2023-03-26 12:59:07,263 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58742.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:59:12,587 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8211, 1.3877, 0.9942, 1.7214, 2.2675, 1.4403, 1.6261, 1.6646], + device='cuda:0'), covar=tensor([0.1526, 0.2113, 0.2035, 0.1210, 0.1854, 0.1854, 0.1467, 0.2030], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0097, 0.0113, 0.0093, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 12:59:35,092 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58775.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:59:36,228 INFO [finetune.py:976] (0/7) Epoch 11, batch 1500, loss[loss=0.2087, simple_loss=0.2846, pruned_loss=0.06643, over 4809.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2636, pruned_loss=0.0661, over 953942.05 frames. ], batch size: 47, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 12:59:57,861 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58797.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:04,298 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58800.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:16,542 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1620, 1.7526, 2.2107, 2.0507, 1.7720, 1.7790, 1.9973, 1.9306], + device='cuda:0'), covar=tensor([0.4370, 0.4984, 0.3622, 0.4660, 0.5712, 0.4266, 0.5808, 0.3774], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0237, 0.0252, 0.0255, 0.0252, 0.0228, 0.0272, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:00:34,186 INFO [finetune.py:976] (0/7) Epoch 11, batch 1550, loss[loss=0.164, simple_loss=0.2342, pruned_loss=0.04695, over 4926.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2637, pruned_loss=0.06554, over 954703.13 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:00:39,951 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.569e+02 1.959e+02 2.197e+02 4.059e+02, threshold=3.918e+02, percent-clipped=1.0 +2023-03-26 13:00:41,220 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58836.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:47,045 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1686, 1.7705, 2.4777, 3.8406, 2.6766, 2.7671, 0.7684, 3.0497], + device='cuda:0'), covar=tensor([0.1619, 0.1480, 0.1268, 0.0551, 0.0733, 0.1684, 0.2056, 0.0540], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0131, 0.0162, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:00:47,623 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:49,460 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:01:07,920 INFO [finetune.py:976] (0/7) Epoch 11, batch 1600, loss[loss=0.1908, simple_loss=0.2587, pruned_loss=0.06142, over 4790.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2627, pruned_loss=0.06538, over 953502.19 frames. ], batch size: 29, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:01:28,384 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:01:50,384 INFO [finetune.py:976] (0/7) Epoch 11, batch 1650, loss[loss=0.1761, simple_loss=0.2413, pruned_loss=0.05546, over 4874.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2607, pruned_loss=0.06512, over 954520.52 frames. ], batch size: 34, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:01:55,256 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.664e+02 1.923e+02 2.390e+02 4.121e+02, threshold=3.846e+02, percent-clipped=1.0 +2023-03-26 13:02:18,274 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58967.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:02:19,506 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5805, 2.4435, 2.0284, 1.0465, 2.2907, 1.9582, 1.8034, 2.1423], + device='cuda:0'), covar=tensor([0.0805, 0.0634, 0.1466, 0.1885, 0.1284, 0.1980, 0.1845, 0.0950], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0200, 0.0202, 0.0188, 0.0215, 0.0209, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:02:24,167 INFO [finetune.py:976] (0/7) Epoch 11, batch 1700, loss[loss=0.1793, simple_loss=0.2501, pruned_loss=0.0542, over 4911.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2579, pruned_loss=0.06407, over 956373.45 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:02:57,898 INFO [finetune.py:976] (0/7) Epoch 11, batch 1750, loss[loss=0.1769, simple_loss=0.2422, pruned_loss=0.05579, over 4777.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2604, pruned_loss=0.0651, over 956820.97 frames. ], batch size: 28, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:03:02,755 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.620e+02 1.895e+02 2.249e+02 5.052e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-03-26 13:03:04,068 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:03:29,777 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6581, 1.2589, 0.9269, 1.5656, 2.0128, 1.4374, 1.4914, 1.4724], + device='cuda:0'), covar=tensor([0.1655, 0.2232, 0.2102, 0.1415, 0.2204, 0.2114, 0.1521, 0.2308], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0093, 0.0120, 0.0095, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:03:33,678 INFO [finetune.py:976] (0/7) Epoch 11, batch 1800, loss[loss=0.2001, simple_loss=0.271, pruned_loss=0.06455, over 4749.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.263, pruned_loss=0.06529, over 957773.47 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:04:04,735 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9432, 1.8652, 1.5871, 1.6389, 1.7443, 1.6895, 1.7168, 2.4735], + device='cuda:0'), covar=tensor([0.4257, 0.4567, 0.3441, 0.4252, 0.4308, 0.2515, 0.4106, 0.1649], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0260, 0.0225, 0.0280, 0.0244, 0.0211, 0.0247, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:04:19,595 INFO [finetune.py:976] (0/7) Epoch 11, batch 1850, loss[loss=0.229, simple_loss=0.2859, pruned_loss=0.08612, over 4883.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2654, pruned_loss=0.0663, over 957862.53 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:04:22,092 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59131.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:04:24,430 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.668e+02 2.065e+02 2.636e+02 4.490e+02, threshold=4.130e+02, percent-clipped=5.0 +2023-03-26 13:04:36,458 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-26 13:04:44,002 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:04:57,370 INFO [finetune.py:976] (0/7) Epoch 11, batch 1900, loss[loss=0.2563, simple_loss=0.3043, pruned_loss=0.1042, over 4802.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2676, pruned_loss=0.06732, over 958665.73 frames. ], batch size: 45, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:04:58,072 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8859, 3.6633, 3.4482, 1.7835, 3.8028, 2.8020, 0.7423, 2.4164], + device='cuda:0'), covar=tensor([0.2145, 0.1544, 0.1612, 0.3327, 0.1057, 0.1043, 0.4726, 0.1563], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0174, 0.0160, 0.0129, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 13:05:46,812 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:05:47,935 INFO [finetune.py:976] (0/7) Epoch 11, batch 1950, loss[loss=0.1984, simple_loss=0.2555, pruned_loss=0.07068, over 4876.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2658, pruned_loss=0.06652, over 956362.31 frames. ], batch size: 31, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:05:59,308 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.570e+02 1.817e+02 2.294e+02 4.310e+02, threshold=3.633e+02, percent-clipped=1.0 +2023-03-26 13:06:29,862 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59262.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:06:51,936 INFO [finetune.py:976] (0/7) Epoch 11, batch 2000, loss[loss=0.1764, simple_loss=0.2464, pruned_loss=0.05322, over 4915.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2617, pruned_loss=0.06508, over 956523.43 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:07:05,194 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-03-26 13:07:15,259 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 13:07:29,294 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7805, 1.6028, 2.1999, 3.4093, 2.3078, 2.4310, 1.0292, 2.7423], + device='cuda:0'), covar=tensor([0.1657, 0.1395, 0.1273, 0.0668, 0.0823, 0.1335, 0.1914, 0.0569], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0137, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:07:37,498 INFO [finetune.py:976] (0/7) Epoch 11, batch 2050, loss[loss=0.1668, simple_loss=0.2388, pruned_loss=0.04744, over 4802.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2578, pruned_loss=0.06375, over 957641.11 frames. ], batch size: 51, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:07:42,276 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.122e+01 1.513e+02 1.843e+02 2.174e+02 3.611e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-26 13:07:44,093 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:08:17,284 INFO [finetune.py:976] (0/7) Epoch 11, batch 2100, loss[loss=0.236, simple_loss=0.3055, pruned_loss=0.0833, over 4927.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2573, pruned_loss=0.06419, over 956328.11 frames. ], batch size: 38, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:08:22,238 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:09:08,920 INFO [finetune.py:976] (0/7) Epoch 11, batch 2150, loss[loss=0.1901, simple_loss=0.2748, pruned_loss=0.05265, over 4825.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2606, pruned_loss=0.0654, over 955123.27 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:09:11,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6908, 1.6293, 1.9708, 1.3175, 1.6955, 1.9548, 1.4741, 2.1131], + device='cuda:0'), covar=tensor([0.1528, 0.2243, 0.1726, 0.2272, 0.1092, 0.1628, 0.3007, 0.1008], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0208, 0.0195, 0.0192, 0.0179, 0.0217, 0.0220, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:09:13,353 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59431.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:09:15,687 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.596e+02 1.893e+02 2.254e+02 5.168e+02, threshold=3.786e+02, percent-clipped=3.0 +2023-03-26 13:09:26,825 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 13:09:54,869 INFO [finetune.py:976] (0/7) Epoch 11, batch 2200, loss[loss=0.2332, simple_loss=0.291, pruned_loss=0.08768, over 4825.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.2641, pruned_loss=0.06679, over 953390.21 frames. ], batch size: 45, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:09:56,162 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:08,280 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8879, 1.5494, 2.1722, 3.5824, 2.4021, 2.4189, 1.1939, 2.8554], + device='cuda:0'), covar=tensor([0.1680, 0.1518, 0.1326, 0.0559, 0.0820, 0.1402, 0.1684, 0.0538], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:10:13,064 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:25,064 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:30,586 INFO [finetune.py:976] (0/7) Epoch 11, batch 2250, loss[loss=0.2066, simple_loss=0.2718, pruned_loss=0.07072, over 4754.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2645, pruned_loss=0.06698, over 952802.60 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:10:37,560 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.729e+02 2.023e+02 2.518e+02 3.990e+02, threshold=4.047e+02, percent-clipped=2.0 +2023-03-26 13:10:39,602 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.85 vs. limit=2.0 +2023-03-26 13:11:02,684 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59562.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:03,355 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:13,470 INFO [finetune.py:976] (0/7) Epoch 11, batch 2300, loss[loss=0.1875, simple_loss=0.2478, pruned_loss=0.06363, over 4918.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2644, pruned_loss=0.0664, over 951528.21 frames. ], batch size: 42, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:11:16,562 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5213, 1.3501, 1.2702, 1.5456, 1.6209, 1.5569, 0.8972, 1.2833], + device='cuda:0'), covar=tensor([0.2190, 0.2109, 0.1994, 0.1625, 0.1627, 0.1248, 0.2673, 0.1928], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0206, 0.0208, 0.0189, 0.0241, 0.0181, 0.0212, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:11:25,617 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 13:11:35,293 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:47,097 INFO [finetune.py:976] (0/7) Epoch 11, batch 2350, loss[loss=0.1898, simple_loss=0.2539, pruned_loss=0.06284, over 4837.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2613, pruned_loss=0.06495, over 950383.82 frames. ], batch size: 44, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:11:52,461 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.309e+01 1.451e+02 1.728e+02 2.097e+02 4.600e+02, threshold=3.455e+02, percent-clipped=1.0 +2023-03-26 13:11:58,632 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-03-26 13:12:02,180 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5182, 1.4603, 1.4283, 1.4677, 1.0187, 2.8860, 1.0969, 1.6602], + device='cuda:0'), covar=tensor([0.3128, 0.2347, 0.2039, 0.2356, 0.1855, 0.0268, 0.2849, 0.1251], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0099, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:12:04,608 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5051, 1.4219, 1.4340, 1.3904, 0.9147, 2.2875, 0.8080, 1.3419], + device='cuda:0'), covar=tensor([0.3380, 0.2485, 0.2084, 0.2511, 0.1879, 0.0410, 0.2747, 0.1354], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0099, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:12:19,963 INFO [finetune.py:976] (0/7) Epoch 11, batch 2400, loss[loss=0.2248, simple_loss=0.2794, pruned_loss=0.08512, over 4913.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.258, pruned_loss=0.06375, over 951510.64 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:12:35,005 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 13:12:53,270 INFO [finetune.py:976] (0/7) Epoch 11, batch 2450, loss[loss=0.1557, simple_loss=0.2166, pruned_loss=0.04745, over 4757.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2557, pruned_loss=0.06295, over 952087.91 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:13:01,210 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.641e+02 1.877e+02 2.149e+02 5.374e+02, threshold=3.753e+02, percent-clipped=2.0 +2023-03-26 13:13:37,051 INFO [finetune.py:976] (0/7) Epoch 11, batch 2500, loss[loss=0.1575, simple_loss=0.2273, pruned_loss=0.04388, over 4760.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2569, pruned_loss=0.06341, over 953838.88 frames. ], batch size: 27, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:14:12,596 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3410, 1.1537, 1.6273, 2.4297, 1.5693, 2.0623, 0.8015, 1.9911], + device='cuda:0'), covar=tensor([0.1706, 0.1644, 0.1126, 0.0703, 0.0982, 0.1213, 0.1610, 0.0658], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0164, 0.0102, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:14:25,748 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59820.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:14:33,891 INFO [finetune.py:976] (0/7) Epoch 11, batch 2550, loss[loss=0.2448, simple_loss=0.3026, pruned_loss=0.09354, over 4878.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2611, pruned_loss=0.06449, over 954998.14 frames. ], batch size: 34, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:14:40,182 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.636e+02 1.885e+02 2.323e+02 4.849e+02, threshold=3.771e+02, percent-clipped=2.0 +2023-03-26 13:14:42,167 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8923, 2.5171, 2.4508, 1.3225, 2.5585, 2.1065, 1.9586, 2.3340], + device='cuda:0'), covar=tensor([0.1004, 0.0909, 0.1625, 0.2193, 0.1780, 0.1799, 0.2074, 0.1201], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0199, 0.0201, 0.0187, 0.0215, 0.0208, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:14:49,216 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:14:57,192 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59858.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:03,799 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59868.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:09,218 INFO [finetune.py:976] (0/7) Epoch 11, batch 2600, loss[loss=0.1365, simple_loss=0.2023, pruned_loss=0.03537, over 4788.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2627, pruned_loss=0.0651, over 955916.53 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:15:18,125 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59889.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:31,230 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:42,443 INFO [finetune.py:976] (0/7) Epoch 11, batch 2650, loss[loss=0.2037, simple_loss=0.279, pruned_loss=0.0642, over 4801.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2631, pruned_loss=0.06433, over 954035.47 frames. ], batch size: 40, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:15:47,336 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.549e+02 1.976e+02 2.444e+02 3.877e+02, threshold=3.952e+02, percent-clipped=1.0 +2023-03-26 13:15:54,324 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5214, 1.5790, 1.3631, 1.5487, 1.9284, 1.7239, 1.5807, 1.3790], + device='cuda:0'), covar=tensor([0.0328, 0.0285, 0.0595, 0.0291, 0.0181, 0.0485, 0.0304, 0.0364], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0108, 0.0139, 0.0114, 0.0101, 0.0104, 0.0092, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.0900e-05, 8.4433e-05, 1.1084e-04, 8.9037e-05, 7.9359e-05, 7.6876e-05, + 6.9983e-05, 8.2854e-05], device='cuda:0') +2023-03-26 13:16:03,053 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:16:08,360 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7156, 1.2698, 1.0521, 1.6328, 2.0015, 1.3254, 1.4911, 1.6192], + device='cuda:0'), covar=tensor([0.1295, 0.1881, 0.1812, 0.1065, 0.1869, 0.2009, 0.1257, 0.1693], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0094, 0.0122, 0.0096, 0.0101, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 13:16:29,561 INFO [finetune.py:976] (0/7) Epoch 11, batch 2700, loss[loss=0.1771, simple_loss=0.2421, pruned_loss=0.05607, over 4906.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2629, pruned_loss=0.06452, over 953368.00 frames. ], batch size: 36, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:16:30,880 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7859, 3.2531, 3.4808, 3.6802, 3.5577, 3.3593, 3.8498, 1.2878], + device='cuda:0'), covar=tensor([0.0763, 0.0955, 0.0884, 0.0897, 0.1124, 0.1554, 0.0888, 0.5037], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0242, 0.0274, 0.0291, 0.0328, 0.0284, 0.0299, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:16:45,700 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-60000.pt +2023-03-26 13:16:46,831 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9734, 2.1038, 2.0829, 1.3819, 2.1306, 2.1545, 2.1365, 1.8156], + device='cuda:0'), covar=tensor([0.0604, 0.0572, 0.0631, 0.0954, 0.0623, 0.0662, 0.0596, 0.0992], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0134, 0.0143, 0.0126, 0.0121, 0.0144, 0.0146, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:16:49,320 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5609, 1.6911, 1.7121, 0.9428, 1.7971, 1.9466, 1.8798, 1.4913], + device='cuda:0'), covar=tensor([0.1013, 0.0682, 0.0489, 0.0588, 0.0431, 0.0508, 0.0310, 0.0690], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0155, 0.0122, 0.0133, 0.0131, 0.0125, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.5315e-05, 1.1325e-04, 8.8122e-05, 9.6395e-05, 9.3460e-05, 9.1113e-05, + 1.0572e-04, 1.0761e-04], device='cuda:0') +2023-03-26 13:17:04,323 INFO [finetune.py:976] (0/7) Epoch 11, batch 2750, loss[loss=0.1546, simple_loss=0.2375, pruned_loss=0.0359, over 4891.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2604, pruned_loss=0.06362, over 954299.27 frames. ], batch size: 35, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:17:09,208 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.603e+02 1.823e+02 2.284e+02 4.397e+02, threshold=3.646e+02, percent-clipped=1.0 +2023-03-26 13:17:12,384 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:21,587 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:30,379 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:37,372 INFO [finetune.py:976] (0/7) Epoch 11, batch 2800, loss[loss=0.1456, simple_loss=0.2154, pruned_loss=0.03792, over 4779.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2573, pruned_loss=0.06292, over 952838.23 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:17:38,086 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:54,527 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60101.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 13:18:02,773 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60113.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 13:18:10,606 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:11,114 INFO [finetune.py:976] (0/7) Epoch 11, batch 2850, loss[loss=0.2017, simple_loss=0.2622, pruned_loss=0.07064, over 4692.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.256, pruned_loss=0.06323, over 950090.88 frames. ], batch size: 23, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:18:17,974 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.579e+02 1.818e+02 2.348e+02 4.165e+02, threshold=3.636e+02, percent-clipped=3.0 +2023-03-26 13:18:20,406 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6870, 1.4714, 1.9161, 1.8258, 1.5592, 3.5294, 1.3863, 1.6222], + device='cuda:0'), covar=tensor([0.0899, 0.1764, 0.1139, 0.0969, 0.1606, 0.0253, 0.1441, 0.1679], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:18:21,044 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60139.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:33,358 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5829, 1.7059, 1.3899, 1.6318, 2.0102, 1.8157, 1.5859, 1.4405], + device='cuda:0'), covar=tensor([0.0281, 0.0275, 0.0528, 0.0274, 0.0193, 0.0459, 0.0395, 0.0372], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0109, 0.0140, 0.0114, 0.0102, 0.0104, 0.0093, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.1103e-05, 8.4938e-05, 1.1142e-04, 8.9295e-05, 7.9718e-05, 7.6972e-05, + 7.0418e-05, 8.3216e-05], device='cuda:0') +2023-03-26 13:18:39,230 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7214, 1.2920, 0.7635, 1.6358, 2.0608, 1.3577, 1.6057, 1.6652], + device='cuda:0'), covar=tensor([0.1532, 0.2134, 0.2147, 0.1195, 0.2014, 0.1900, 0.1393, 0.1908], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0098, 0.0115, 0.0094, 0.0122, 0.0096, 0.0101, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 13:18:39,241 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:51,666 INFO [finetune.py:976] (0/7) Epoch 11, batch 2900, loss[loss=0.2314, simple_loss=0.2927, pruned_loss=0.08506, over 4774.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2599, pruned_loss=0.06484, over 951663.41 frames. ], batch size: 54, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:19:12,020 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60198.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:21,310 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:22,529 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:51,303 INFO [finetune.py:976] (0/7) Epoch 11, batch 2950, loss[loss=0.2111, simple_loss=0.2715, pruned_loss=0.07541, over 4852.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2623, pruned_loss=0.06549, over 951247.46 frames. ], batch size: 44, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:19:52,023 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6805, 1.4778, 1.0391, 0.2331, 1.2640, 1.4882, 1.4255, 1.3696], + device='cuda:0'), covar=tensor([0.0842, 0.0898, 0.1295, 0.2069, 0.1518, 0.2342, 0.2372, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0199, 0.0201, 0.0186, 0.0215, 0.0208, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:20:00,143 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.723e+02 2.035e+02 2.444e+02 4.360e+02, threshold=4.070e+02, percent-clipped=6.0 +2023-03-26 13:20:06,675 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:20:16,616 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:20:28,429 INFO [finetune.py:976] (0/7) Epoch 11, batch 3000, loss[loss=0.1958, simple_loss=0.26, pruned_loss=0.06582, over 4829.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2634, pruned_loss=0.06566, over 951885.48 frames. ], batch size: 47, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:20:28,430 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 13:20:38,899 INFO [finetune.py:1010] (0/7) Epoch 11, validation: loss=0.1572, simple_loss=0.2284, pruned_loss=0.04301, over 2265189.00 frames. +2023-03-26 13:20:38,899 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6332MB +2023-03-26 13:20:54,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8057, 1.6400, 2.0565, 1.9270, 1.7971, 4.1717, 1.5014, 1.7531], + device='cuda:0'), covar=tensor([0.0989, 0.2108, 0.1388, 0.1167, 0.1843, 0.0235, 0.1797, 0.2084], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:20:56,507 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9415, 2.0409, 2.0664, 1.3077, 1.9689, 2.0750, 2.1261, 1.6702], + device='cuda:0'), covar=tensor([0.0645, 0.0620, 0.0655, 0.1001, 0.0674, 0.0724, 0.0580, 0.1142], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0136, 0.0144, 0.0127, 0.0122, 0.0146, 0.0147, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:21:10,749 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1767, 1.3001, 1.1570, 1.4107, 1.3488, 2.4283, 1.2284, 1.4199], + device='cuda:0'), covar=tensor([0.0972, 0.1835, 0.1240, 0.0949, 0.1702, 0.0377, 0.1497, 0.1668], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:21:13,696 INFO [finetune.py:976] (0/7) Epoch 11, batch 3050, loss[loss=0.1542, simple_loss=0.2344, pruned_loss=0.03702, over 4786.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2642, pruned_loss=0.06526, over 953687.41 frames. ], batch size: 29, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:21:19,479 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.587e+02 1.939e+02 2.482e+02 4.597e+02, threshold=3.877e+02, percent-clipped=2.0 +2023-03-26 13:21:22,663 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7351, 1.2437, 0.9935, 1.6836, 2.1430, 1.3520, 1.6419, 1.5868], + device='cuda:0'), covar=tensor([0.1473, 0.2050, 0.1828, 0.1173, 0.1807, 0.1756, 0.1337, 0.2038], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0094, 0.0121, 0.0096, 0.0100, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:21:26,381 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 13:21:56,081 INFO [finetune.py:976] (0/7) Epoch 11, batch 3100, loss[loss=0.1652, simple_loss=0.2332, pruned_loss=0.04855, over 4770.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2613, pruned_loss=0.0641, over 954589.59 frames. ], batch size: 27, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:22:08,706 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60396.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 13:22:08,790 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8265, 1.3307, 1.9239, 1.7665, 1.5611, 1.5208, 1.7024, 1.6728], + device='cuda:0'), covar=tensor([0.4448, 0.4710, 0.3785, 0.4410, 0.5593, 0.4343, 0.5216, 0.3793], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0239, 0.0254, 0.0259, 0.0255, 0.0230, 0.0273, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:22:16,692 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60408.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 13:22:17,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8273, 1.5967, 2.1136, 3.4494, 2.4128, 2.4356, 1.1186, 2.7847], + device='cuda:0'), covar=tensor([0.1663, 0.1475, 0.1339, 0.0548, 0.0718, 0.1259, 0.1857, 0.0536], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:22:25,055 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:22:29,561 INFO [finetune.py:976] (0/7) Epoch 11, batch 3150, loss[loss=0.2111, simple_loss=0.2614, pruned_loss=0.0804, over 4805.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2593, pruned_loss=0.06403, over 954050.59 frames. ], batch size: 51, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:22:34,362 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60434.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:22:34,872 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.624e+02 1.838e+02 2.200e+02 4.980e+02, threshold=3.676e+02, percent-clipped=1.0 +2023-03-26 13:22:49,352 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7721, 3.2360, 3.4187, 3.6255, 3.5573, 3.2877, 3.8137, 1.2009], + device='cuda:0'), covar=tensor([0.0844, 0.0916, 0.0907, 0.1092, 0.1195, 0.1646, 0.0861, 0.5301], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0276, 0.0292, 0.0331, 0.0284, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:22:58,897 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6480, 1.4968, 1.1208, 0.2909, 1.3770, 1.4802, 1.4246, 1.5460], + device='cuda:0'), covar=tensor([0.0877, 0.0844, 0.1258, 0.1923, 0.1288, 0.2194, 0.2156, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0201, 0.0203, 0.0188, 0.0217, 0.0210, 0.0225, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:23:01,693 INFO [finetune.py:976] (0/7) Epoch 11, batch 3200, loss[loss=0.1603, simple_loss=0.2309, pruned_loss=0.04486, over 4901.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2562, pruned_loss=0.06315, over 955853.95 frames. ], batch size: 35, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:23:20,576 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:23:37,313 INFO [finetune.py:976] (0/7) Epoch 11, batch 3250, loss[loss=0.1992, simple_loss=0.2711, pruned_loss=0.06362, over 4819.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2556, pruned_loss=0.06277, over 955413.87 frames. ], batch size: 40, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:23:48,950 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.626e+02 1.982e+02 2.397e+02 3.737e+02, threshold=3.964e+02, percent-clipped=1.0 +2023-03-26 13:23:59,840 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:04,029 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:05,272 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60554.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:11,474 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 13:24:27,367 INFO [finetune.py:976] (0/7) Epoch 11, batch 3300, loss[loss=0.1795, simple_loss=0.2481, pruned_loss=0.05543, over 4760.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.261, pruned_loss=0.06481, over 952357.88 frames. ], batch size: 59, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:24:37,765 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.48 vs. limit=5.0 +2023-03-26 13:24:45,612 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:57,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5747, 1.7279, 1.3244, 1.6213, 2.0382, 1.7905, 1.5976, 1.4550], + device='cuda:0'), covar=tensor([0.0309, 0.0262, 0.0592, 0.0274, 0.0166, 0.0506, 0.0290, 0.0402], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0108, 0.0140, 0.0114, 0.0102, 0.0103, 0.0093, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.0772e-05, 8.4379e-05, 1.1111e-04, 8.8976e-05, 7.9376e-05, 7.6722e-05, + 7.0209e-05, 8.2633e-05], device='cuda:0') +2023-03-26 13:25:28,936 INFO [finetune.py:976] (0/7) Epoch 11, batch 3350, loss[loss=0.1747, simple_loss=0.2453, pruned_loss=0.05203, over 4813.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2635, pruned_loss=0.06571, over 952923.53 frames. ], batch size: 40, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:25:34,910 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.701e+02 2.036e+02 2.469e+02 3.577e+02, threshold=4.071e+02, percent-clipped=0.0 +2023-03-26 13:25:45,023 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6522, 1.1792, 0.8611, 1.5740, 1.9782, 1.3831, 1.2946, 1.5911], + device='cuda:0'), covar=tensor([0.1499, 0.2231, 0.2070, 0.1260, 0.1985, 0.2070, 0.1564, 0.1892], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0120, 0.0095, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:25:53,435 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5040, 1.4405, 1.5518, 0.7865, 1.5535, 1.5156, 1.4169, 1.3177], + device='cuda:0'), covar=tensor([0.0587, 0.0793, 0.0697, 0.0993, 0.0794, 0.0750, 0.0690, 0.1284], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0141, 0.0124, 0.0120, 0.0143, 0.0143, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:26:02,929 INFO [finetune.py:976] (0/7) Epoch 11, batch 3400, loss[loss=0.1967, simple_loss=0.2627, pruned_loss=0.06534, over 4888.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2649, pruned_loss=0.0665, over 952329.62 frames. ], batch size: 32, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:26:13,120 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 13:26:16,535 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:24,734 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:30,775 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:32,557 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60721.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:36,601 INFO [finetune.py:976] (0/7) Epoch 11, batch 3450, loss[loss=0.1797, simple_loss=0.2426, pruned_loss=0.05838, over 4816.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2637, pruned_loss=0.06579, over 953419.23 frames. ], batch size: 40, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:26:40,473 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 13:26:41,026 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60734.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:41,510 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.902e+01 1.594e+02 1.892e+02 2.253e+02 3.493e+02, threshold=3.783e+02, percent-clipped=0.0 +2023-03-26 13:26:52,715 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:06,421 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:25,456 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60769.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:36,412 INFO [finetune.py:976] (0/7) Epoch 11, batch 3500, loss[loss=0.1757, simple_loss=0.2407, pruned_loss=0.05538, over 4890.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2608, pruned_loss=0.0649, over 953552.57 frames. ], batch size: 32, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:27:37,732 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60779.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:39,436 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60782.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:28:15,227 INFO [finetune.py:976] (0/7) Epoch 11, batch 3550, loss[loss=0.201, simple_loss=0.2714, pruned_loss=0.06535, over 4686.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2591, pruned_loss=0.06476, over 953891.59 frames. ], batch size: 23, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:28:20,663 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.566e+02 1.863e+02 2.348e+02 4.575e+02, threshold=3.726e+02, percent-clipped=4.0 +2023-03-26 13:28:34,183 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60854.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:28:34,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1618, 1.9378, 1.4808, 0.5862, 1.6417, 1.7601, 1.5937, 1.8174], + device='cuda:0'), covar=tensor([0.0772, 0.0798, 0.1358, 0.2009, 0.1313, 0.2135, 0.2163, 0.0823], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0200, 0.0202, 0.0186, 0.0215, 0.0209, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:28:42,607 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7305, 0.7090, 1.7160, 1.5886, 1.5055, 1.4469, 1.5040, 1.6044], + device='cuda:0'), covar=tensor([0.3827, 0.4197, 0.3527, 0.3759, 0.4539, 0.3746, 0.4548, 0.3306], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0239, 0.0254, 0.0259, 0.0255, 0.0231, 0.0274, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:28:49,076 INFO [finetune.py:976] (0/7) Epoch 11, batch 3600, loss[loss=0.2039, simple_loss=0.2668, pruned_loss=0.07043, over 4811.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2551, pruned_loss=0.06278, over 951558.44 frames. ], batch size: 39, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:29:17,735 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:29:26,610 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7658, 3.2751, 3.4280, 3.6287, 3.5117, 3.2751, 3.8257, 1.2047], + device='cuda:0'), covar=tensor([0.0925, 0.0978, 0.0979, 0.1136, 0.1476, 0.1689, 0.0927, 0.5530], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0241, 0.0274, 0.0289, 0.0330, 0.0281, 0.0299, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:29:31,320 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4470, 1.4270, 1.7883, 1.7808, 1.5071, 3.3601, 1.3742, 1.5026], + device='cuda:0'), covar=tensor([0.0965, 0.1733, 0.1192, 0.0937, 0.1636, 0.0253, 0.1470, 0.1774], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:29:39,498 INFO [finetune.py:976] (0/7) Epoch 11, batch 3650, loss[loss=0.2527, simple_loss=0.3003, pruned_loss=0.1025, over 4744.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2562, pruned_loss=0.06308, over 949947.89 frames. ], batch size: 54, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:29:44,363 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.638e+02 1.962e+02 2.312e+02 3.604e+02, threshold=3.924e+02, percent-clipped=0.0 +2023-03-26 13:29:52,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2394, 2.0252, 1.4472, 0.5301, 1.6858, 1.9375, 1.7821, 1.7858], + device='cuda:0'), covar=tensor([0.0750, 0.0696, 0.1324, 0.1872, 0.1237, 0.1915, 0.1797, 0.0787], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0200, 0.0203, 0.0187, 0.0216, 0.0210, 0.0223, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:30:33,792 INFO [finetune.py:976] (0/7) Epoch 11, batch 3700, loss[loss=0.2259, simple_loss=0.2928, pruned_loss=0.07948, over 4733.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2598, pruned_loss=0.06365, over 951060.67 frames. ], batch size: 59, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:30:59,266 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2047, 1.3162, 1.3558, 0.7292, 1.2384, 1.5395, 1.5829, 1.2353], + device='cuda:0'), covar=tensor([0.0895, 0.0620, 0.0503, 0.0490, 0.0464, 0.0638, 0.0345, 0.0738], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0154, 0.0122, 0.0133, 0.0131, 0.0125, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.4425e-05, 1.1281e-04, 8.7786e-05, 9.6259e-05, 9.3076e-05, 9.0963e-05, + 1.0510e-04, 1.0751e-04], device='cuda:0') +2023-03-26 13:31:10,015 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 13:31:15,830 INFO [finetune.py:976] (0/7) Epoch 11, batch 3750, loss[loss=0.2553, simple_loss=0.3147, pruned_loss=0.0979, over 4829.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2616, pruned_loss=0.06414, over 951921.13 frames. ], batch size: 47, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:31:20,655 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.587e+02 1.819e+02 2.276e+02 4.586e+02, threshold=3.638e+02, percent-clipped=1.0 +2023-03-26 13:31:38,435 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 13:31:38,940 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61061.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:31:47,191 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:31:49,400 INFO [finetune.py:976] (0/7) Epoch 11, batch 3800, loss[loss=0.1721, simple_loss=0.2359, pruned_loss=0.05418, over 4741.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2629, pruned_loss=0.06456, over 952408.54 frames. ], batch size: 59, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:32:29,721 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:32:32,630 INFO [finetune.py:976] (0/7) Epoch 11, batch 3850, loss[loss=0.1853, simple_loss=0.2416, pruned_loss=0.06448, over 4901.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2614, pruned_loss=0.06403, over 952489.14 frames. ], batch size: 35, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:32:37,924 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.518e+02 1.864e+02 2.279e+02 4.215e+02, threshold=3.727e+02, percent-clipped=1.0 +2023-03-26 13:33:02,470 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0312, 1.7771, 2.7038, 1.4346, 2.1033, 2.3507, 1.6498, 2.4783], + device='cuda:0'), covar=tensor([0.1400, 0.1959, 0.1084, 0.2173, 0.0969, 0.1479, 0.2628, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0205, 0.0194, 0.0192, 0.0179, 0.0216, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:33:05,949 INFO [finetune.py:976] (0/7) Epoch 11, batch 3900, loss[loss=0.1761, simple_loss=0.244, pruned_loss=0.05414, over 4927.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2589, pruned_loss=0.06373, over 953349.90 frames. ], batch size: 38, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:33:39,748 INFO [finetune.py:976] (0/7) Epoch 11, batch 3950, loss[loss=0.1459, simple_loss=0.2191, pruned_loss=0.03635, over 4806.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2556, pruned_loss=0.06233, over 955136.69 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:33:45,061 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.570e+02 1.907e+02 2.309e+02 4.377e+02, threshold=3.813e+02, percent-clipped=3.0 +2023-03-26 13:33:48,952 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 13:33:54,391 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.64 vs. limit=5.0 +2023-03-26 13:34:02,623 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9199, 1.8045, 1.6342, 2.1000, 2.2923, 2.1488, 1.6340, 1.5527], + device='cuda:0'), covar=tensor([0.2178, 0.1988, 0.1981, 0.1557, 0.1744, 0.1090, 0.2414, 0.1903], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0206, 0.0208, 0.0188, 0.0240, 0.0181, 0.0211, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:34:12,004 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 13:34:12,380 INFO [finetune.py:976] (0/7) Epoch 11, batch 4000, loss[loss=0.2124, simple_loss=0.2844, pruned_loss=0.07021, over 4823.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2561, pruned_loss=0.06277, over 956899.57 frames. ], batch size: 40, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:34:55,752 INFO [finetune.py:976] (0/7) Epoch 11, batch 4050, loss[loss=0.1983, simple_loss=0.2793, pruned_loss=0.0587, over 4891.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2603, pruned_loss=0.06451, over 955981.87 frames. ], batch size: 43, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:35:04,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.652e+02 2.086e+02 2.571e+02 4.987e+02, threshold=4.171e+02, percent-clipped=6.0 +2023-03-26 13:35:10,525 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 13:35:23,171 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4422, 2.9512, 2.7479, 1.2150, 2.9629, 2.1879, 0.6411, 1.7903], + device='cuda:0'), covar=tensor([0.2166, 0.2258, 0.2016, 0.3662, 0.1456, 0.1227, 0.4286, 0.1818], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0175, 0.0160, 0.0129, 0.0156, 0.0122, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 13:35:33,869 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0743, 1.2702, 1.1127, 1.3935, 1.3577, 2.4245, 1.1789, 1.3832], + device='cuda:0'), covar=tensor([0.0976, 0.1789, 0.1200, 0.0912, 0.1678, 0.0388, 0.1530, 0.1661], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:35:36,715 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0323, 1.2537, 1.1786, 1.3445, 1.3958, 2.3959, 1.1909, 1.3862], + device='cuda:0'), covar=tensor([0.1065, 0.1957, 0.1180, 0.0952, 0.1643, 0.0394, 0.1569, 0.1792], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:35:36,763 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7415, 1.5787, 1.3702, 1.1931, 1.5611, 1.5168, 1.5176, 2.0681], + device='cuda:0'), covar=tensor([0.4522, 0.4502, 0.3580, 0.4212, 0.4005, 0.2476, 0.3883, 0.1964], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0260, 0.0223, 0.0278, 0.0243, 0.0210, 0.0245, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:35:41,810 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:35:43,536 INFO [finetune.py:976] (0/7) Epoch 11, batch 4100, loss[loss=0.1943, simple_loss=0.2706, pruned_loss=0.059, over 4768.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.263, pruned_loss=0.06525, over 955221.02 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:35:47,824 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8882, 1.3224, 1.6920, 1.7013, 1.5073, 1.5385, 1.6671, 1.6357], + device='cuda:0'), covar=tensor([0.5121, 0.5278, 0.4900, 0.5196, 0.6808, 0.4971, 0.6494, 0.4863], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0239, 0.0254, 0.0259, 0.0256, 0.0231, 0.0275, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:36:11,792 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61417.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:36:20,055 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:36:26,637 INFO [finetune.py:976] (0/7) Epoch 11, batch 4150, loss[loss=0.1666, simple_loss=0.2315, pruned_loss=0.05083, over 4785.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2637, pruned_loss=0.06534, over 954472.80 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:36:32,502 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.629e+02 1.982e+02 2.519e+02 5.426e+02, threshold=3.964e+02, percent-clipped=4.0 +2023-03-26 13:36:59,826 INFO [finetune.py:976] (0/7) Epoch 11, batch 4200, loss[loss=0.1624, simple_loss=0.2235, pruned_loss=0.05069, over 4721.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2641, pruned_loss=0.06582, over 954628.38 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:37:03,885 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9952, 1.6888, 2.3704, 1.4886, 2.0869, 2.2796, 1.6647, 2.4090], + device='cuda:0'), covar=tensor([0.1315, 0.2091, 0.1248, 0.2065, 0.0921, 0.1419, 0.2759, 0.0889], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0206, 0.0194, 0.0192, 0.0179, 0.0217, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:37:35,243 INFO [finetune.py:976] (0/7) Epoch 11, batch 4250, loss[loss=0.1855, simple_loss=0.2497, pruned_loss=0.06063, over 4908.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.261, pruned_loss=0.06478, over 956922.99 frames. ], batch size: 43, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:37:44,259 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6089, 1.6836, 1.6531, 0.9794, 1.6904, 1.8251, 1.9389, 1.4360], + device='cuda:0'), covar=tensor([0.0830, 0.0550, 0.0483, 0.0503, 0.0445, 0.0689, 0.0262, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0154, 0.0122, 0.0133, 0.0131, 0.0126, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.4627e-05, 1.1312e-04, 8.8174e-05, 9.5866e-05, 9.3420e-05, 9.1856e-05, + 1.0528e-04, 1.0734e-04], device='cuda:0') +2023-03-26 13:37:45,943 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.771e+01 1.547e+02 1.858e+02 2.245e+02 5.805e+02, threshold=3.715e+02, percent-clipped=2.0 +2023-03-26 13:37:49,817 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 13:38:11,413 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6188, 2.2094, 3.0004, 1.8531, 2.6938, 3.0484, 2.1493, 3.0355], + device='cuda:0'), covar=tensor([0.1567, 0.2030, 0.1358, 0.2409, 0.0890, 0.1647, 0.2696, 0.0872], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0205, 0.0193, 0.0191, 0.0178, 0.0216, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:38:15,492 INFO [finetune.py:976] (0/7) Epoch 11, batch 4300, loss[loss=0.1664, simple_loss=0.2356, pruned_loss=0.04858, over 4832.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2586, pruned_loss=0.06411, over 956292.65 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:38:48,429 INFO [finetune.py:976] (0/7) Epoch 11, batch 4350, loss[loss=0.1751, simple_loss=0.2303, pruned_loss=0.05994, over 4759.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2547, pruned_loss=0.06275, over 957963.91 frames. ], batch size: 54, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:38:54,827 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.580e+02 1.801e+02 2.212e+02 3.446e+02, threshold=3.603e+02, percent-clipped=0.0 +2023-03-26 13:39:04,080 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 13:39:21,856 INFO [finetune.py:976] (0/7) Epoch 11, batch 4400, loss[loss=0.2562, simple_loss=0.3162, pruned_loss=0.0981, over 4912.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2576, pruned_loss=0.06417, over 958187.71 frames. ], batch size: 36, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:39:31,348 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6442, 1.5461, 2.0674, 1.3146, 1.8498, 2.0137, 1.4281, 2.0660], + device='cuda:0'), covar=tensor([0.1452, 0.2065, 0.1320, 0.2040, 0.0894, 0.1384, 0.2706, 0.0959], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0203, 0.0191, 0.0189, 0.0177, 0.0213, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:39:53,786 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61717.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:04,331 INFO [finetune.py:976] (0/7) Epoch 11, batch 4450, loss[loss=0.2513, simple_loss=0.3131, pruned_loss=0.0948, over 4815.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2612, pruned_loss=0.06554, over 955348.53 frames. ], batch size: 51, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:40:07,480 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:14,298 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.628e+02 1.977e+02 2.534e+02 3.640e+02, threshold=3.954e+02, percent-clipped=2.0 +2023-03-26 13:40:23,019 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:23,054 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0236, 1.2397, 2.0076, 1.8884, 1.7610, 1.7088, 1.7619, 1.8628], + device='cuda:0'), covar=tensor([0.3811, 0.4495, 0.3747, 0.3851, 0.5221, 0.3839, 0.5221, 0.3616], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0239, 0.0254, 0.0259, 0.0256, 0.0231, 0.0275, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:40:45,895 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:57,017 INFO [finetune.py:976] (0/7) Epoch 11, batch 4500, loss[loss=0.1703, simple_loss=0.2465, pruned_loss=0.04708, over 4755.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2625, pruned_loss=0.06589, over 956254.33 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:41:07,864 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:41:16,079 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:41:33,078 INFO [finetune.py:976] (0/7) Epoch 11, batch 4550, loss[loss=0.1882, simple_loss=0.2631, pruned_loss=0.05669, over 4889.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.263, pruned_loss=0.06587, over 955132.88 frames. ], batch size: 35, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:41:43,495 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.535e+01 1.607e+02 1.951e+02 2.245e+02 3.846e+02, threshold=3.902e+02, percent-clipped=0.0 +2023-03-26 13:41:58,773 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1373, 2.0801, 1.7093, 1.8477, 2.0817, 1.8232, 2.3657, 2.1896], + device='cuda:0'), covar=tensor([0.1325, 0.1982, 0.2977, 0.2758, 0.2478, 0.1555, 0.3010, 0.1703], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0186, 0.0233, 0.0253, 0.0239, 0.0197, 0.0212, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:42:05,250 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 13:42:15,220 INFO [finetune.py:976] (0/7) Epoch 11, batch 4600, loss[loss=0.1725, simple_loss=0.2412, pruned_loss=0.05188, over 4888.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2621, pruned_loss=0.065, over 955246.32 frames. ], batch size: 32, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:42:37,999 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1920, 1.3363, 1.2321, 1.4378, 1.4178, 2.2244, 1.1915, 1.4237], + device='cuda:0'), covar=tensor([0.0787, 0.1419, 0.1212, 0.0773, 0.1319, 0.0422, 0.1256, 0.1300], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 13:42:40,458 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61914.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:42:44,746 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-26 13:42:48,594 INFO [finetune.py:976] (0/7) Epoch 11, batch 4650, loss[loss=0.1784, simple_loss=0.2372, pruned_loss=0.05979, over 4901.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2608, pruned_loss=0.06515, over 955790.74 frames. ], batch size: 32, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:42:56,048 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.606e+02 1.934e+02 2.317e+02 5.626e+02, threshold=3.867e+02, percent-clipped=3.0 +2023-03-26 13:43:31,733 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61975.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:43:32,823 INFO [finetune.py:976] (0/7) Epoch 11, batch 4700, loss[loss=0.2064, simple_loss=0.2715, pruned_loss=0.07066, over 4899.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2569, pruned_loss=0.06344, over 956271.85 frames. ], batch size: 36, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:43:44,869 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1238, 2.0336, 1.8900, 2.2200, 2.6581, 2.2816, 2.1234, 1.5992], + device='cuda:0'), covar=tensor([0.2101, 0.1933, 0.1734, 0.1558, 0.1692, 0.1026, 0.1941, 0.1817], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0207, 0.0208, 0.0189, 0.0242, 0.0181, 0.0212, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:43:48,497 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-62000.pt +2023-03-26 13:44:19,772 INFO [finetune.py:976] (0/7) Epoch 11, batch 4750, loss[loss=0.1918, simple_loss=0.2565, pruned_loss=0.06354, over 4823.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2552, pruned_loss=0.06287, over 958144.29 frames. ], batch size: 45, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:44:25,601 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.474e+02 1.769e+02 2.148e+02 4.944e+02, threshold=3.539e+02, percent-clipped=1.0 +2023-03-26 13:44:53,404 INFO [finetune.py:976] (0/7) Epoch 11, batch 4800, loss[loss=0.1805, simple_loss=0.2614, pruned_loss=0.04982, over 4744.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2586, pruned_loss=0.06376, over 958273.78 frames. ], batch size: 59, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:45:06,482 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:13,198 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62099.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:40,250 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 13:45:41,878 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:50,589 INFO [finetune.py:976] (0/7) Epoch 11, batch 4850, loss[loss=0.23, simple_loss=0.2957, pruned_loss=0.08214, over 4840.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2624, pruned_loss=0.06493, over 957009.70 frames. ], batch size: 47, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:46:01,539 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.730e+02 2.037e+02 2.587e+02 8.043e+02, threshold=4.075e+02, percent-clipped=4.0 +2023-03-26 13:46:34,366 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 13:46:45,228 INFO [finetune.py:976] (0/7) Epoch 11, batch 4900, loss[loss=0.2128, simple_loss=0.2849, pruned_loss=0.07029, over 4887.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2628, pruned_loss=0.06518, over 956986.98 frames. ], batch size: 32, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:46:46,630 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 13:46:48,907 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62183.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:47:38,674 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 13:47:49,179 INFO [finetune.py:976] (0/7) Epoch 11, batch 4950, loss[loss=0.1489, simple_loss=0.2074, pruned_loss=0.04519, over 3754.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2637, pruned_loss=0.06535, over 954860.25 frames. ], batch size: 16, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:47:56,653 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.728e+02 2.029e+02 2.471e+02 5.736e+02, threshold=4.057e+02, percent-clipped=2.0 +2023-03-26 13:47:58,483 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8124, 1.2908, 1.8482, 1.7237, 1.5825, 1.5392, 1.6700, 1.6724], + device='cuda:0'), covar=tensor([0.4280, 0.4561, 0.3671, 0.4314, 0.4895, 0.4162, 0.4892, 0.3741], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0241, 0.0256, 0.0261, 0.0258, 0.0232, 0.0277, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:48:18,908 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:48:21,277 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:48:24,035 INFO [finetune.py:976] (0/7) Epoch 11, batch 5000, loss[loss=0.1972, simple_loss=0.2619, pruned_loss=0.06627, over 4832.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2601, pruned_loss=0.0636, over 954287.64 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:48:57,126 INFO [finetune.py:976] (0/7) Epoch 11, batch 5050, loss[loss=0.1715, simple_loss=0.238, pruned_loss=0.05253, over 4928.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.258, pruned_loss=0.06344, over 954441.50 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 16.0 +2023-03-26 13:49:02,470 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62334.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:04,171 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.504e+02 1.759e+02 2.068e+02 4.473e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-26 13:49:30,523 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 13:49:31,100 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1541, 2.0772, 1.7390, 2.2308, 2.0076, 2.0066, 1.9335, 2.9868], + device='cuda:0'), covar=tensor([0.4482, 0.5816, 0.3741, 0.5213, 0.5204, 0.2563, 0.5327, 0.1781], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0223, 0.0277, 0.0243, 0.0210, 0.0247, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:49:32,190 INFO [finetune.py:976] (0/7) Epoch 11, batch 5100, loss[loss=0.1817, simple_loss=0.2472, pruned_loss=0.0581, over 4726.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.254, pruned_loss=0.06159, over 954383.31 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 16.0 +2023-03-26 13:49:39,551 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 13:49:40,035 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:42,300 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62391.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:47,648 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:52,553 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9475, 1.5682, 2.3535, 3.5572, 2.4636, 2.5706, 1.0054, 2.7787], + device='cuda:0'), covar=tensor([0.1621, 0.1445, 0.1248, 0.0516, 0.0765, 0.1918, 0.1828, 0.0495], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0114, 0.0133, 0.0162, 0.0099, 0.0136, 0.0124, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:50:00,599 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9509, 1.7918, 1.6681, 1.5197, 1.9822, 1.7072, 1.8443, 1.9103], + device='cuda:0'), covar=tensor([0.1422, 0.2037, 0.3056, 0.2434, 0.2496, 0.1712, 0.2560, 0.1894], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0185, 0.0230, 0.0251, 0.0238, 0.0195, 0.0209, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:50:03,007 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.7861, 4.1961, 4.3695, 4.5484, 4.5085, 4.2708, 4.8996, 1.5790], + device='cuda:0'), covar=tensor([0.0847, 0.0858, 0.0794, 0.1105, 0.1378, 0.1511, 0.0606, 0.5585], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0243, 0.0277, 0.0290, 0.0331, 0.0284, 0.0302, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:50:03,057 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4241, 1.7979, 1.5280, 1.5421, 2.1414, 1.9380, 1.9302, 1.8182], + device='cuda:0'), covar=tensor([0.0521, 0.0353, 0.0534, 0.0347, 0.0287, 0.0603, 0.0367, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0109, 0.0141, 0.0114, 0.0102, 0.0104, 0.0092, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.1667e-05, 8.4805e-05, 1.1192e-04, 8.9117e-05, 7.9433e-05, 7.7231e-05, + 6.9598e-05, 8.3118e-05], device='cuda:0') +2023-03-26 13:50:05,688 INFO [finetune.py:976] (0/7) Epoch 11, batch 5150, loss[loss=0.2268, simple_loss=0.2955, pruned_loss=0.07909, over 4905.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2547, pruned_loss=0.06252, over 952539.72 frames. ], batch size: 35, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:50:12,136 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:12,674 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.578e+02 2.001e+02 2.432e+02 3.455e+02, threshold=4.003e+02, percent-clipped=0.0 +2023-03-26 13:50:26,762 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:30,455 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62452.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:53,081 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 13:50:55,269 INFO [finetune.py:976] (0/7) Epoch 11, batch 5200, loss[loss=0.1728, simple_loss=0.2365, pruned_loss=0.05451, over 4768.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2587, pruned_loss=0.06366, over 951529.44 frames. ], batch size: 28, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:50:56,438 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:51:36,855 INFO [finetune.py:976] (0/7) Epoch 11, batch 5250, loss[loss=0.1848, simple_loss=0.2407, pruned_loss=0.06446, over 4216.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2595, pruned_loss=0.06348, over 952036.79 frames. ], batch size: 18, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:51:54,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.618e+02 1.949e+02 2.406e+02 7.235e+02, threshold=3.897e+02, percent-clipped=3.0 +2023-03-26 13:52:03,520 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:19,512 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62570.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:23,690 INFO [finetune.py:976] (0/7) Epoch 11, batch 5300, loss[loss=0.2305, simple_loss=0.2962, pruned_loss=0.08242, over 4751.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2634, pruned_loss=0.06565, over 952971.00 frames. ], batch size: 54, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:52:29,589 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:44,287 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:52,192 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62618.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:57,602 INFO [finetune.py:976] (0/7) Epoch 11, batch 5350, loss[loss=0.1736, simple_loss=0.2418, pruned_loss=0.0527, over 4888.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2622, pruned_loss=0.06436, over 953093.17 frames. ], batch size: 32, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:52:58,901 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:04,200 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.504e+02 1.845e+02 2.238e+02 3.589e+02, threshold=3.690e+02, percent-clipped=0.0 +2023-03-26 13:53:10,792 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:24,270 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:30,762 INFO [finetune.py:976] (0/7) Epoch 11, batch 5400, loss[loss=0.212, simple_loss=0.2764, pruned_loss=0.0738, over 4811.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2589, pruned_loss=0.06293, over 950890.43 frames. ], batch size: 41, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:54:04,661 INFO [finetune.py:976] (0/7) Epoch 11, batch 5450, loss[loss=0.1944, simple_loss=0.2571, pruned_loss=0.06583, over 4859.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2563, pruned_loss=0.06221, over 952223.93 frames. ], batch size: 49, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:54:04,776 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:10,763 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.513e+01 1.463e+02 1.876e+02 2.335e+02 4.427e+02, threshold=3.751e+02, percent-clipped=2.0 +2023-03-26 13:54:17,814 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:36,205 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:38,553 INFO [finetune.py:976] (0/7) Epoch 11, batch 5500, loss[loss=0.1775, simple_loss=0.241, pruned_loss=0.05705, over 4731.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2533, pruned_loss=0.06129, over 950209.24 frames. ], batch size: 54, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:54:39,245 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:55:06,643 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9755, 1.6919, 2.3978, 3.7740, 2.6616, 2.6112, 0.9527, 3.0481], + device='cuda:0'), covar=tensor([0.1821, 0.1591, 0.1420, 0.0536, 0.0828, 0.1855, 0.1949, 0.0557], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0163, 0.0100, 0.0137, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:55:12,312 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:55:12,912 INFO [finetune.py:976] (0/7) Epoch 11, batch 5550, loss[loss=0.1394, simple_loss=0.2194, pruned_loss=0.02969, over 4796.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2548, pruned_loss=0.06195, over 951891.51 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:55:13,139 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-03-26 13:55:16,488 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 13:55:17,857 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:55:19,877 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.580e+02 1.841e+02 2.336e+02 5.980e+02, threshold=3.683e+02, percent-clipped=6.0 +2023-03-26 13:55:40,002 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-03-26 13:56:03,684 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7429, 1.5076, 2.0124, 3.4134, 2.3770, 2.2975, 0.9197, 2.6944], + device='cuda:0'), covar=tensor([0.1679, 0.1495, 0.1436, 0.0539, 0.0740, 0.1459, 0.1938, 0.0499], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0163, 0.0100, 0.0138, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 13:56:07,689 INFO [finetune.py:976] (0/7) Epoch 11, batch 5600, loss[loss=0.1615, simple_loss=0.2415, pruned_loss=0.04075, over 4825.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2594, pruned_loss=0.06323, over 953607.64 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:56:13,065 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0786, 1.7370, 2.0747, 1.9732, 1.7074, 1.7467, 1.9120, 1.8780], + device='cuda:0'), covar=tensor([0.4488, 0.4762, 0.3587, 0.4714, 0.5904, 0.4302, 0.6264, 0.3748], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0238, 0.0254, 0.0259, 0.0255, 0.0230, 0.0274, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:56:22,195 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62901.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:56:37,252 INFO [finetune.py:976] (0/7) Epoch 11, batch 5650, loss[loss=0.1664, simple_loss=0.2201, pruned_loss=0.05633, over 4203.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2625, pruned_loss=0.06405, over 954210.36 frames. ], batch size: 18, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:56:38,485 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:56:48,737 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.326e+01 1.606e+02 1.910e+02 2.279e+02 4.497e+02, threshold=3.820e+02, percent-clipped=2.0 +2023-03-26 13:56:51,135 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:57:23,402 INFO [finetune.py:976] (0/7) Epoch 11, batch 5700, loss[loss=0.18, simple_loss=0.2365, pruned_loss=0.06172, over 4254.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2585, pruned_loss=0.06316, over 936821.65 frames. ], batch size: 18, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:57:23,432 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:57:40,551 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-11.pt +2023-03-26 13:57:54,982 INFO [finetune.py:976] (0/7) Epoch 12, batch 0, loss[loss=0.2168, simple_loss=0.2839, pruned_loss=0.07492, over 4890.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.2839, pruned_loss=0.07492, over 4890.00 frames. ], batch size: 37, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:57:54,984 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 13:58:03,813 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0850, 1.7668, 1.7035, 1.7244, 1.7666, 1.6829, 1.7163, 2.4139], + device='cuda:0'), covar=tensor([0.4058, 0.5482, 0.3623, 0.4082, 0.4354, 0.2687, 0.4540, 0.1830], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0260, 0.0222, 0.0275, 0.0242, 0.0209, 0.0245, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:58:11,588 INFO [finetune.py:1010] (0/7) Epoch 12, validation: loss=0.16, simple_loss=0.2305, pruned_loss=0.04472, over 2265189.00 frames. +2023-03-26 13:58:11,589 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 13:58:19,010 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0632, 2.1069, 2.0565, 1.5279, 2.1135, 2.2118, 2.2113, 1.8108], + device='cuda:0'), covar=tensor([0.0659, 0.0627, 0.0873, 0.0952, 0.0739, 0.0730, 0.0640, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0125, 0.0121, 0.0143, 0.0143, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:58:22,059 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63022.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:58:37,034 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.590e+02 1.966e+02 2.351e+02 4.424e+02, threshold=3.931e+02, percent-clipped=2.0 +2023-03-26 13:58:37,198 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9197, 1.7230, 1.5305, 1.5556, 1.6869, 1.6347, 1.6969, 2.3675], + device='cuda:0'), covar=tensor([0.4349, 0.4757, 0.3488, 0.4486, 0.4262, 0.2516, 0.4055, 0.1786], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0222, 0.0275, 0.0242, 0.0208, 0.0245, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:58:48,198 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 13:58:49,949 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63047.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:59:00,522 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 13:59:00,859 INFO [finetune.py:976] (0/7) Epoch 12, batch 50, loss[loss=0.2569, simple_loss=0.3037, pruned_loss=0.1051, over 4901.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2688, pruned_loss=0.06909, over 217242.48 frames. ], batch size: 36, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:59:10,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0593, 1.8355, 2.0810, 1.3745, 2.1105, 2.1208, 2.0915, 1.3807], + device='cuda:0'), covar=tensor([0.0731, 0.0852, 0.0764, 0.1039, 0.0683, 0.0806, 0.0731, 0.1785], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0134, 0.0141, 0.0125, 0.0121, 0.0144, 0.0143, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:59:42,087 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.1922, 4.3883, 4.6755, 5.0026, 4.8945, 4.5878, 5.2741, 1.5510], + device='cuda:0'), covar=tensor([0.0601, 0.0772, 0.0706, 0.0769, 0.0973, 0.1339, 0.0451, 0.5423], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0243, 0.0275, 0.0289, 0.0328, 0.0282, 0.0300, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 13:59:42,645 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63095.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:59:54,683 INFO [finetune.py:976] (0/7) Epoch 12, batch 100, loss[loss=0.1463, simple_loss=0.2196, pruned_loss=0.03651, over 4824.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2604, pruned_loss=0.06469, over 381631.95 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:00:15,388 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:00:21,134 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.723e+02 1.978e+02 2.544e+02 5.107e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-03-26 14:00:21,331 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-03-26 14:00:50,143 INFO [finetune.py:976] (0/7) Epoch 12, batch 150, loss[loss=0.1402, simple_loss=0.2083, pruned_loss=0.03608, over 4757.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2568, pruned_loss=0.06459, over 510533.82 frames. ], batch size: 28, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:00:53,218 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1186, 1.6819, 2.1346, 1.9980, 1.7637, 1.8015, 1.9538, 1.9184], + device='cuda:0'), covar=tensor([0.4684, 0.5152, 0.3947, 0.4890, 0.6137, 0.4361, 0.5934, 0.3916], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0238, 0.0255, 0.0260, 0.0256, 0.0232, 0.0275, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:01:47,576 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63201.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:01:56,082 INFO [finetune.py:976] (0/7) Epoch 12, batch 200, loss[loss=0.1774, simple_loss=0.2432, pruned_loss=0.05575, over 4746.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2541, pruned_loss=0.06323, over 609647.82 frames. ], batch size: 26, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:02:17,471 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63221.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:32,829 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.551e+02 1.870e+02 2.223e+02 3.918e+02, threshold=3.740e+02, percent-clipped=0.0 +2023-03-26 14:02:41,470 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:46,790 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63249.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:51,346 INFO [finetune.py:976] (0/7) Epoch 12, batch 250, loss[loss=0.1563, simple_loss=0.2238, pruned_loss=0.0444, over 4811.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2573, pruned_loss=0.0639, over 686427.95 frames. ], batch size: 25, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:03:08,671 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:09,478 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 14:03:13,376 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:23,972 INFO [finetune.py:976] (0/7) Epoch 12, batch 300, loss[loss=0.1978, simple_loss=0.2737, pruned_loss=0.06094, over 4932.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2612, pruned_loss=0.06406, over 748456.58 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:03:28,776 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0134, 1.7849, 1.7385, 2.0190, 2.5913, 2.1215, 1.9073, 1.5424], + device='cuda:0'), covar=tensor([0.2425, 0.2375, 0.2123, 0.1870, 0.1958, 0.1240, 0.2377, 0.2018], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0210, 0.0189, 0.0243, 0.0183, 0.0213, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:03:40,223 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:51,184 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.663e+02 2.076e+02 2.406e+02 5.777e+02, threshold=4.151e+02, percent-clipped=4.0 +2023-03-26 14:04:08,571 INFO [finetune.py:976] (0/7) Epoch 12, batch 350, loss[loss=0.1688, simple_loss=0.2427, pruned_loss=0.04746, over 4753.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.262, pruned_loss=0.06419, over 795583.35 frames. ], batch size: 27, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:04:27,601 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63370.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:04:31,247 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2894, 1.7853, 2.2324, 2.1622, 1.9268, 1.8946, 2.0719, 2.0433], + device='cuda:0'), covar=tensor([0.3981, 0.4966, 0.3864, 0.4484, 0.5978, 0.4576, 0.5904, 0.3851], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0239, 0.0256, 0.0261, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:04:59,619 INFO [finetune.py:976] (0/7) Epoch 12, batch 400, loss[loss=0.205, simple_loss=0.2646, pruned_loss=0.07271, over 4816.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2619, pruned_loss=0.06405, over 830929.10 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:05:02,022 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4717, 1.3045, 1.7083, 1.7729, 1.5257, 3.3448, 1.2076, 1.4949], + device='cuda:0'), covar=tensor([0.1017, 0.1867, 0.1196, 0.0988, 0.1584, 0.0262, 0.1596, 0.1848], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0082, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 14:05:08,481 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8896, 1.3436, 1.9441, 1.7252, 1.5804, 1.5630, 1.7383, 1.7255], + device='cuda:0'), covar=tensor([0.3956, 0.4511, 0.3432, 0.4156, 0.5105, 0.3921, 0.4827, 0.3494], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0239, 0.0256, 0.0261, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:05:10,711 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:11,913 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:16,633 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:21,322 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.591e+02 1.854e+02 2.332e+02 4.296e+02, threshold=3.709e+02, percent-clipped=1.0 +2023-03-26 14:05:38,140 INFO [finetune.py:976] (0/7) Epoch 12, batch 450, loss[loss=0.1929, simple_loss=0.2508, pruned_loss=0.06748, over 4706.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2601, pruned_loss=0.06328, over 859289.58 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:05:57,260 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:59,772 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:06:00,970 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:06:08,821 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1191, 1.5376, 0.7242, 1.9991, 2.5050, 1.7418, 1.6795, 1.9987], + device='cuda:0'), covar=tensor([0.1486, 0.2033, 0.2417, 0.1199, 0.1792, 0.2079, 0.1485, 0.1930], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0096, 0.0114, 0.0093, 0.0121, 0.0096, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:06:15,170 INFO [finetune.py:976] (0/7) Epoch 12, batch 500, loss[loss=0.2081, simple_loss=0.2736, pruned_loss=0.07127, over 4902.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2591, pruned_loss=0.0634, over 882653.62 frames. ], batch size: 43, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:06:37,049 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.336e+01 1.553e+02 1.855e+02 2.331e+02 4.193e+02, threshold=3.711e+02, percent-clipped=1.0 +2023-03-26 14:06:37,736 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.0365, 4.4094, 4.7014, 4.5915, 4.5494, 4.3519, 5.1360, 1.7223], + device='cuda:0'), covar=tensor([0.0871, 0.1322, 0.1128, 0.1604, 0.1742, 0.1985, 0.0820, 0.7117], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0244, 0.0276, 0.0290, 0.0330, 0.0283, 0.0301, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:06:48,875 INFO [finetune.py:976] (0/7) Epoch 12, batch 550, loss[loss=0.158, simple_loss=0.224, pruned_loss=0.04597, over 4830.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2551, pruned_loss=0.06186, over 899557.23 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:06:58,438 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:03,795 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63577.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:10,288 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:17,000 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-26 14:07:22,327 INFO [finetune.py:976] (0/7) Epoch 12, batch 600, loss[loss=0.1876, simple_loss=0.2531, pruned_loss=0.06102, over 4915.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2547, pruned_loss=0.06175, over 911968.79 frames. ], batch size: 36, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:07:40,193 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:44,849 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.685e+02 2.017e+02 2.531e+02 3.696e+02, threshold=4.034e+02, percent-clipped=0.0 +2023-03-26 14:07:51,116 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63647.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:56,396 INFO [finetune.py:976] (0/7) Epoch 12, batch 650, loss[loss=0.212, simple_loss=0.2965, pruned_loss=0.06373, over 4826.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2588, pruned_loss=0.06311, over 922879.86 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:08:29,866 INFO [finetune.py:976] (0/7) Epoch 12, batch 700, loss[loss=0.1924, simple_loss=0.2696, pruned_loss=0.05765, over 4808.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2602, pruned_loss=0.06305, over 928846.56 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:08:59,836 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.754e+02 2.049e+02 2.499e+02 4.974e+02, threshold=4.098e+02, percent-clipped=3.0 +2023-03-26 14:09:05,237 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8319, 4.1503, 4.4027, 4.5937, 4.5706, 4.2808, 4.9579, 1.4877], + device='cuda:0'), covar=tensor([0.0708, 0.0794, 0.0759, 0.0892, 0.1143, 0.1483, 0.0524, 0.5302], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0244, 0.0276, 0.0290, 0.0330, 0.0283, 0.0300, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:09:11,206 INFO [finetune.py:976] (0/7) Epoch 12, batch 750, loss[loss=0.2229, simple_loss=0.291, pruned_loss=0.0774, over 4824.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2613, pruned_loss=0.06345, over 935494.65 frames. ], batch size: 30, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:09:25,542 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:09:26,757 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:09:56,449 INFO [finetune.py:976] (0/7) Epoch 12, batch 800, loss[loss=0.2377, simple_loss=0.288, pruned_loss=0.09369, over 4815.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2625, pruned_loss=0.06394, over 940476.28 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:10:04,394 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:26,008 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.587e+02 1.868e+02 2.134e+02 3.136e+02, threshold=3.736e+02, percent-clipped=1.0 +2023-03-26 14:10:32,453 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:38,486 INFO [finetune.py:976] (0/7) Epoch 12, batch 850, loss[loss=0.2024, simple_loss=0.2631, pruned_loss=0.07084, over 4906.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2601, pruned_loss=0.06353, over 942664.31 frames. ], batch size: 37, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:10:51,321 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:54,967 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63877.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:22,731 INFO [finetune.py:976] (0/7) Epoch 12, batch 900, loss[loss=0.1932, simple_loss=0.2575, pruned_loss=0.06448, over 4850.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2559, pruned_loss=0.0616, over 944479.84 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:11:23,448 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:29,448 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:35,890 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:35,900 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:44,045 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.611e+02 1.873e+02 2.372e+02 4.297e+02, threshold=3.747e+02, percent-clipped=2.0 +2023-03-26 14:11:47,183 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63942.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:56,453 INFO [finetune.py:976] (0/7) Epoch 12, batch 950, loss[loss=0.2014, simple_loss=0.2685, pruned_loss=0.06717, over 4935.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2548, pruned_loss=0.0615, over 946911.15 frames. ], batch size: 38, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:12:09,059 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6324, 2.0904, 2.8325, 1.9431, 2.6196, 2.8582, 1.9776, 2.8868], + device='cuda:0'), covar=tensor([0.1424, 0.2435, 0.1554, 0.2116, 0.0879, 0.1652, 0.2792, 0.0927], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0206, 0.0194, 0.0191, 0.0179, 0.0215, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:12:10,862 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:12:26,965 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-64000.pt +2023-03-26 14:12:31,104 INFO [finetune.py:976] (0/7) Epoch 12, batch 1000, loss[loss=0.237, simple_loss=0.3027, pruned_loss=0.08567, over 4827.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2575, pruned_loss=0.0626, over 949784.59 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:12:43,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64023.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:12:51,952 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.649e+02 1.875e+02 2.259e+02 3.443e+02, threshold=3.751e+02, percent-clipped=0.0 +2023-03-26 14:12:57,321 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5436, 1.0411, 0.7796, 1.3495, 1.9661, 0.6932, 1.2288, 1.3152], + device='cuda:0'), covar=tensor([0.1401, 0.2187, 0.1835, 0.1265, 0.1687, 0.2016, 0.1621, 0.1970], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0093, 0.0120, 0.0095, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:13:04,254 INFO [finetune.py:976] (0/7) Epoch 12, batch 1050, loss[loss=0.1747, simple_loss=0.2426, pruned_loss=0.05337, over 4837.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2603, pruned_loss=0.06304, over 952413.59 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:13:17,523 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:19,224 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:22,959 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64084.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:29,460 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4998, 1.3171, 1.9652, 1.3309, 1.5907, 1.8090, 1.3098, 1.9554], + device='cuda:0'), covar=tensor([0.1432, 0.2678, 0.1186, 0.1722, 0.0994, 0.1314, 0.3167, 0.0866], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0207, 0.0196, 0.0192, 0.0180, 0.0216, 0.0219, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:13:37,904 INFO [finetune.py:976] (0/7) Epoch 12, batch 1100, loss[loss=0.2484, simple_loss=0.3103, pruned_loss=0.09323, over 4858.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2619, pruned_loss=0.06435, over 950703.53 frames. ], batch size: 31, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:13:53,895 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64124.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:55,103 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:05,897 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.584e+02 1.925e+02 2.329e+02 4.054e+02, threshold=3.850e+02, percent-clipped=2.0 +2023-03-26 14:14:17,902 INFO [finetune.py:976] (0/7) Epoch 12, batch 1150, loss[loss=0.1908, simple_loss=0.2698, pruned_loss=0.05588, over 4875.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2626, pruned_loss=0.06481, over 951174.03 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:14:25,688 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:35,656 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5995, 1.4274, 1.3023, 1.5794, 1.5667, 1.6039, 0.9718, 1.3798], + device='cuda:0'), covar=tensor([0.2013, 0.2020, 0.1846, 0.1595, 0.1624, 0.1164, 0.2448, 0.1743], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0207, 0.0209, 0.0189, 0.0241, 0.0182, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:14:48,850 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64201.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:56,319 INFO [finetune.py:976] (0/7) Epoch 12, batch 1200, loss[loss=0.211, simple_loss=0.2661, pruned_loss=0.07789, over 4827.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2609, pruned_loss=0.06391, over 953251.65 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:15:14,919 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:24,852 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:31,343 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.575e+02 1.833e+02 2.193e+02 5.344e+02, threshold=3.667e+02, percent-clipped=2.0 +2023-03-26 14:15:34,437 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64242.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:43,192 INFO [finetune.py:976] (0/7) Epoch 12, batch 1250, loss[loss=0.1757, simple_loss=0.2372, pruned_loss=0.05707, over 4819.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2582, pruned_loss=0.06333, over 951990.85 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:15:55,101 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:55,714 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:09,180 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:11,097 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:12,418 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-03-26 14:16:27,224 INFO [finetune.py:976] (0/7) Epoch 12, batch 1300, loss[loss=0.1624, simple_loss=0.2208, pruned_loss=0.05198, over 4774.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2561, pruned_loss=0.06279, over 953000.96 frames. ], batch size: 26, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:16:48,477 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.610e+02 1.842e+02 2.244e+02 4.381e+02, threshold=3.684e+02, percent-clipped=1.0 +2023-03-26 14:16:53,411 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:59,917 INFO [finetune.py:976] (0/7) Epoch 12, batch 1350, loss[loss=0.2307, simple_loss=0.2899, pruned_loss=0.08572, over 4894.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2561, pruned_loss=0.06275, over 954879.54 frames. ], batch size: 32, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:17:05,381 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1839, 1.8276, 1.2798, 2.0634, 2.5645, 1.8486, 2.0660, 2.1620], + device='cuda:0'), covar=tensor([0.1272, 0.1781, 0.1870, 0.1061, 0.1612, 0.1612, 0.1292, 0.1592], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0119, 0.0094, 0.0099, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:17:16,040 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:17:19,801 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9779, 1.8953, 1.3632, 1.7734, 1.9232, 1.6420, 2.5443, 1.9428], + device='cuda:0'), covar=tensor([0.1388, 0.2025, 0.3556, 0.3108, 0.2869, 0.1781, 0.2267, 0.2071], + device='cuda:0'), in_proj_covar=tensor([0.0176, 0.0187, 0.0233, 0.0254, 0.0241, 0.0198, 0.0212, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:17:33,430 INFO [finetune.py:976] (0/7) Epoch 12, batch 1400, loss[loss=0.289, simple_loss=0.3404, pruned_loss=0.1187, over 4257.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2603, pruned_loss=0.06388, over 954542.76 frames. ], batch size: 65, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:17:34,172 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7878, 1.3457, 1.0314, 1.7089, 2.2354, 1.4479, 1.6083, 1.7520], + device='cuda:0'), covar=tensor([0.1621, 0.2133, 0.2033, 0.1247, 0.1933, 0.2024, 0.1454, 0.1939], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0113, 0.0092, 0.0120, 0.0094, 0.0099, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:17:34,208 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64406.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:17:54,256 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.617e+02 1.936e+02 2.295e+02 3.610e+02, threshold=3.872e+02, percent-clipped=0.0 +2023-03-26 14:18:06,657 INFO [finetune.py:976] (0/7) Epoch 12, batch 1450, loss[loss=0.1807, simple_loss=0.2489, pruned_loss=0.05622, over 4906.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2613, pruned_loss=0.06424, over 954409.50 frames. ], batch size: 35, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:18:08,007 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9557, 1.4803, 2.0111, 1.8674, 1.7025, 1.6903, 1.8339, 1.8296], + device='cuda:0'), covar=tensor([0.4030, 0.4331, 0.3443, 0.4170, 0.5004, 0.3769, 0.4789, 0.3328], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0237, 0.0254, 0.0259, 0.0256, 0.0230, 0.0273, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:18:13,327 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64465.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:13,904 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:25,575 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0970, 1.9224, 1.5593, 1.6869, 1.8391, 1.7708, 1.8285, 2.6092], + device='cuda:0'), covar=tensor([0.3942, 0.4542, 0.3563, 0.4540, 0.4242, 0.2536, 0.4438, 0.1682], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0258, 0.0223, 0.0275, 0.0242, 0.0209, 0.0246, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:18:31,615 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8893, 1.4534, 0.7331, 1.7756, 2.2496, 1.4211, 1.5726, 1.6533], + device='cuda:0'), covar=tensor([0.1541, 0.2102, 0.2235, 0.1186, 0.1842, 0.1948, 0.1471, 0.2040], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0119, 0.0093, 0.0098, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:18:37,434 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:39,730 INFO [finetune.py:976] (0/7) Epoch 12, batch 1500, loss[loss=0.2499, simple_loss=0.3037, pruned_loss=0.09809, over 4836.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2631, pruned_loss=0.06548, over 954612.63 frames. ], batch size: 49, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:18:46,086 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:54,914 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64526.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:01,494 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.738e+02 2.083e+02 2.672e+02 4.064e+02, threshold=4.165e+02, percent-clipped=1.0 +2023-03-26 14:19:15,861 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64549.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:19,447 INFO [finetune.py:976] (0/7) Epoch 12, batch 1550, loss[loss=0.1723, simple_loss=0.2374, pruned_loss=0.0536, over 4824.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2631, pruned_loss=0.06495, over 955494.77 frames. ], batch size: 38, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:19:20,765 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1593, 3.5297, 3.7994, 3.9222, 3.9232, 3.6758, 4.2300, 1.2934], + device='cuda:0'), covar=tensor([0.0684, 0.0796, 0.0763, 0.0880, 0.0997, 0.1333, 0.0616, 0.5159], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0243, 0.0275, 0.0291, 0.0331, 0.0282, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:19:33,880 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64572.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:45,160 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:45,845 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:54,177 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6195, 1.7474, 1.8856, 1.0972, 1.8400, 2.0669, 1.9714, 1.5913], + device='cuda:0'), covar=tensor([0.1073, 0.0670, 0.0490, 0.0654, 0.0490, 0.0632, 0.0390, 0.0703], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0151, 0.0121, 0.0131, 0.0129, 0.0125, 0.0142, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.3039e-05, 1.1066e-04, 8.7072e-05, 9.4446e-05, 9.1833e-05, 9.1041e-05, + 1.0390e-04, 1.0575e-04], device='cuda:0') +2023-03-26 14:19:56,454 INFO [finetune.py:976] (0/7) Epoch 12, batch 1600, loss[loss=0.2292, simple_loss=0.2768, pruned_loss=0.09081, over 4694.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2594, pruned_loss=0.06344, over 954088.94 frames. ], batch size: 59, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:20:05,239 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.9981, 1.0197, 0.8764, 1.1058, 1.1683, 1.0961, 1.0106, 0.9189], + device='cuda:0'), covar=tensor([0.0350, 0.0285, 0.0615, 0.0264, 0.0258, 0.0433, 0.0313, 0.0385], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0109, 0.0140, 0.0114, 0.0102, 0.0104, 0.0094, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2122e-05, 8.4568e-05, 1.1129e-04, 8.8874e-05, 7.9796e-05, 7.7285e-05, + 7.0960e-05, 8.3672e-05], device='cuda:0') +2023-03-26 14:20:08,115 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64620.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:24,564 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6147, 3.9039, 4.1784, 4.3803, 4.3302, 4.0416, 4.6865, 1.4847], + device='cuda:0'), covar=tensor([0.0704, 0.0889, 0.0741, 0.1008, 0.1085, 0.1514, 0.0578, 0.5515], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0244, 0.0276, 0.0291, 0.0332, 0.0283, 0.0302, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:20:30,242 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.633e+02 1.922e+02 2.431e+02 4.177e+02, threshold=3.845e+02, percent-clipped=1.0 +2023-03-26 14:20:41,540 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64648.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:42,766 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:49,612 INFO [finetune.py:976] (0/7) Epoch 12, batch 1650, loss[loss=0.2024, simple_loss=0.2584, pruned_loss=0.07315, over 4830.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.257, pruned_loss=0.0631, over 953557.44 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:21:05,197 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:05,241 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6100, 1.7119, 1.4673, 1.6446, 2.0550, 1.8774, 1.6810, 1.5062], + device='cuda:0'), covar=tensor([0.0320, 0.0300, 0.0492, 0.0297, 0.0174, 0.0475, 0.0329, 0.0386], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0109, 0.0140, 0.0114, 0.0102, 0.0105, 0.0094, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2246e-05, 8.4639e-05, 1.1134e-04, 8.8915e-05, 8.0000e-05, 7.7566e-05, + 7.1083e-05, 8.3690e-05], device='cuda:0') +2023-03-26 14:21:19,530 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:21,787 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64703.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:22,888 INFO [finetune.py:976] (0/7) Epoch 12, batch 1700, loss[loss=0.2434, simple_loss=0.3069, pruned_loss=0.08999, over 4817.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2552, pruned_loss=0.06231, over 954965.07 frames. ], batch size: 39, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:21:27,418 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64709.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:46,813 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:53,446 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.591e+02 1.930e+02 2.225e+02 5.420e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 14:21:58,922 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:03,104 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64752.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:05,307 INFO [finetune.py:976] (0/7) Epoch 12, batch 1750, loss[loss=0.2673, simple_loss=0.3361, pruned_loss=0.09931, over 4838.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2582, pruned_loss=0.06363, over 954310.05 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:22:10,838 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64764.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:35,806 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2405, 2.0857, 1.7692, 1.9721, 2.2166, 1.9078, 2.4512, 2.1723], + device='cuda:0'), covar=tensor([0.1325, 0.1973, 0.2909, 0.2636, 0.2397, 0.1702, 0.2535, 0.1922], + device='cuda:0'), in_proj_covar=tensor([0.0177, 0.0188, 0.0233, 0.0255, 0.0241, 0.0198, 0.0213, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:22:38,078 INFO [finetune.py:976] (0/7) Epoch 12, batch 1800, loss[loss=0.228, simple_loss=0.2977, pruned_loss=0.07919, over 4831.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.26, pruned_loss=0.06344, over 954336.91 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:22:38,805 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:43,522 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64813.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:48,364 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:58,985 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.622e+02 1.968e+02 2.271e+02 4.247e+02, threshold=3.936e+02, percent-clipped=1.0 +2023-03-26 14:23:11,389 INFO [finetune.py:976] (0/7) Epoch 12, batch 1850, loss[loss=0.2323, simple_loss=0.2925, pruned_loss=0.08604, over 4814.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2613, pruned_loss=0.064, over 954368.17 frames. ], batch size: 40, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:23:33,508 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:23:45,140 INFO [finetune.py:976] (0/7) Epoch 12, batch 1900, loss[loss=0.1675, simple_loss=0.237, pruned_loss=0.04897, over 4788.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2617, pruned_loss=0.06384, over 952870.34 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:24:05,579 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:06,596 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.653e+02 1.911e+02 2.364e+02 4.358e+02, threshold=3.822e+02, percent-clipped=3.0 +2023-03-26 14:24:12,013 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:18,921 INFO [finetune.py:976] (0/7) Epoch 12, batch 1950, loss[loss=0.1641, simple_loss=0.2339, pruned_loss=0.04717, over 4907.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2598, pruned_loss=0.06319, over 954584.05 frames. ], batch size: 37, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:24:40,804 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:58,092 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65001.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:59,915 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:00,451 INFO [finetune.py:976] (0/7) Epoch 12, batch 2000, loss[loss=0.1658, simple_loss=0.2393, pruned_loss=0.0461, over 4848.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2567, pruned_loss=0.06159, over 953833.49 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:25:04,253 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 14:25:21,644 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.609e+02 1.880e+02 2.233e+02 7.388e+02, threshold=3.760e+02, percent-clipped=1.0 +2023-03-26 14:25:21,787 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65037.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:25:34,387 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:42,375 INFO [finetune.py:976] (0/7) Epoch 12, batch 2050, loss[loss=0.1825, simple_loss=0.2616, pruned_loss=0.0517, over 4794.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2542, pruned_loss=0.06127, over 954098.51 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:25:42,470 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:43,061 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5688, 3.6111, 3.3850, 1.5497, 3.7628, 2.8456, 0.8966, 2.5126], + device='cuda:0'), covar=tensor([0.2802, 0.2057, 0.1941, 0.3848, 0.1106, 0.1018, 0.4573, 0.1625], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0176, 0.0161, 0.0129, 0.0157, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 14:25:45,413 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65059.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:21,860 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:24,690 INFO [finetune.py:976] (0/7) Epoch 12, batch 2100, loss[loss=0.1712, simple_loss=0.2262, pruned_loss=0.05813, over 4229.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2541, pruned_loss=0.06168, over 953204.39 frames. ], batch size: 18, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:26:27,107 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:31,282 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0321, 1.7483, 2.1712, 2.0259, 1.8103, 1.7896, 1.9437, 1.9473], + device='cuda:0'), covar=tensor([0.4724, 0.4712, 0.3746, 0.4700, 0.5693, 0.4312, 0.5908, 0.3801], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0238, 0.0256, 0.0260, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:26:32,505 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:34,263 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8756, 1.4301, 0.7622, 1.6366, 2.1338, 1.4193, 1.5787, 1.7127], + device='cuda:0'), covar=tensor([0.1487, 0.2084, 0.2181, 0.1264, 0.2002, 0.2092, 0.1507, 0.1991], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0120, 0.0094, 0.0099, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:26:35,467 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:52,597 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.674e+02 1.985e+02 2.398e+02 5.597e+02, threshold=3.971e+02, percent-clipped=1.0 +2023-03-26 14:27:08,172 INFO [finetune.py:976] (0/7) Epoch 12, batch 2150, loss[loss=0.1807, simple_loss=0.2532, pruned_loss=0.0541, over 4835.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2576, pruned_loss=0.06277, over 953806.09 frames. ], batch size: 30, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:27:11,813 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8671, 1.1651, 1.7159, 1.7242, 1.5152, 1.5113, 1.6050, 1.6072], + device='cuda:0'), covar=tensor([0.4411, 0.4568, 0.3891, 0.4145, 0.5277, 0.4317, 0.5219, 0.3815], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0238, 0.0257, 0.0261, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:27:17,729 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:27:41,486 INFO [finetune.py:976] (0/7) Epoch 12, batch 2200, loss[loss=0.1907, simple_loss=0.2648, pruned_loss=0.05827, over 4923.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2609, pruned_loss=0.06355, over 954416.24 frames. ], batch size: 29, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:28:03,262 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.670e+02 2.055e+02 2.491e+02 4.530e+02, threshold=4.111e+02, percent-clipped=2.0 +2023-03-26 14:28:08,815 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:15,261 INFO [finetune.py:976] (0/7) Epoch 12, batch 2250, loss[loss=0.1891, simple_loss=0.2563, pruned_loss=0.06093, over 4752.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2624, pruned_loss=0.06378, over 955328.67 frames. ], batch size: 27, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:28:27,722 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5361, 1.4401, 1.3234, 1.3733, 1.7440, 1.7078, 1.4946, 1.3170], + device='cuda:0'), covar=tensor([0.0277, 0.0255, 0.0600, 0.0291, 0.0223, 0.0459, 0.0328, 0.0370], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0108, 0.0139, 0.0114, 0.0102, 0.0104, 0.0093, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.1310e-05, 8.4070e-05, 1.1085e-04, 8.8622e-05, 7.9826e-05, 7.6731e-05, + 7.0512e-05, 8.3052e-05], device='cuda:0') +2023-03-26 14:28:36,006 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65285.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:36,176 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 14:28:39,791 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 14:28:41,280 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:48,507 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:49,046 INFO [finetune.py:976] (0/7) Epoch 12, batch 2300, loss[loss=0.1827, simple_loss=0.2625, pruned_loss=0.05145, over 4812.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2631, pruned_loss=0.06401, over 956265.55 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:28:50,947 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1535, 3.6048, 3.7969, 3.9986, 3.8886, 3.6026, 4.2189, 1.3799], + device='cuda:0'), covar=tensor([0.0750, 0.0818, 0.0814, 0.0904, 0.1241, 0.1646, 0.0715, 0.5215], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0243, 0.0275, 0.0291, 0.0330, 0.0282, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:29:07,485 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65332.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:29:10,444 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.717e+01 1.673e+02 1.949e+02 2.267e+02 6.743e+02, threshold=3.897e+02, percent-clipped=1.0 +2023-03-26 14:29:16,550 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:29:20,076 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65352.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:29:22,342 INFO [finetune.py:976] (0/7) Epoch 12, batch 2350, loss[loss=0.1729, simple_loss=0.2368, pruned_loss=0.0545, over 4813.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2614, pruned_loss=0.06412, over 957491.64 frames. ], batch size: 40, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:29:24,809 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65359.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:29:38,192 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4786, 2.3196, 2.8381, 1.6918, 2.6098, 2.5329, 2.1361, 2.8211], + device='cuda:0'), covar=tensor([0.1475, 0.1892, 0.1685, 0.2514, 0.0975, 0.2015, 0.2581, 0.0906], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0208, 0.0197, 0.0194, 0.0179, 0.0217, 0.0219, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:30:02,651 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65401.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:04,996 INFO [finetune.py:976] (0/7) Epoch 12, batch 2400, loss[loss=0.1902, simple_loss=0.2489, pruned_loss=0.06576, over 4859.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2577, pruned_loss=0.06289, over 958755.14 frames. ], batch size: 49, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:30:06,250 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:06,313 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5989, 1.4575, 1.6169, 0.8578, 1.5998, 1.9050, 1.7480, 1.4603], + device='cuda:0'), covar=tensor([0.1098, 0.1035, 0.0547, 0.0708, 0.0499, 0.0551, 0.0451, 0.0831], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0152, 0.0121, 0.0131, 0.0130, 0.0126, 0.0142, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.2970e-05, 1.1143e-04, 8.6971e-05, 9.4606e-05, 9.2613e-05, 9.1432e-05, + 1.0393e-04, 1.0595e-04], device='cuda:0') +2023-03-26 14:30:07,395 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:09,181 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65411.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:26,324 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.534e+02 1.885e+02 2.327e+02 5.518e+02, threshold=3.771e+02, percent-clipped=1.0 +2023-03-26 14:30:34,673 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65449.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:38,806 INFO [finetune.py:976] (0/7) Epoch 12, batch 2450, loss[loss=0.1723, simple_loss=0.2415, pruned_loss=0.05153, over 4899.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2554, pruned_loss=0.06239, over 959397.27 frames. ], batch size: 35, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:30:39,468 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:31:31,331 INFO [finetune.py:976] (0/7) Epoch 12, batch 2500, loss[loss=0.2072, simple_loss=0.2856, pruned_loss=0.06435, over 4907.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2584, pruned_loss=0.06395, over 958242.62 frames. ], batch size: 37, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:31:48,012 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4127, 2.0883, 2.8460, 1.7303, 2.5596, 2.5756, 1.9625, 2.7415], + device='cuda:0'), covar=tensor([0.1240, 0.1897, 0.1293, 0.2096, 0.0790, 0.1632, 0.2410, 0.0929], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0207, 0.0195, 0.0192, 0.0179, 0.0216, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:31:49,233 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65532.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:31:52,644 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.473e+01 1.650e+02 2.020e+02 2.341e+02 4.049e+02, threshold=4.040e+02, percent-clipped=1.0 +2023-03-26 14:32:06,903 INFO [finetune.py:976] (0/7) Epoch 12, batch 2550, loss[loss=0.2197, simple_loss=0.2821, pruned_loss=0.07862, over 4757.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2609, pruned_loss=0.06416, over 957552.09 frames. ], batch size: 54, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:32:27,322 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0658, 1.8988, 1.6636, 2.0158, 2.6728, 2.0526, 1.9446, 1.5717], + device='cuda:0'), covar=tensor([0.2297, 0.2086, 0.2010, 0.1758, 0.1795, 0.1139, 0.2203, 0.1964], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0208, 0.0210, 0.0189, 0.0242, 0.0183, 0.0213, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:32:40,910 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65593.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:32:43,299 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0440, 1.5685, 2.4790, 1.5314, 2.2022, 2.2245, 1.6173, 2.4197], + device='cuda:0'), covar=tensor([0.1311, 0.2237, 0.1246, 0.1961, 0.0819, 0.1624, 0.2789, 0.0831], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0206, 0.0194, 0.0191, 0.0179, 0.0216, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:32:48,714 INFO [finetune.py:976] (0/7) Epoch 12, batch 2600, loss[loss=0.214, simple_loss=0.2883, pruned_loss=0.06985, over 4923.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2618, pruned_loss=0.06472, over 957233.60 frames. ], batch size: 38, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:32:56,502 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1611, 3.6293, 3.7827, 4.0079, 3.8959, 3.6504, 4.2215, 1.3619], + device='cuda:0'), covar=tensor([0.0722, 0.0850, 0.0849, 0.0899, 0.1160, 0.1617, 0.0663, 0.5258], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0242, 0.0276, 0.0291, 0.0330, 0.0282, 0.0300, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:33:06,630 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65632.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:08,924 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4135, 2.3940, 2.4628, 1.7763, 2.4785, 2.6266, 2.4399, 2.1729], + device='cuda:0'), covar=tensor([0.0556, 0.0558, 0.0658, 0.0889, 0.0855, 0.0555, 0.0588, 0.0887], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0135, 0.0142, 0.0125, 0.0123, 0.0143, 0.0145, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:33:10,011 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.524e+01 1.645e+02 2.049e+02 2.494e+02 4.393e+02, threshold=4.097e+02, percent-clipped=1.0 +2023-03-26 14:33:10,745 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3865, 1.6182, 1.7195, 0.9098, 1.5681, 1.9046, 1.8792, 1.5076], + device='cuda:0'), covar=tensor([0.0956, 0.0641, 0.0445, 0.0555, 0.0473, 0.0595, 0.0331, 0.0607], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0153, 0.0122, 0.0132, 0.0130, 0.0126, 0.0143, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.3271e-05, 1.1166e-04, 8.7761e-05, 9.4853e-05, 9.2776e-05, 9.1318e-05, + 1.0414e-04, 1.0645e-04], device='cuda:0') +2023-03-26 14:33:12,530 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:22,370 INFO [finetune.py:976] (0/7) Epoch 12, batch 2650, loss[loss=0.1805, simple_loss=0.2544, pruned_loss=0.05334, over 4690.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2626, pruned_loss=0.06492, over 955763.49 frames. ], batch size: 59, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:33:38,619 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:43,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1019, 1.8771, 2.5042, 3.8042, 2.7645, 2.7153, 0.8111, 3.1235], + device='cuda:0'), covar=tensor([0.1658, 0.1428, 0.1366, 0.0506, 0.0736, 0.1625, 0.2133, 0.0466], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0135, 0.0165, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:33:47,588 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 14:33:55,609 INFO [finetune.py:976] (0/7) Epoch 12, batch 2700, loss[loss=0.2213, simple_loss=0.2876, pruned_loss=0.0775, over 4898.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2619, pruned_loss=0.06418, over 957014.48 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:33:59,759 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65711.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:34:17,009 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.547e+02 1.884e+02 2.200e+02 3.210e+02, threshold=3.769e+02, percent-clipped=0.0 +2023-03-26 14:34:30,258 INFO [finetune.py:976] (0/7) Epoch 12, batch 2750, loss[loss=0.1475, simple_loss=0.2183, pruned_loss=0.03828, over 4824.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2588, pruned_loss=0.06345, over 956842.75 frames. ], batch size: 30, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:34:38,002 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:35:20,816 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8885, 3.4509, 3.5742, 3.7447, 3.6572, 3.4369, 3.9641, 1.3611], + device='cuda:0'), covar=tensor([0.0953, 0.0860, 0.0937, 0.1086, 0.1431, 0.1534, 0.0851, 0.5210], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0277, 0.0292, 0.0332, 0.0283, 0.0302, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:35:30,774 INFO [finetune.py:976] (0/7) Epoch 12, batch 2800, loss[loss=0.1998, simple_loss=0.253, pruned_loss=0.0733, over 4794.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2557, pruned_loss=0.06193, over 957617.72 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:35:52,230 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.914e+01 1.578e+02 1.887e+02 2.176e+02 5.167e+02, threshold=3.774e+02, percent-clipped=1.0 +2023-03-26 14:35:52,954 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65838.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:36:04,208 INFO [finetune.py:976] (0/7) Epoch 12, batch 2850, loss[loss=0.1851, simple_loss=0.2705, pruned_loss=0.04988, over 4855.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2543, pruned_loss=0.06136, over 957251.71 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:36:36,963 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65888.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:36:45,225 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65899.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:36:48,785 INFO [finetune.py:976] (0/7) Epoch 12, batch 2900, loss[loss=0.2188, simple_loss=0.2904, pruned_loss=0.07356, over 4801.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2575, pruned_loss=0.06307, over 955853.22 frames. ], batch size: 45, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:37:10,264 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.527e+02 1.842e+02 2.377e+02 4.547e+02, threshold=3.684e+02, percent-clipped=3.0 +2023-03-26 14:37:10,407 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2594, 2.0585, 1.4775, 0.5689, 1.7086, 1.8207, 1.7119, 1.7898], + device='cuda:0'), covar=tensor([0.0786, 0.0851, 0.1652, 0.2140, 0.1484, 0.2421, 0.2508, 0.0906], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0198, 0.0200, 0.0185, 0.0215, 0.0207, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:37:12,785 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:37:27,692 INFO [finetune.py:976] (0/7) Epoch 12, batch 2950, loss[loss=0.1692, simple_loss=0.2481, pruned_loss=0.04514, over 4913.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2591, pruned_loss=0.06296, over 956984.98 frames. ], batch size: 36, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:37:42,570 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-03-26 14:38:02,424 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:38:21,896 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-66000.pt +2023-03-26 14:38:26,402 INFO [finetune.py:976] (0/7) Epoch 12, batch 3000, loss[loss=0.2127, simple_loss=0.2782, pruned_loss=0.07362, over 4831.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2605, pruned_loss=0.06363, over 955321.06 frames. ], batch size: 47, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:38:26,403 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 14:38:31,307 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2353, 2.0627, 1.4936, 0.5620, 1.7528, 1.8163, 1.7265, 1.8506], + device='cuda:0'), covar=tensor([0.0997, 0.0806, 0.1606, 0.2114, 0.1501, 0.2641, 0.2301, 0.0888], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0198, 0.0201, 0.0185, 0.0214, 0.0207, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:38:37,073 INFO [finetune.py:1010] (0/7) Epoch 12, validation: loss=0.1571, simple_loss=0.2281, pruned_loss=0.04309, over 2265189.00 frames. +2023-03-26 14:38:37,073 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 14:38:53,471 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 14:38:58,511 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.619e+02 1.943e+02 2.343e+02 4.325e+02, threshold=3.886e+02, percent-clipped=3.0 +2023-03-26 14:39:21,449 INFO [finetune.py:976] (0/7) Epoch 12, batch 3050, loss[loss=0.1907, simple_loss=0.257, pruned_loss=0.06223, over 4852.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.262, pruned_loss=0.06403, over 957342.06 frames. ], batch size: 31, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:39:32,828 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-03-26 14:39:53,014 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 14:40:36,352 INFO [finetune.py:976] (0/7) Epoch 12, batch 3100, loss[loss=0.1542, simple_loss=0.2195, pruned_loss=0.04444, over 4820.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2597, pruned_loss=0.06347, over 956455.05 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:41:00,819 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8670, 1.7018, 1.4917, 1.8608, 2.2501, 1.9162, 1.5036, 1.5222], + device='cuda:0'), covar=tensor([0.2043, 0.2028, 0.1920, 0.1606, 0.1697, 0.1152, 0.2485, 0.1811], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0211, 0.0190, 0.0243, 0.0185, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:41:19,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.628e+02 1.972e+02 2.398e+02 4.316e+02, threshold=3.945e+02, percent-clipped=3.0 +2023-03-26 14:41:27,575 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:41:40,145 INFO [finetune.py:976] (0/7) Epoch 12, batch 3150, loss[loss=0.221, simple_loss=0.2724, pruned_loss=0.08477, over 4938.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2559, pruned_loss=0.0617, over 955739.42 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:41:43,350 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66160.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:42:24,560 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66188.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:42:32,941 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66194.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:42:40,452 INFO [finetune.py:976] (0/7) Epoch 12, batch 3200, loss[loss=0.2052, simple_loss=0.2607, pruned_loss=0.07486, over 4858.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2524, pruned_loss=0.06033, over 955345.76 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:42:40,558 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66205.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:42:45,935 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1166, 1.8817, 2.4748, 3.7653, 2.6772, 2.7222, 1.4818, 2.9654], + device='cuda:0'), covar=tensor([0.1656, 0.1418, 0.1399, 0.0612, 0.0713, 0.1284, 0.1790, 0.0593], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0117, 0.0136, 0.0167, 0.0102, 0.0139, 0.0127, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 14:42:51,180 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66221.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:43:01,314 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66236.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:43:01,823 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.602e+02 1.878e+02 2.419e+02 4.134e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-26 14:43:02,558 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1011, 3.2168, 3.1360, 2.5097, 3.1969, 3.4964, 3.3136, 2.9880], + device='cuda:0'), covar=tensor([0.0468, 0.0489, 0.0608, 0.0682, 0.0677, 0.0501, 0.0497, 0.0783], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0141, 0.0123, 0.0122, 0.0142, 0.0142, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:43:14,215 INFO [finetune.py:976] (0/7) Epoch 12, batch 3250, loss[loss=0.1921, simple_loss=0.2763, pruned_loss=0.05401, over 4838.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2528, pruned_loss=0.06063, over 953696.35 frames. ], batch size: 47, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:43:16,221 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66258.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:43:25,200 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9214, 1.8895, 1.9230, 1.2839, 1.9812, 2.0935, 1.9744, 1.6409], + device='cuda:0'), covar=tensor([0.0661, 0.0666, 0.0752, 0.0914, 0.0642, 0.0628, 0.0625, 0.1100], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0140, 0.0122, 0.0121, 0.0141, 0.0141, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:43:31,219 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2810, 2.2413, 1.9633, 1.0809, 2.1398, 1.8503, 1.6789, 2.0843], + device='cuda:0'), covar=tensor([0.1094, 0.0760, 0.1603, 0.1990, 0.1342, 0.2102, 0.2122, 0.1039], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0199, 0.0201, 0.0185, 0.0214, 0.0207, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:43:48,267 INFO [finetune.py:976] (0/7) Epoch 12, batch 3300, loss[loss=0.2256, simple_loss=0.2965, pruned_loss=0.07738, over 4800.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2576, pruned_loss=0.06211, over 954654.90 frames. ], batch size: 41, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:43:56,993 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66319.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:44:14,652 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.642e+02 1.904e+02 2.370e+02 4.024e+02, threshold=3.808e+02, percent-clipped=2.0 +2023-03-26 14:44:29,722 INFO [finetune.py:976] (0/7) Epoch 12, batch 3350, loss[loss=0.2713, simple_loss=0.332, pruned_loss=0.1053, over 4900.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2603, pruned_loss=0.06312, over 953211.58 frames. ], batch size: 43, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:45:02,613 INFO [finetune.py:976] (0/7) Epoch 12, batch 3400, loss[loss=0.2004, simple_loss=0.2753, pruned_loss=0.0628, over 4789.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2622, pruned_loss=0.06404, over 952897.71 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:45:24,436 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.693e+02 1.994e+02 2.432e+02 3.824e+02, threshold=3.988e+02, percent-clipped=2.0 +2023-03-26 14:45:36,092 INFO [finetune.py:976] (0/7) Epoch 12, batch 3450, loss[loss=0.1707, simple_loss=0.2409, pruned_loss=0.05026, over 4768.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2601, pruned_loss=0.0628, over 951842.34 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:46:02,768 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66494.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:46:06,883 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66500.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:46:09,848 INFO [finetune.py:976] (0/7) Epoch 12, batch 3500, loss[loss=0.2265, simple_loss=0.2755, pruned_loss=0.08872, over 4727.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2576, pruned_loss=0.06249, over 950491.55 frames. ], batch size: 59, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:46:15,049 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4388, 2.2680, 1.7948, 2.5131, 2.3131, 1.9372, 2.8671, 2.3856], + device='cuda:0'), covar=tensor([0.1387, 0.2719, 0.3429, 0.2885, 0.2669, 0.1798, 0.3760, 0.2028], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0188, 0.0233, 0.0255, 0.0242, 0.0198, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:46:20,799 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66516.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:46:23,824 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66521.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:46:36,432 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.655e+02 1.937e+02 2.486e+02 6.010e+02, threshold=3.875e+02, percent-clipped=2.0 +2023-03-26 14:46:39,999 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66542.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:46:57,616 INFO [finetune.py:976] (0/7) Epoch 12, batch 3550, loss[loss=0.1847, simple_loss=0.2503, pruned_loss=0.05955, over 4833.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2565, pruned_loss=0.06247, over 952906.75 frames. ], batch size: 40, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:47:23,070 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66582.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:47:34,941 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5454, 2.4972, 2.6092, 1.3370, 2.8637, 3.1043, 2.7081, 2.4244], + device='cuda:0'), covar=tensor([0.1299, 0.0856, 0.0367, 0.0700, 0.0421, 0.0747, 0.0387, 0.0666], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0124, 0.0132, 0.0131, 0.0127, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.3913e-05, 1.1222e-04, 8.9220e-05, 9.5215e-05, 9.2719e-05, 9.2041e-05, + 1.0526e-04, 1.0707e-04], device='cuda:0') +2023-03-26 14:47:43,651 INFO [finetune.py:976] (0/7) Epoch 12, batch 3600, loss[loss=0.1596, simple_loss=0.2308, pruned_loss=0.04423, over 4822.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2533, pruned_loss=0.06086, over 954791.42 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:47:53,380 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66614.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:48:08,740 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.581e+02 1.999e+02 2.430e+02 3.919e+02, threshold=3.999e+02, percent-clipped=1.0 +2023-03-26 14:48:21,117 INFO [finetune.py:976] (0/7) Epoch 12, batch 3650, loss[loss=0.2706, simple_loss=0.3358, pruned_loss=0.1027, over 4812.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2564, pruned_loss=0.06246, over 953494.25 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:48:54,888 INFO [finetune.py:976] (0/7) Epoch 12, batch 3700, loss[loss=0.1865, simple_loss=0.2647, pruned_loss=0.05419, over 4852.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2595, pruned_loss=0.06266, over 955034.54 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:49:14,923 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66731.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:49:18,458 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.675e+02 1.999e+02 2.466e+02 3.717e+02, threshold=3.997e+02, percent-clipped=0.0 +2023-03-26 14:49:38,724 INFO [finetune.py:976] (0/7) Epoch 12, batch 3750, loss[loss=0.1829, simple_loss=0.2675, pruned_loss=0.04912, over 4858.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2618, pruned_loss=0.06346, over 955844.13 frames. ], batch size: 31, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:03,078 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66792.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:50:09,002 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66800.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:50:12,436 INFO [finetune.py:976] (0/7) Epoch 12, batch 3800, loss[loss=0.2212, simple_loss=0.2736, pruned_loss=0.08434, over 4782.00 frames. ], tot_loss[loss=0.1954, simple_loss=0.2629, pruned_loss=0.06397, over 954223.01 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:19,256 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66816.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:50:33,176 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5598, 1.5171, 1.3700, 1.4086, 1.8406, 1.7236, 1.5067, 1.3057], + device='cuda:0'), covar=tensor([0.0295, 0.0258, 0.0600, 0.0282, 0.0198, 0.0409, 0.0322, 0.0352], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0110, 0.0141, 0.0115, 0.0103, 0.0105, 0.0095, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.2604e-05, 8.5300e-05, 1.1204e-04, 8.9374e-05, 8.0706e-05, 7.8019e-05, + 7.1969e-05, 8.4380e-05], device='cuda:0') +2023-03-26 14:50:33,876 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-26 14:50:34,225 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.175e+02 1.668e+02 2.101e+02 2.666e+02 4.038e+02, threshold=4.202e+02, percent-clipped=1.0 +2023-03-26 14:50:40,356 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:50:45,462 INFO [finetune.py:976] (0/7) Epoch 12, batch 3850, loss[loss=0.1656, simple_loss=0.2347, pruned_loss=0.0482, over 4760.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2606, pruned_loss=0.06283, over 953125.05 frames. ], batch size: 26, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:51,372 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66864.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:51:00,116 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66877.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:51:04,704 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8344, 1.5919, 1.5058, 1.7663, 2.4872, 1.8579, 1.6163, 1.4526], + device='cuda:0'), covar=tensor([0.2225, 0.2412, 0.2125, 0.1839, 0.1611, 0.1355, 0.2516, 0.2191], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0207, 0.0211, 0.0190, 0.0241, 0.0183, 0.0213, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:51:17,301 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9841, 1.6913, 2.3328, 1.5822, 1.9547, 2.3066, 1.6405, 2.3742], + device='cuda:0'), covar=tensor([0.1247, 0.1979, 0.1371, 0.1886, 0.0891, 0.1315, 0.2568, 0.0779], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0207, 0.0195, 0.0193, 0.0180, 0.0216, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:51:18,885 INFO [finetune.py:976] (0/7) Epoch 12, batch 3900, loss[loss=0.1832, simple_loss=0.2492, pruned_loss=0.05857, over 4825.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2573, pruned_loss=0.06205, over 955629.88 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:51:24,964 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66914.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:51:40,885 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.325e+01 1.578e+02 1.785e+02 2.294e+02 5.103e+02, threshold=3.570e+02, percent-clipped=1.0 +2023-03-26 14:51:51,222 INFO [finetune.py:976] (0/7) Epoch 12, batch 3950, loss[loss=0.156, simple_loss=0.2246, pruned_loss=0.04376, over 4802.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2543, pruned_loss=0.06031, over 956597.02 frames. ], batch size: 29, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:51:53,039 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:51:58,481 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66962.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:52:09,534 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 14:52:21,080 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0916, 2.0143, 2.2523, 1.3787, 2.2117, 2.2790, 2.1102, 1.8589], + device='cuda:0'), covar=tensor([0.0589, 0.0617, 0.0583, 0.0914, 0.0750, 0.0623, 0.0593, 0.1048], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0140, 0.0122, 0.0121, 0.0141, 0.0140, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:52:28,886 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.4510, 4.7718, 5.0421, 5.2916, 5.2078, 4.8802, 5.5775, 1.7669], + device='cuda:0'), covar=tensor([0.0748, 0.0739, 0.0685, 0.0852, 0.1172, 0.1448, 0.0512, 0.5757], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0242, 0.0275, 0.0290, 0.0329, 0.0281, 0.0300, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:52:47,108 INFO [finetune.py:976] (0/7) Epoch 12, batch 4000, loss[loss=0.1358, simple_loss=0.2093, pruned_loss=0.03108, over 4824.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2544, pruned_loss=0.06064, over 956400.53 frames. ], batch size: 25, lr: 3.63e-03, grad_scale: 32.0 +2023-03-26 14:53:04,486 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67018.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:53:14,099 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6115, 1.6267, 1.9166, 1.9863, 1.7449, 3.0159, 1.5242, 1.7280], + device='cuda:0'), covar=tensor([0.0889, 0.1553, 0.1336, 0.0825, 0.1363, 0.0297, 0.1331, 0.1474], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0075, 0.0078, 0.0091, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 14:53:18,612 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.442e+01 1.595e+02 2.014e+02 2.521e+02 4.335e+02, threshold=4.027e+02, percent-clipped=3.0 +2023-03-26 14:53:28,873 INFO [finetune.py:976] (0/7) Epoch 12, batch 4050, loss[loss=0.2228, simple_loss=0.2875, pruned_loss=0.07908, over 4861.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.258, pruned_loss=0.0627, over 956319.55 frames. ], batch size: 34, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:53:38,337 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9038, 1.6438, 2.0787, 1.3873, 1.8499, 2.1463, 2.0213, 1.3147], + device='cuda:0'), covar=tensor([0.0678, 0.0869, 0.0674, 0.0942, 0.0737, 0.0632, 0.0687, 0.1704], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0140, 0.0123, 0.0121, 0.0141, 0.0141, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:53:45,060 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9652, 1.8402, 1.5120, 1.6347, 1.9636, 1.6410, 2.1522, 1.9441], + device='cuda:0'), covar=tensor([0.1383, 0.2164, 0.3220, 0.2830, 0.2598, 0.1740, 0.3531, 0.1824], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0189, 0.0235, 0.0257, 0.0244, 0.0200, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:53:47,877 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:53:50,802 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67087.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:54:02,029 INFO [finetune.py:976] (0/7) Epoch 12, batch 4100, loss[loss=0.1822, simple_loss=0.2537, pruned_loss=0.05532, over 4761.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2605, pruned_loss=0.06349, over 952429.68 frames. ], batch size: 28, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:54:07,502 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3123, 1.2189, 1.6831, 2.4646, 1.6361, 2.1652, 0.8920, 2.0393], + device='cuda:0'), covar=tensor([0.1773, 0.1544, 0.1121, 0.0730, 0.0902, 0.1183, 0.1587, 0.0660], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0135, 0.0165, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:54:29,987 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.725e+02 1.998e+02 2.409e+02 3.172e+02, threshold=3.997e+02, percent-clipped=0.0 +2023-03-26 14:54:34,073 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:54:44,210 INFO [finetune.py:976] (0/7) Epoch 12, batch 4150, loss[loss=0.1921, simple_loss=0.2532, pruned_loss=0.06552, over 4732.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2637, pruned_loss=0.06506, over 952816.84 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:54:45,003 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 14:54:59,055 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67177.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:55:17,547 INFO [finetune.py:976] (0/7) Epoch 12, batch 4200, loss[loss=0.1771, simple_loss=0.2469, pruned_loss=0.05361, over 4827.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.263, pruned_loss=0.06401, over 953761.32 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:55:30,586 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67225.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:55:39,457 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.549e+02 1.852e+02 2.427e+02 4.145e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-26 14:55:50,530 INFO [finetune.py:976] (0/7) Epoch 12, batch 4250, loss[loss=0.1651, simple_loss=0.2281, pruned_loss=0.0511, over 4147.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2601, pruned_loss=0.06289, over 954473.83 frames. ], batch size: 18, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:56:08,563 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-26 14:56:32,185 INFO [finetune.py:976] (0/7) Epoch 12, batch 4300, loss[loss=0.2035, simple_loss=0.2757, pruned_loss=0.06565, over 4917.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2573, pruned_loss=0.06224, over 953423.75 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:56:37,189 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67313.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:56:43,972 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8327, 1.1576, 1.9751, 1.8201, 1.6204, 1.5655, 1.6727, 1.7238], + device='cuda:0'), covar=tensor([0.3669, 0.4009, 0.3096, 0.3640, 0.4399, 0.3694, 0.4306, 0.3169], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0238, 0.0254, 0.0261, 0.0257, 0.0233, 0.0274, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:56:47,468 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2265, 3.6582, 3.8598, 4.0525, 4.0159, 3.7699, 4.3121, 1.4679], + device='cuda:0'), covar=tensor([0.0743, 0.0813, 0.0804, 0.1030, 0.1049, 0.1346, 0.0664, 0.5124], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0242, 0.0275, 0.0290, 0.0327, 0.0280, 0.0300, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:56:51,678 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1288, 1.9522, 1.6926, 1.9891, 1.8961, 1.8638, 1.9074, 2.6630], + device='cuda:0'), covar=tensor([0.4319, 0.5151, 0.3651, 0.4451, 0.4528, 0.2790, 0.4590, 0.1802], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0226, 0.0279, 0.0245, 0.0212, 0.0249, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 14:56:54,399 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.954e+01 1.550e+02 1.913e+02 2.348e+02 5.397e+02, threshold=3.825e+02, percent-clipped=3.0 +2023-03-26 14:57:05,085 INFO [finetune.py:976] (0/7) Epoch 12, batch 4350, loss[loss=0.1771, simple_loss=0.2442, pruned_loss=0.055, over 4835.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2535, pruned_loss=0.06098, over 950235.75 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:57:14,789 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6258, 1.1146, 0.8832, 1.5668, 1.9844, 1.2515, 1.3046, 1.6169], + device='cuda:0'), covar=tensor([0.1475, 0.2224, 0.1982, 0.1211, 0.2025, 0.2009, 0.1556, 0.1867], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0120, 0.0094, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 14:57:28,528 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67387.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:57:39,911 INFO [finetune.py:976] (0/7) Epoch 12, batch 4400, loss[loss=0.2565, simple_loss=0.3161, pruned_loss=0.09849, over 4893.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2559, pruned_loss=0.06244, over 951846.90 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:57:59,499 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4511, 1.2803, 1.3064, 1.2646, 1.6065, 1.5120, 1.3946, 1.2403], + device='cuda:0'), covar=tensor([0.0271, 0.0307, 0.0528, 0.0288, 0.0215, 0.0481, 0.0334, 0.0382], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0109, 0.0139, 0.0113, 0.0102, 0.0104, 0.0094, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.1985e-05, 8.4404e-05, 1.1035e-04, 8.8298e-05, 7.9599e-05, 7.7110e-05, + 7.1107e-05, 8.3550e-05], device='cuda:0') +2023-03-26 14:58:04,820 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5832, 1.4713, 1.4761, 1.5006, 0.9967, 3.0057, 1.1143, 1.5800], + device='cuda:0'), covar=tensor([0.3325, 0.2587, 0.2149, 0.2452, 0.1999, 0.0252, 0.2689, 0.1347], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 14:58:14,137 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67435.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:58:16,973 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.656e+02 1.972e+02 2.339e+02 4.406e+02, threshold=3.944e+02, percent-clipped=2.0 +2023-03-26 14:58:17,067 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67439.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:58:30,809 INFO [finetune.py:976] (0/7) Epoch 12, batch 4450, loss[loss=0.1885, simple_loss=0.2625, pruned_loss=0.05729, over 4706.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2584, pruned_loss=0.06328, over 948916.24 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:59:03,972 INFO [finetune.py:976] (0/7) Epoch 12, batch 4500, loss[loss=0.228, simple_loss=0.2858, pruned_loss=0.08512, over 4900.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2617, pruned_loss=0.06431, over 951672.53 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:59:09,862 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67513.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:59:12,936 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 14:59:26,032 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.686e+02 1.980e+02 2.352e+02 4.001e+02, threshold=3.961e+02, percent-clipped=1.0 +2023-03-26 14:59:37,241 INFO [finetune.py:976] (0/7) Epoch 12, batch 4550, loss[loss=0.2143, simple_loss=0.2727, pruned_loss=0.07793, over 4902.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2631, pruned_loss=0.06489, over 952858.15 frames. ], batch size: 36, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:59:45,882 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3683, 1.2784, 1.3193, 1.3326, 0.7555, 2.2657, 0.7361, 1.2058], + device='cuda:0'), covar=tensor([0.3510, 0.2551, 0.2208, 0.2407, 0.2152, 0.0366, 0.2955, 0.1385], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 14:59:56,258 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67574.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:00:19,986 INFO [finetune.py:976] (0/7) Epoch 12, batch 4600, loss[loss=0.1912, simple_loss=0.2554, pruned_loss=0.06355, over 4896.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2628, pruned_loss=0.06426, over 954416.99 frames. ], batch size: 36, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:00:21,339 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6376, 1.5163, 1.5107, 1.5852, 1.2244, 3.4986, 1.4222, 1.8653], + device='cuda:0'), covar=tensor([0.3230, 0.2428, 0.2053, 0.2279, 0.1697, 0.0189, 0.2591, 0.1205], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0116, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:00:24,945 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:00:42,114 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.477e+02 1.878e+02 2.272e+02 4.960e+02, threshold=3.756e+02, percent-clipped=1.0 +2023-03-26 15:00:48,164 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 15:00:53,234 INFO [finetune.py:976] (0/7) Epoch 12, batch 4650, loss[loss=0.1756, simple_loss=0.2248, pruned_loss=0.06313, over 4157.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2597, pruned_loss=0.06317, over 953900.06 frames. ], batch size: 65, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:00:56,977 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67661.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:10,452 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:17,286 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 15:01:17,715 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:31,302 INFO [finetune.py:976] (0/7) Epoch 12, batch 4700, loss[loss=0.1886, simple_loss=0.2399, pruned_loss=0.06865, over 4262.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2562, pruned_loss=0.06199, over 951972.92 frames. ], batch size: 66, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:01:32,597 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3852, 3.7953, 3.9822, 4.2030, 4.1627, 3.8743, 4.4364, 1.3887], + device='cuda:0'), covar=tensor([0.0675, 0.0852, 0.0902, 0.0781, 0.1053, 0.1550, 0.0713, 0.5702], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0243, 0.0276, 0.0290, 0.0329, 0.0280, 0.0301, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:01:40,005 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 15:01:44,283 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 15:01:56,970 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.553e+02 1.823e+02 2.116e+02 3.808e+02, threshold=3.646e+02, percent-clipped=1.0 +2023-03-26 15:01:57,094 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67739.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:58,304 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:02:05,942 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67753.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:02:07,554 INFO [finetune.py:976] (0/7) Epoch 12, batch 4750, loss[loss=0.207, simple_loss=0.2741, pruned_loss=0.06993, over 4761.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2541, pruned_loss=0.06107, over 951457.56 frames. ], batch size: 26, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:02:28,907 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67787.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:02:40,332 INFO [finetune.py:976] (0/7) Epoch 12, batch 4800, loss[loss=0.2766, simple_loss=0.3241, pruned_loss=0.1146, over 4815.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2562, pruned_loss=0.06221, over 951736.51 frames. ], batch size: 40, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:03:07,510 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.756e+02 1.975e+02 2.556e+02 4.813e+02, threshold=3.950e+02, percent-clipped=3.0 +2023-03-26 15:03:25,933 INFO [finetune.py:976] (0/7) Epoch 12, batch 4850, loss[loss=0.1864, simple_loss=0.2501, pruned_loss=0.06135, over 4834.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2578, pruned_loss=0.06192, over 948591.27 frames. ], batch size: 39, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:03:39,903 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67869.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:04:03,160 INFO [finetune.py:976] (0/7) Epoch 12, batch 4900, loss[loss=0.1715, simple_loss=0.231, pruned_loss=0.05599, over 3943.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2603, pruned_loss=0.06301, over 950021.32 frames. ], batch size: 17, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:04:15,492 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6611, 1.4648, 2.0848, 3.4068, 2.4007, 2.3903, 1.1626, 2.6020], + device='cuda:0'), covar=tensor([0.1750, 0.1473, 0.1326, 0.0562, 0.0765, 0.1536, 0.1741, 0.0619], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0134, 0.0165, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:04:26,941 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.717e+02 1.971e+02 2.418e+02 4.222e+02, threshold=3.942e+02, percent-clipped=1.0 +2023-03-26 15:04:36,657 INFO [finetune.py:976] (0/7) Epoch 12, batch 4950, loss[loss=0.188, simple_loss=0.2639, pruned_loss=0.05601, over 4888.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2619, pruned_loss=0.06328, over 951724.39 frames. ], batch size: 43, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:04:53,969 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:16,837 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-68000.pt +2023-03-26 15:05:20,898 INFO [finetune.py:976] (0/7) Epoch 12, batch 5000, loss[loss=0.1736, simple_loss=0.23, pruned_loss=0.05866, over 4837.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2601, pruned_loss=0.06255, over 952226.18 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:05:32,892 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 15:05:41,203 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:43,413 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.543e+02 1.867e+02 2.301e+02 3.447e+02, threshold=3.734e+02, percent-clipped=0.0 +2023-03-26 15:05:46,338 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68042.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:50,380 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68048.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:05:54,510 INFO [finetune.py:976] (0/7) Epoch 12, batch 5050, loss[loss=0.1707, simple_loss=0.2363, pruned_loss=0.05253, over 4825.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2574, pruned_loss=0.06182, over 954428.14 frames. ], batch size: 25, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:06:15,501 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68087.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:06:27,689 INFO [finetune.py:976] (0/7) Epoch 12, batch 5100, loss[loss=0.2461, simple_loss=0.2926, pruned_loss=0.09976, over 4757.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2547, pruned_loss=0.06102, over 955663.95 frames. ], batch size: 54, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:06:52,464 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5407, 1.0189, 0.8664, 1.5257, 2.0273, 1.0649, 1.3633, 1.4857], + device='cuda:0'), covar=tensor([0.1994, 0.2991, 0.2505, 0.1564, 0.2415, 0.2603, 0.2163, 0.2680], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0094, 0.0100, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:06:53,056 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9259, 3.4318, 3.5861, 3.7939, 3.7268, 3.4367, 3.9533, 1.2230], + device='cuda:0'), covar=tensor([0.0819, 0.0832, 0.0918, 0.0971, 0.1204, 0.1546, 0.0867, 0.5438], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0240, 0.0275, 0.0289, 0.0327, 0.0281, 0.0299, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:06:59,402 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.565e+02 1.837e+02 2.198e+02 4.078e+02, threshold=3.675e+02, percent-clipped=2.0 +2023-03-26 15:07:05,457 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68148.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 15:07:06,505 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4099, 2.2987, 1.9731, 2.4823, 2.1741, 2.1825, 2.2007, 3.1065], + device='cuda:0'), covar=tensor([0.3919, 0.4896, 0.3549, 0.4759, 0.4791, 0.2556, 0.4564, 0.1790], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0224, 0.0276, 0.0244, 0.0211, 0.0247, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:07:10,948 INFO [finetune.py:976] (0/7) Epoch 12, batch 5150, loss[loss=0.1964, simple_loss=0.2652, pruned_loss=0.06383, over 4766.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2546, pruned_loss=0.0611, over 953222.08 frames. ], batch size: 54, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:07:19,519 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:07:28,352 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1115, 1.0175, 1.0545, 0.5003, 0.9300, 1.1944, 1.2066, 1.0308], + device='cuda:0'), covar=tensor([0.0883, 0.0600, 0.0520, 0.0568, 0.0520, 0.0725, 0.0386, 0.0686], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0156, 0.0125, 0.0133, 0.0133, 0.0129, 0.0147, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.4838e-05, 1.1439e-04, 9.0108e-05, 9.6116e-05, 9.4463e-05, 9.3324e-05, + 1.0684e-04, 1.0797e-04], device='cuda:0') +2023-03-26 15:07:31,443 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 15:07:43,719 INFO [finetune.py:976] (0/7) Epoch 12, batch 5200, loss[loss=0.2544, simple_loss=0.3287, pruned_loss=0.09008, over 4745.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2581, pruned_loss=0.06217, over 953847.20 frames. ], batch size: 59, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:07:51,086 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:08:05,786 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.664e+02 1.889e+02 2.252e+02 3.665e+02, threshold=3.778e+02, percent-clipped=0.0 +2023-03-26 15:08:13,735 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0679, 0.9659, 0.9961, 0.2767, 0.8868, 1.1867, 1.1883, 1.0139], + device='cuda:0'), covar=tensor([0.1016, 0.0877, 0.0552, 0.0657, 0.0640, 0.0701, 0.0486, 0.0776], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0156, 0.0125, 0.0133, 0.0133, 0.0128, 0.0146, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.4218e-05, 1.1376e-04, 8.9815e-05, 9.5708e-05, 9.4280e-05, 9.3039e-05, + 1.0650e-04, 1.0755e-04], device='cuda:0') +2023-03-26 15:08:15,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4604, 1.5878, 1.3096, 1.5925, 1.8719, 1.6956, 1.5244, 1.3160], + device='cuda:0'), covar=tensor([0.0344, 0.0252, 0.0537, 0.0248, 0.0190, 0.0425, 0.0290, 0.0372], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0109, 0.0140, 0.0113, 0.0102, 0.0105, 0.0095, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2514e-05, 8.4401e-05, 1.1117e-04, 8.8462e-05, 7.9894e-05, 7.7776e-05, + 7.2053e-05, 8.4118e-05], device='cuda:0') +2023-03-26 15:08:16,523 INFO [finetune.py:976] (0/7) Epoch 12, batch 5250, loss[loss=0.209, simple_loss=0.2771, pruned_loss=0.07046, over 4146.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2598, pruned_loss=0.06253, over 954109.32 frames. ], batch size: 65, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:08:35,656 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 15:09:03,119 INFO [finetune.py:976] (0/7) Epoch 12, batch 5300, loss[loss=0.1836, simple_loss=0.2549, pruned_loss=0.0562, over 4820.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2603, pruned_loss=0.06254, over 954479.04 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:09:16,891 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6759, 1.5242, 1.0084, 0.2711, 1.3398, 1.4673, 1.5362, 1.4926], + device='cuda:0'), covar=tensor([0.0821, 0.0796, 0.1208, 0.1954, 0.1236, 0.2275, 0.1982, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0195, 0.0197, 0.0182, 0.0210, 0.0205, 0.0219, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:09:24,973 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:25,572 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:26,706 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.294e+02 1.838e+02 2.123e+02 2.651e+02 4.524e+02, threshold=4.245e+02, percent-clipped=5.0 +2023-03-26 15:09:27,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0401, 1.8881, 1.7259, 1.9577, 2.0099, 1.7622, 2.2154, 1.9998], + device='cuda:0'), covar=tensor([0.1312, 0.2153, 0.2849, 0.2136, 0.2220, 0.1505, 0.2831, 0.1733], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0186, 0.0232, 0.0255, 0.0241, 0.0198, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:09:32,251 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:36,484 INFO [finetune.py:976] (0/7) Epoch 12, batch 5350, loss[loss=0.1425, simple_loss=0.2097, pruned_loss=0.03767, over 4762.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.26, pruned_loss=0.06229, over 954843.56 frames. ], batch size: 26, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:09:50,627 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2184, 2.0156, 1.7172, 1.9507, 2.1276, 1.8411, 2.3477, 2.1471], + device='cuda:0'), covar=tensor([0.1429, 0.2382, 0.3358, 0.2769, 0.2601, 0.1709, 0.3519, 0.1937], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0186, 0.0232, 0.0255, 0.0242, 0.0198, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:09:55,984 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:04,681 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:06,563 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:10,285 INFO [finetune.py:976] (0/7) Epoch 12, batch 5400, loss[loss=0.1867, simple_loss=0.2508, pruned_loss=0.06135, over 4901.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2576, pruned_loss=0.06145, over 952502.30 frames. ], batch size: 36, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:10:36,142 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0354, 1.8187, 1.6717, 2.1192, 2.2630, 2.0472, 1.5491, 1.7145], + device='cuda:0'), covar=tensor([0.2166, 0.1996, 0.1926, 0.1599, 0.1800, 0.1127, 0.2455, 0.1992], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0207, 0.0210, 0.0190, 0.0239, 0.0182, 0.0213, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:10:40,850 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.541e+02 1.801e+02 2.082e+02 4.267e+02, threshold=3.602e+02, percent-clipped=1.0 +2023-03-26 15:10:44,356 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68443.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:10:51,598 INFO [finetune.py:976] (0/7) Epoch 12, batch 5450, loss[loss=0.1976, simple_loss=0.258, pruned_loss=0.06857, over 4865.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2547, pruned_loss=0.06032, over 952911.86 frames. ], batch size: 31, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:10:54,131 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8675, 1.5936, 2.0512, 2.0473, 1.8172, 3.6025, 1.4713, 1.6878], + device='cuda:0'), covar=tensor([0.0948, 0.1923, 0.1187, 0.1043, 0.1684, 0.0261, 0.1659, 0.2044], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:10:54,743 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68460.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:11:00,548 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7404, 3.7158, 3.5370, 1.7231, 3.8554, 2.7588, 0.6827, 2.4325], + device='cuda:0'), covar=tensor([0.2135, 0.2146, 0.1609, 0.3418, 0.1143, 0.1172, 0.4597, 0.1865], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0174, 0.0159, 0.0128, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 15:11:12,624 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0469, 1.8526, 1.6478, 1.6342, 1.7933, 1.8364, 1.7584, 2.4791], + device='cuda:0'), covar=tensor([0.4663, 0.4919, 0.3891, 0.4546, 0.4723, 0.2679, 0.4381, 0.2056], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0223, 0.0275, 0.0244, 0.0210, 0.0246, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:11:24,505 INFO [finetune.py:976] (0/7) Epoch 12, batch 5500, loss[loss=0.1957, simple_loss=0.2646, pruned_loss=0.06336, over 4859.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2518, pruned_loss=0.0593, over 953960.13 frames. ], batch size: 44, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:11:47,044 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.509e+02 1.942e+02 2.407e+02 6.603e+02, threshold=3.884e+02, percent-clipped=3.0 +2023-03-26 15:11:59,913 INFO [finetune.py:976] (0/7) Epoch 12, batch 5550, loss[loss=0.2518, simple_loss=0.3069, pruned_loss=0.09836, over 4903.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2539, pruned_loss=0.06041, over 953307.57 frames. ], batch size: 43, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:12:00,646 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4979, 1.3151, 1.3396, 1.3375, 1.7066, 1.6051, 1.3753, 1.2948], + device='cuda:0'), covar=tensor([0.0354, 0.0294, 0.0619, 0.0320, 0.0246, 0.0373, 0.0384, 0.0431], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0109, 0.0140, 0.0113, 0.0102, 0.0105, 0.0095, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2836e-05, 8.4379e-05, 1.1106e-04, 8.8444e-05, 7.9629e-05, 7.8120e-05, + 7.1965e-05, 8.3814e-05], device='cuda:0') +2023-03-26 15:12:23,720 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68578.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:12:39,652 INFO [finetune.py:976] (0/7) Epoch 12, batch 5600, loss[loss=0.2262, simple_loss=0.283, pruned_loss=0.08467, over 4936.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2576, pruned_loss=0.0622, over 952992.93 frames. ], batch size: 42, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:12:43,950 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 15:12:58,305 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68637.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:12:59,421 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.664e+02 1.965e+02 2.319e+02 3.885e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-26 15:12:59,532 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:09,174 INFO [finetune.py:976] (0/7) Epoch 12, batch 5650, loss[loss=0.232, simple_loss=0.2999, pruned_loss=0.08205, over 4915.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2605, pruned_loss=0.06308, over 953104.64 frames. ], batch size: 36, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:13:27,855 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:27,894 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:41,829 INFO [finetune.py:976] (0/7) Epoch 12, batch 5700, loss[loss=0.1752, simple_loss=0.2239, pruned_loss=0.06326, over 4406.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2581, pruned_loss=0.06336, over 938187.61 frames. ], batch size: 19, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:14:12,777 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-12.pt +2023-03-26 15:14:27,874 INFO [finetune.py:976] (0/7) Epoch 13, batch 0, loss[loss=0.2159, simple_loss=0.2843, pruned_loss=0.07381, over 4898.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2843, pruned_loss=0.07381, over 4898.00 frames. ], batch size: 36, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:14:27,875 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 15:14:42,136 INFO [finetune.py:1010] (0/7) Epoch 13, validation: loss=0.1598, simple_loss=0.23, pruned_loss=0.04482, over 2265189.00 frames. +2023-03-26 15:14:42,136 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 15:14:47,268 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.546e+02 1.915e+02 2.253e+02 4.332e+02, threshold=3.830e+02, percent-clipped=1.0 +2023-03-26 15:14:49,805 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68743.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:14:51,653 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68746.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:14:56,348 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6804, 1.6519, 1.6577, 0.8213, 1.7617, 1.9387, 1.9393, 1.5102], + device='cuda:0'), covar=tensor([0.0919, 0.0638, 0.0431, 0.0636, 0.0390, 0.0517, 0.0268, 0.0616], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0155, 0.0123, 0.0132, 0.0132, 0.0127, 0.0145, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.3373e-05, 1.1311e-04, 8.8848e-05, 9.5292e-05, 9.3380e-05, 9.2446e-05, + 1.0547e-04, 1.0665e-04], device='cuda:0') +2023-03-26 15:14:58,507 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68755.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:15:13,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8759, 1.7211, 1.9263, 1.0671, 1.8787, 1.9757, 1.7864, 1.5695], + device='cuda:0'), covar=tensor([0.0620, 0.0737, 0.0733, 0.1054, 0.0737, 0.0718, 0.0675, 0.1231], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0133, 0.0142, 0.0125, 0.0122, 0.0141, 0.0142, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:15:15,981 INFO [finetune.py:976] (0/7) Epoch 13, batch 50, loss[loss=0.1833, simple_loss=0.2417, pruned_loss=0.06244, over 4189.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2614, pruned_loss=0.06432, over 217360.64 frames. ], batch size: 66, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:15:21,852 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68791.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:15:22,505 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6859, 1.5208, 1.1303, 0.3193, 1.2986, 1.5152, 1.4188, 1.5620], + device='cuda:0'), covar=tensor([0.0911, 0.0747, 0.1221, 0.1911, 0.1317, 0.2068, 0.2251, 0.0742], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0195, 0.0196, 0.0184, 0.0211, 0.0205, 0.0220, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:15:27,189 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2153, 1.7875, 2.2594, 2.1039, 1.8157, 1.8550, 2.0308, 1.9724], + device='cuda:0'), covar=tensor([0.4207, 0.4997, 0.3722, 0.4533, 0.5721, 0.4476, 0.5644, 0.3774], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0237, 0.0254, 0.0260, 0.0256, 0.0232, 0.0273, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:15:50,384 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1657, 1.8972, 1.9997, 0.8357, 2.2302, 2.3428, 2.0972, 1.8733], + device='cuda:0'), covar=tensor([0.1040, 0.0833, 0.0680, 0.0861, 0.0645, 0.0777, 0.0628, 0.0816], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0154, 0.0123, 0.0132, 0.0131, 0.0127, 0.0145, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.3108e-05, 1.1261e-04, 8.8659e-05, 9.4959e-05, 9.3203e-05, 9.2186e-05, + 1.0545e-04, 1.0637e-04], device='cuda:0') +2023-03-26 15:15:57,664 INFO [finetune.py:976] (0/7) Epoch 13, batch 100, loss[loss=0.1634, simple_loss=0.2358, pruned_loss=0.0455, over 4783.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2546, pruned_loss=0.06128, over 381336.84 frames. ], batch size: 29, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:16:02,753 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.681e+02 1.901e+02 2.429e+02 4.753e+02, threshold=3.802e+02, percent-clipped=2.0 +2023-03-26 15:16:05,431 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 15:16:26,743 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4563, 3.8515, 4.0547, 4.2525, 4.1963, 3.9702, 4.5578, 1.3927], + device='cuda:0'), covar=tensor([0.0811, 0.0879, 0.0898, 0.0994, 0.1250, 0.1492, 0.0658, 0.5574], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0242, 0.0275, 0.0291, 0.0328, 0.0282, 0.0299, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:16:31,426 INFO [finetune.py:976] (0/7) Epoch 13, batch 150, loss[loss=0.1834, simple_loss=0.2484, pruned_loss=0.05916, over 4900.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2507, pruned_loss=0.06009, over 508860.64 frames. ], batch size: 32, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:16:59,136 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7086, 1.6451, 1.5671, 1.6900, 1.0648, 3.4074, 1.3172, 1.7790], + device='cuda:0'), covar=tensor([0.3269, 0.2410, 0.2119, 0.2405, 0.1972, 0.0203, 0.2562, 0.1317], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:17:05,107 INFO [finetune.py:976] (0/7) Epoch 13, batch 200, loss[loss=0.183, simple_loss=0.2451, pruned_loss=0.06051, over 4766.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2485, pruned_loss=0.05939, over 608282.05 frames. ], batch size: 28, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:17:05,767 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:17:09,211 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.222e+02 1.603e+02 1.930e+02 2.189e+02 8.191e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 15:17:46,315 INFO [finetune.py:976] (0/7) Epoch 13, batch 250, loss[loss=0.2153, simple_loss=0.2879, pruned_loss=0.07138, over 4866.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2534, pruned_loss=0.061, over 685122.31 frames. ], batch size: 34, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:18:05,424 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7040, 1.1199, 0.7557, 1.5985, 1.9688, 1.4498, 1.2883, 1.4806], + device='cuda:0'), covar=tensor([0.1936, 0.2922, 0.2566, 0.1585, 0.2516, 0.2488, 0.2155, 0.2806], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0091, 0.0119, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:18:19,709 INFO [finetune.py:976] (0/7) Epoch 13, batch 300, loss[loss=0.1909, simple_loss=0.2614, pruned_loss=0.06024, over 4792.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2578, pruned_loss=0.06295, over 747007.13 frames. ], batch size: 51, lr: 3.62e-03, grad_scale: 32.0 +2023-03-26 15:18:23,315 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.585e+02 1.877e+02 2.328e+02 4.201e+02, threshold=3.755e+02, percent-clipped=2.0 +2023-03-26 15:18:24,577 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:25,859 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69043.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:30,423 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6328, 1.5090, 1.4513, 1.5253, 1.8304, 1.7136, 1.5203, 1.3711], + device='cuda:0'), covar=tensor([0.0376, 0.0295, 0.0501, 0.0294, 0.0246, 0.0465, 0.0360, 0.0415], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0113, 0.0101, 0.0104, 0.0094, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2202e-05, 8.3653e-05, 1.1029e-04, 8.7830e-05, 7.8874e-05, 7.7218e-05, + 7.1315e-05, 8.3284e-05], device='cuda:0') +2023-03-26 15:18:34,523 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:55,352 INFO [finetune.py:976] (0/7) Epoch 13, batch 350, loss[loss=0.2091, simple_loss=0.2703, pruned_loss=0.07393, over 4837.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2602, pruned_loss=0.06387, over 792859.56 frames. ], batch size: 47, lr: 3.62e-03, grad_scale: 32.0 +2023-03-26 15:19:17,518 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 15:19:18,505 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:19:19,646 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69104.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:19:41,449 INFO [finetune.py:976] (0/7) Epoch 13, batch 400, loss[loss=0.1675, simple_loss=0.2486, pruned_loss=0.04318, over 4831.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2597, pruned_loss=0.06268, over 827208.89 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:19:50,066 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.689e+02 1.999e+02 2.345e+02 4.076e+02, threshold=3.998e+02, percent-clipped=3.0 +2023-03-26 15:20:00,985 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1575, 1.7361, 2.1545, 2.0224, 1.7975, 1.8661, 1.9779, 1.9312], + device='cuda:0'), covar=tensor([0.4387, 0.4858, 0.3873, 0.4758, 0.5831, 0.4158, 0.6352, 0.3794], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0239, 0.0255, 0.0261, 0.0258, 0.0233, 0.0275, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:20:09,732 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69162.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:20:13,410 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:20:23,367 INFO [finetune.py:976] (0/7) Epoch 13, batch 450, loss[loss=0.2272, simple_loss=0.2791, pruned_loss=0.08766, over 4825.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2585, pruned_loss=0.06245, over 855185.45 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:20:29,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8720, 1.8683, 1.5862, 1.9639, 2.3775, 1.9934, 1.6769, 1.5046], + device='cuda:0'), covar=tensor([0.2185, 0.1945, 0.1962, 0.1612, 0.1707, 0.1159, 0.2317, 0.1960], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0206, 0.0209, 0.0190, 0.0239, 0.0181, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:20:30,147 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 15:21:04,189 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69223.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:05,415 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:07,849 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:10,197 INFO [finetune.py:976] (0/7) Epoch 13, batch 500, loss[loss=0.1732, simple_loss=0.2371, pruned_loss=0.05459, over 4144.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2565, pruned_loss=0.06162, over 878094.93 frames. ], batch size: 18, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:21:10,902 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:14,297 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.659e+02 1.928e+02 2.205e+02 4.798e+02, threshold=3.855e+02, percent-clipped=1.0 +2023-03-26 15:21:17,693 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 15:21:18,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69245.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:21:21,802 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3012, 2.1258, 1.8359, 2.3018, 2.0337, 2.0565, 1.9918, 3.0279], + device='cuda:0'), covar=tensor([0.4436, 0.5709, 0.3881, 0.5251, 0.5155, 0.2995, 0.5101, 0.1798], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0223, 0.0275, 0.0244, 0.0211, 0.0245, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:21:37,037 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:43,320 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:43,877 INFO [finetune.py:976] (0/7) Epoch 13, batch 550, loss[loss=0.1542, simple_loss=0.2092, pruned_loss=0.04961, over 4693.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2529, pruned_loss=0.06039, over 893859.44 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:21:45,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8299, 3.3248, 3.5063, 3.6860, 3.6235, 3.3844, 3.8867, 1.2631], + device='cuda:0'), covar=tensor([0.0763, 0.0853, 0.0852, 0.0978, 0.1101, 0.1413, 0.0758, 0.4919], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0242, 0.0275, 0.0291, 0.0327, 0.0281, 0.0300, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:21:45,836 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69286.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:59,416 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69306.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 15:22:08,844 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:17,555 INFO [finetune.py:976] (0/7) Epoch 13, batch 600, loss[loss=0.1753, simple_loss=0.2507, pruned_loss=0.04996, over 4829.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2518, pruned_loss=0.05979, over 903623.07 frames. ], batch size: 30, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:22:18,290 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69334.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:22:21,204 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.536e+02 1.861e+02 2.296e+02 3.946e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 15:22:21,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2077, 2.2159, 1.8055, 1.1107, 2.0167, 1.9108, 1.7641, 2.0966], + device='cuda:0'), covar=tensor([0.0913, 0.0579, 0.1472, 0.1653, 0.1212, 0.1615, 0.1760, 0.0783], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0196, 0.0199, 0.0185, 0.0214, 0.0207, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:22:22,506 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69341.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:49,533 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69368.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:58,838 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:59,983 INFO [finetune.py:976] (0/7) Epoch 13, batch 650, loss[loss=0.1655, simple_loss=0.2461, pruned_loss=0.04246, over 4780.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2558, pruned_loss=0.06142, over 914757.35 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:23:03,693 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:06,763 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:10,344 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:30,675 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:33,430 INFO [finetune.py:976] (0/7) Epoch 13, batch 700, loss[loss=0.1747, simple_loss=0.2509, pruned_loss=0.04927, over 4781.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2589, pruned_loss=0.06262, over 924986.61 frames. ], batch size: 29, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:23:36,600 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 15:23:37,530 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.702e+02 1.957e+02 2.425e+02 4.096e+02, threshold=3.913e+02, percent-clipped=2.0 +2023-03-26 15:23:47,852 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:24:06,514 INFO [finetune.py:976] (0/7) Epoch 13, batch 750, loss[loss=0.2343, simple_loss=0.2995, pruned_loss=0.08456, over 4822.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2584, pruned_loss=0.06246, over 926392.81 frames. ], batch size: 39, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:24:40,551 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69518.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:24:44,562 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:24:50,486 INFO [finetune.py:976] (0/7) Epoch 13, batch 800, loss[loss=0.1883, simple_loss=0.2529, pruned_loss=0.06182, over 4738.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2573, pruned_loss=0.06168, over 932993.92 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:24:57,839 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.694e+02 1.982e+02 2.355e+02 4.334e+02, threshold=3.964e+02, percent-clipped=1.0 +2023-03-26 15:25:47,577 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69581.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:25:48,750 INFO [finetune.py:976] (0/7) Epoch 13, batch 850, loss[loss=0.1689, simple_loss=0.2328, pruned_loss=0.0525, over 4804.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2553, pruned_loss=0.0608, over 936618.37 frames. ], batch size: 25, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:26:03,113 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69601.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:26:21,344 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69629.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:26:24,227 INFO [finetune.py:976] (0/7) Epoch 13, batch 900, loss[loss=0.1886, simple_loss=0.254, pruned_loss=0.06157, over 4941.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2526, pruned_loss=0.05993, over 941190.11 frames. ], batch size: 33, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:26:27,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.604e+02 1.856e+02 2.224e+02 3.601e+02, threshold=3.711e+02, percent-clipped=0.0 +2023-03-26 15:26:37,709 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 15:26:55,513 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69674.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:26:56,707 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69676.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:27:06,462 INFO [finetune.py:976] (0/7) Epoch 13, batch 950, loss[loss=0.1847, simple_loss=0.2518, pruned_loss=0.05884, over 4802.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2516, pruned_loss=0.06017, over 942982.00 frames. ], batch size: 45, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:27:29,362 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:02,965 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69724.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:08,356 INFO [finetune.py:976] (0/7) Epoch 13, batch 1000, loss[loss=0.1559, simple_loss=0.2202, pruned_loss=0.04577, over 4046.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2537, pruned_loss=0.06121, over 945216.82 frames. ], batch size: 17, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:28:10,189 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69735.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 15:28:12,999 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.598e+02 1.856e+02 2.406e+02 4.029e+02, threshold=3.712e+02, percent-clipped=2.0 +2023-03-26 15:28:17,951 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:20,307 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69750.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:52,947 INFO [finetune.py:976] (0/7) Epoch 13, batch 1050, loss[loss=0.1956, simple_loss=0.2565, pruned_loss=0.06735, over 4905.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2567, pruned_loss=0.06164, over 945885.19 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:28:54,487 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 15:29:27,247 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5474, 2.0919, 1.8729, 0.9036, 2.0792, 1.9218, 1.4602, 1.9557], + device='cuda:0'), covar=tensor([0.0840, 0.1192, 0.1961, 0.2532, 0.1673, 0.2103, 0.2789, 0.1291], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0199, 0.0184, 0.0213, 0.0207, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:29:29,048 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5598, 1.3939, 1.3171, 1.5639, 1.8039, 1.6504, 1.5118, 1.2834], + device='cuda:0'), covar=tensor([0.0274, 0.0297, 0.0589, 0.0256, 0.0191, 0.0372, 0.0298, 0.0382], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0139, 0.0113, 0.0101, 0.0105, 0.0095, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2934e-05, 8.4197e-05, 1.1038e-04, 8.7667e-05, 7.8989e-05, 7.7664e-05, + 7.1472e-05, 8.2952e-05], device='cuda:0') +2023-03-26 15:29:38,102 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69818.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:29:48,094 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:29:59,171 INFO [finetune.py:976] (0/7) Epoch 13, batch 1100, loss[loss=0.2139, simple_loss=0.2785, pruned_loss=0.0746, over 4792.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2592, pruned_loss=0.06233, over 948570.28 frames. ], batch size: 29, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:29:59,381 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 15:30:01,779 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0668, 1.8101, 2.3835, 1.6391, 2.2251, 2.3970, 1.8185, 2.5380], + device='cuda:0'), covar=tensor([0.1476, 0.1963, 0.1465, 0.2099, 0.0883, 0.1444, 0.2545, 0.0898], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0207, 0.0196, 0.0193, 0.0179, 0.0216, 0.0219, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:30:02,885 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.609e+02 1.898e+02 2.282e+02 6.010e+02, threshold=3.795e+02, percent-clipped=2.0 +2023-03-26 15:30:35,884 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69866.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:43,313 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:51,891 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69881.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:53,490 INFO [finetune.py:976] (0/7) Epoch 13, batch 1150, loss[loss=0.1825, simple_loss=0.2522, pruned_loss=0.0564, over 4788.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2596, pruned_loss=0.06226, over 950160.90 frames. ], batch size: 29, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:31:12,311 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69901.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:31:42,205 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:31:42,256 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69929.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:31:44,609 INFO [finetune.py:976] (0/7) Epoch 13, batch 1200, loss[loss=0.1782, simple_loss=0.2507, pruned_loss=0.05287, over 4784.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2587, pruned_loss=0.06232, over 952408.86 frames. ], batch size: 26, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:31:48,756 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.603e+02 1.893e+02 2.321e+02 3.158e+02, threshold=3.786e+02, percent-clipped=0.0 +2023-03-26 15:31:55,829 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69949.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:32:13,198 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:13,757 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:17,837 INFO [finetune.py:976] (0/7) Epoch 13, batch 1250, loss[loss=0.1573, simple_loss=0.217, pruned_loss=0.04877, over 4236.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2568, pruned_loss=0.06169, over 952954.20 frames. ], batch size: 65, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:32:29,735 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-70000.pt +2023-03-26 15:32:45,359 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5499, 1.6984, 1.7542, 0.9144, 1.7761, 2.0221, 1.8952, 1.5206], + device='cuda:0'), covar=tensor([0.0896, 0.0707, 0.0495, 0.0595, 0.0443, 0.0640, 0.0327, 0.0715], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0120, 0.0128, 0.0129, 0.0125, 0.0141, 0.0143], + device='cuda:0'), out_proj_covar=tensor([9.1747e-05, 1.1010e-04, 8.6705e-05, 9.2450e-05, 9.1264e-05, 9.0504e-05, + 1.0303e-04, 1.0419e-04], device='cuda:0') +2023-03-26 15:32:46,510 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:46,536 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:47,238 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.52 vs. limit=5.0 +2023-03-26 15:32:50,161 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70030.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:32:52,397 INFO [finetune.py:976] (0/7) Epoch 13, batch 1300, loss[loss=0.1629, simple_loss=0.2222, pruned_loss=0.05183, over 4761.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2531, pruned_loss=0.06057, over 950367.50 frames. ], batch size: 59, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:32:52,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1896, 2.1397, 2.2425, 0.8583, 2.5067, 2.6073, 2.2702, 2.0810], + device='cuda:0'), covar=tensor([0.0936, 0.0718, 0.0507, 0.0757, 0.0487, 0.0592, 0.0416, 0.0732], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0120, 0.0128, 0.0129, 0.0125, 0.0141, 0.0143], + device='cuda:0'), out_proj_covar=tensor([9.1692e-05, 1.1003e-04, 8.6630e-05, 9.2386e-05, 9.1150e-05, 9.0448e-05, + 1.0294e-04, 1.0406e-04], device='cuda:0') +2023-03-26 15:32:56,055 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.649e+02 1.897e+02 2.309e+02 4.234e+02, threshold=3.795e+02, percent-clipped=2.0 +2023-03-26 15:33:03,827 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:06,688 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8687, 1.9889, 2.0293, 1.4277, 2.0363, 2.0727, 2.0153, 1.6795], + device='cuda:0'), covar=tensor([0.0574, 0.0593, 0.0609, 0.0860, 0.0682, 0.0607, 0.0570, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0130, 0.0140, 0.0123, 0.0122, 0.0140, 0.0140, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:33:10,859 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5732, 3.3350, 3.2290, 1.5783, 3.4547, 2.5786, 0.8699, 2.2234], + device='cuda:0'), covar=tensor([0.2403, 0.2402, 0.1532, 0.3392, 0.1223, 0.1125, 0.4230, 0.1790], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0175, 0.0160, 0.0129, 0.0156, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 15:33:16,384 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 15:33:19,135 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:25,839 INFO [finetune.py:976] (0/7) Epoch 13, batch 1350, loss[loss=0.1627, simple_loss=0.2332, pruned_loss=0.0461, over 4762.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2541, pruned_loss=0.06141, over 951891.00 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:33:36,432 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:53,146 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:34:08,077 INFO [finetune.py:976] (0/7) Epoch 13, batch 1400, loss[loss=0.2293, simple_loss=0.3039, pruned_loss=0.07734, over 4874.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2566, pruned_loss=0.06186, over 951610.41 frames. ], batch size: 34, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:34:12,153 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.588e+02 1.939e+02 2.393e+02 8.943e+02, threshold=3.877e+02, percent-clipped=1.0 +2023-03-26 15:34:31,286 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6913, 3.6466, 3.5440, 1.8677, 3.8594, 2.9160, 1.2836, 2.5125], + device='cuda:0'), covar=tensor([0.2262, 0.1719, 0.1359, 0.2929, 0.0992, 0.0898, 0.3626, 0.1371], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0175, 0.0160, 0.0129, 0.0157, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 15:34:34,206 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70171.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:34:41,771 INFO [finetune.py:976] (0/7) Epoch 13, batch 1450, loss[loss=0.2068, simple_loss=0.2865, pruned_loss=0.06352, over 4815.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2586, pruned_loss=0.06191, over 954424.52 frames. ], batch size: 47, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:34:47,781 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5560, 1.3889, 1.8571, 2.9740, 2.0472, 2.1757, 1.0140, 2.3814], + device='cuda:0'), covar=tensor([0.1872, 0.1576, 0.1254, 0.0606, 0.0868, 0.1334, 0.1841, 0.0591], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0166, 0.0102, 0.0139, 0.0128, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 15:35:00,656 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9150, 3.4251, 3.5925, 3.7812, 3.7016, 3.5166, 3.9736, 1.3052], + device='cuda:0'), covar=tensor([0.0901, 0.0923, 0.0970, 0.1104, 0.1449, 0.1604, 0.0862, 0.5307], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0244, 0.0278, 0.0293, 0.0331, 0.0283, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:35:05,324 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7802, 1.5713, 2.1578, 3.5757, 2.4180, 2.4370, 1.0956, 2.7221], + device='cuda:0'), covar=tensor([0.1802, 0.1513, 0.1306, 0.0560, 0.0805, 0.1297, 0.1926, 0.0581], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0166, 0.0102, 0.0139, 0.0128, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 15:35:16,323 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5120, 1.5083, 1.6279, 1.7949, 1.6154, 2.6933, 1.3859, 1.5798], + device='cuda:0'), covar=tensor([0.0829, 0.1409, 0.1239, 0.0752, 0.1304, 0.0320, 0.1241, 0.1401], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:35:26,432 INFO [finetune.py:976] (0/7) Epoch 13, batch 1500, loss[loss=0.2318, simple_loss=0.2894, pruned_loss=0.08704, over 4864.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2598, pruned_loss=0.06254, over 953680.22 frames. ], batch size: 31, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:35:30,134 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.613e+02 1.899e+02 2.364e+02 4.350e+02, threshold=3.798e+02, percent-clipped=1.0 +2023-03-26 15:35:46,865 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:36:10,535 INFO [finetune.py:976] (0/7) Epoch 13, batch 1550, loss[loss=0.1608, simple_loss=0.2388, pruned_loss=0.04134, over 4870.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2595, pruned_loss=0.06147, over 956239.21 frames. ], batch size: 32, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:36:14,341 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7786, 1.8215, 1.6426, 1.9608, 2.1548, 1.9873, 1.5804, 1.4515], + device='cuda:0'), covar=tensor([0.2250, 0.2051, 0.1936, 0.1606, 0.1857, 0.1202, 0.2447, 0.2073], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0206, 0.0209, 0.0189, 0.0239, 0.0182, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:36:42,060 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2916, 1.8789, 2.2964, 2.1801, 1.9810, 1.9821, 2.1497, 2.0990], + device='cuda:0'), covar=tensor([0.4062, 0.4392, 0.3508, 0.4192, 0.5105, 0.4061, 0.5043, 0.3476], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0238, 0.0255, 0.0262, 0.0259, 0.0234, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:36:49,635 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:36:58,940 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70330.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:37:00,630 INFO [finetune.py:976] (0/7) Epoch 13, batch 1600, loss[loss=0.1925, simple_loss=0.2488, pruned_loss=0.06809, over 4909.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2567, pruned_loss=0.06052, over 956189.67 frames. ], batch size: 36, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:37:04,736 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.529e+02 1.873e+02 2.318e+02 5.550e+02, threshold=3.745e+02, percent-clipped=4.0 +2023-03-26 15:37:30,801 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70378.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:37:34,183 INFO [finetune.py:976] (0/7) Epoch 13, batch 1650, loss[loss=0.1971, simple_loss=0.2564, pruned_loss=0.06887, over 4833.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2543, pruned_loss=0.05989, over 956548.75 frames. ], batch size: 33, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:37:59,997 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 15:38:08,087 INFO [finetune.py:976] (0/7) Epoch 13, batch 1700, loss[loss=0.1785, simple_loss=0.2305, pruned_loss=0.0632, over 4326.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2525, pruned_loss=0.0597, over 956916.35 frames. ], batch size: 65, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:38:11,732 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.610e+02 1.926e+02 2.276e+02 4.227e+02, threshold=3.852e+02, percent-clipped=1.0 +2023-03-26 15:38:16,643 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 15:38:30,202 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:38:32,137 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9554, 1.7812, 1.5868, 1.6998, 1.7129, 1.6628, 1.6987, 2.4395], + device='cuda:0'), covar=tensor([0.4275, 0.4627, 0.3370, 0.4145, 0.4127, 0.2514, 0.4055, 0.1800], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0259, 0.0224, 0.0276, 0.0246, 0.0212, 0.0247, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:38:41,456 INFO [finetune.py:976] (0/7) Epoch 13, batch 1750, loss[loss=0.1873, simple_loss=0.2442, pruned_loss=0.06516, over 4126.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2544, pruned_loss=0.06048, over 956783.80 frames. ], batch size: 65, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:39:24,240 INFO [finetune.py:976] (0/7) Epoch 13, batch 1800, loss[loss=0.1405, simple_loss=0.2119, pruned_loss=0.03456, over 4762.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2575, pruned_loss=0.06083, over 956874.49 frames. ], batch size: 27, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:39:28,350 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.597e+02 2.051e+02 2.548e+02 3.844e+02, threshold=4.101e+02, percent-clipped=0.0 +2023-03-26 15:39:38,093 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2788, 3.7014, 3.9074, 4.0882, 4.0103, 3.8239, 4.3794, 1.3396], + device='cuda:0'), covar=tensor([0.0824, 0.0828, 0.0782, 0.1013, 0.1265, 0.1551, 0.0669, 0.5753], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0246, 0.0279, 0.0294, 0.0333, 0.0284, 0.0305, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:39:58,058 INFO [finetune.py:976] (0/7) Epoch 13, batch 1850, loss[loss=0.2145, simple_loss=0.2789, pruned_loss=0.07506, over 4903.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2595, pruned_loss=0.06175, over 957218.52 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:40:26,915 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:40:42,677 INFO [finetune.py:976] (0/7) Epoch 13, batch 1900, loss[loss=0.192, simple_loss=0.2708, pruned_loss=0.05659, over 4884.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2601, pruned_loss=0.06219, over 957607.91 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:40:43,926 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9693, 1.9539, 1.6417, 1.7818, 1.8038, 1.7420, 1.8017, 2.5297], + device='cuda:0'), covar=tensor([0.4408, 0.4624, 0.3490, 0.4349, 0.4236, 0.2620, 0.3952, 0.1729], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0224, 0.0277, 0.0245, 0.0212, 0.0247, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:40:46,778 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.570e+02 1.884e+02 2.217e+02 6.026e+02, threshold=3.769e+02, percent-clipped=2.0 +2023-03-26 15:41:27,334 INFO [finetune.py:976] (0/7) Epoch 13, batch 1950, loss[loss=0.1745, simple_loss=0.253, pruned_loss=0.04801, over 4922.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.26, pruned_loss=0.06252, over 956665.71 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:41:28,077 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0612, 1.9315, 1.7083, 1.9409, 1.8715, 1.7978, 1.8033, 2.6113], + device='cuda:0'), covar=tensor([0.3978, 0.4460, 0.3285, 0.4014, 0.4071, 0.2463, 0.4112, 0.1686], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0224, 0.0277, 0.0245, 0.0211, 0.0247, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:42:06,899 INFO [finetune.py:976] (0/7) Epoch 13, batch 2000, loss[loss=0.1594, simple_loss=0.2271, pruned_loss=0.04589, over 4816.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.257, pruned_loss=0.06196, over 956279.58 frames. ], batch size: 39, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:42:15,810 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.535e+02 1.807e+02 2.194e+02 3.140e+02, threshold=3.615e+02, percent-clipped=0.0 +2023-03-26 15:42:29,831 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7991, 1.3064, 1.9299, 1.7667, 1.6080, 1.5249, 1.7319, 1.6930], + device='cuda:0'), covar=tensor([0.4256, 0.4290, 0.3495, 0.4035, 0.5130, 0.3839, 0.4650, 0.3473], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0239, 0.0257, 0.0263, 0.0261, 0.0235, 0.0277, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:42:36,892 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:42:48,489 INFO [finetune.py:976] (0/7) Epoch 13, batch 2050, loss[loss=0.165, simple_loss=0.2359, pruned_loss=0.04702, over 4822.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2545, pruned_loss=0.06138, over 957525.05 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:43:09,365 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70814.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:43:22,315 INFO [finetune.py:976] (0/7) Epoch 13, batch 2100, loss[loss=0.2771, simple_loss=0.3382, pruned_loss=0.108, over 4856.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2538, pruned_loss=0.06114, over 956511.52 frames. ], batch size: 44, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:43:26,466 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.827e+01 1.609e+02 1.892e+02 2.240e+02 3.187e+02, threshold=3.783e+02, percent-clipped=0.0 +2023-03-26 15:43:56,100 INFO [finetune.py:976] (0/7) Epoch 13, batch 2150, loss[loss=0.2374, simple_loss=0.3022, pruned_loss=0.08631, over 4865.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2571, pruned_loss=0.06231, over 956966.34 frames. ], batch size: 31, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:44:05,275 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 15:44:35,039 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:44:46,796 INFO [finetune.py:976] (0/7) Epoch 13, batch 2200, loss[loss=0.1869, simple_loss=0.2479, pruned_loss=0.06298, over 4922.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2585, pruned_loss=0.06211, over 956559.19 frames. ], batch size: 42, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:44:50,484 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.701e+02 1.958e+02 2.316e+02 4.574e+02, threshold=3.916e+02, percent-clipped=1.0 +2023-03-26 15:44:53,605 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 15:45:01,194 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7679, 1.7165, 1.4905, 1.8654, 2.2493, 1.8974, 1.4869, 1.4203], + device='cuda:0'), covar=tensor([0.2202, 0.2057, 0.2037, 0.1707, 0.1642, 0.1158, 0.2394, 0.1969], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0205, 0.0207, 0.0188, 0.0237, 0.0180, 0.0211, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:45:07,657 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:45:19,189 INFO [finetune.py:976] (0/7) Epoch 13, batch 2250, loss[loss=0.2404, simple_loss=0.3063, pruned_loss=0.08723, over 4890.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2603, pruned_loss=0.0626, over 956145.37 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:45:26,134 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:46:02,058 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:46:03,707 INFO [finetune.py:976] (0/7) Epoch 13, batch 2300, loss[loss=0.2371, simple_loss=0.2974, pruned_loss=0.0884, over 4849.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2615, pruned_loss=0.06305, over 956917.05 frames. ], batch size: 44, lr: 3.60e-03, grad_scale: 64.0 +2023-03-26 15:46:08,250 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.237e+01 1.685e+02 2.000e+02 2.324e+02 3.629e+02, threshold=3.999e+02, percent-clipped=0.0 +2023-03-26 15:46:23,767 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71053.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:46:47,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4352, 2.3643, 2.1993, 2.5355, 1.9928, 4.9685, 2.1787, 2.7333], + device='cuda:0'), covar=tensor([0.2835, 0.2142, 0.1859, 0.2002, 0.1483, 0.0100, 0.2066, 0.1108], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:46:47,504 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.11 vs. limit=5.0 +2023-03-26 15:46:59,593 INFO [finetune.py:976] (0/7) Epoch 13, batch 2350, loss[loss=0.1648, simple_loss=0.2233, pruned_loss=0.05313, over 4028.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2586, pruned_loss=0.0621, over 954199.36 frames. ], batch size: 17, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:47:10,246 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71091.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:47:39,532 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3110, 2.6666, 2.2714, 1.7378, 2.5837, 2.7355, 2.6317, 2.3999], + device='cuda:0'), covar=tensor([0.0719, 0.0634, 0.0919, 0.0964, 0.0805, 0.0750, 0.0692, 0.0964], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0132, 0.0142, 0.0124, 0.0124, 0.0142, 0.0141, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:47:39,639 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 15:48:00,798 INFO [finetune.py:976] (0/7) Epoch 13, batch 2400, loss[loss=0.1701, simple_loss=0.2319, pruned_loss=0.05415, over 4823.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.256, pruned_loss=0.06113, over 952229.29 frames. ], batch size: 39, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:48:06,380 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8102, 2.4817, 2.0915, 1.1237, 2.3163, 2.1248, 1.9123, 2.1270], + device='cuda:0'), covar=tensor([0.0747, 0.0789, 0.1427, 0.1918, 0.1360, 0.1999, 0.1963, 0.0988], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0199, 0.0185, 0.0213, 0.0207, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:48:09,284 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.076e+01 1.502e+02 1.791e+02 2.104e+02 3.987e+02, threshold=3.583e+02, percent-clipped=0.0 +2023-03-26 15:48:28,463 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3204, 2.3695, 2.3741, 1.8077, 2.2336, 2.5034, 2.3878, 2.1312], + device='cuda:0'), covar=tensor([0.0527, 0.0447, 0.0534, 0.0736, 0.0892, 0.0512, 0.0486, 0.0741], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0132, 0.0143, 0.0125, 0.0124, 0.0143, 0.0142, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:49:05,621 INFO [finetune.py:976] (0/7) Epoch 13, batch 2450, loss[loss=0.1834, simple_loss=0.2511, pruned_loss=0.05786, over 4761.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2524, pruned_loss=0.05975, over 951735.14 frames. ], batch size: 54, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:49:13,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9373, 2.4105, 2.0423, 1.1364, 2.3025, 2.1715, 1.6839, 2.1371], + device='cuda:0'), covar=tensor([0.0815, 0.0975, 0.1697, 0.2109, 0.1384, 0.1916, 0.2518, 0.1140], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0195, 0.0198, 0.0184, 0.0212, 0.0206, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:49:23,585 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1743, 2.1400, 1.8218, 2.2883, 2.0168, 2.0312, 1.9793, 2.9951], + device='cuda:0'), covar=tensor([0.4468, 0.5459, 0.3883, 0.4571, 0.5129, 0.2708, 0.5369, 0.1744], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0225, 0.0279, 0.0247, 0.0213, 0.0248, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:49:31,856 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0482, 1.7762, 2.0672, 1.9574, 1.7357, 1.7273, 1.9882, 1.8952], + device='cuda:0'), covar=tensor([0.4083, 0.4596, 0.3450, 0.4623, 0.5311, 0.4114, 0.4999, 0.3298], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0240, 0.0257, 0.0264, 0.0262, 0.0236, 0.0277, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:50:04,541 INFO [finetune.py:976] (0/7) Epoch 13, batch 2500, loss[loss=0.1785, simple_loss=0.2405, pruned_loss=0.05822, over 4731.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2522, pruned_loss=0.05942, over 951552.53 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:50:08,815 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.629e+02 1.890e+02 2.415e+02 4.682e+02, threshold=3.780e+02, percent-clipped=4.0 +2023-03-26 15:50:32,592 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5470, 1.4807, 2.0476, 3.2704, 2.2818, 2.2755, 1.2533, 2.5289], + device='cuda:0'), covar=tensor([0.1711, 0.1437, 0.1268, 0.0552, 0.0746, 0.1531, 0.1663, 0.0522], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0115, 0.0133, 0.0164, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:50:36,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 15:50:41,417 INFO [finetune.py:976] (0/7) Epoch 13, batch 2550, loss[loss=0.175, simple_loss=0.2551, pruned_loss=0.04746, over 4901.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2567, pruned_loss=0.061, over 952565.67 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:50:43,487 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 15:51:22,574 INFO [finetune.py:976] (0/7) Epoch 13, batch 2600, loss[loss=0.2132, simple_loss=0.265, pruned_loss=0.08075, over 4776.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2582, pruned_loss=0.06132, over 951345.74 frames. ], batch size: 28, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:51:26,871 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.678e+02 1.922e+02 2.428e+02 5.321e+02, threshold=3.843e+02, percent-clipped=3.0 +2023-03-26 15:51:31,778 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:51:42,200 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7097, 1.5518, 1.5797, 1.6362, 1.1006, 3.3520, 1.2501, 1.8614], + device='cuda:0'), covar=tensor([0.3522, 0.2593, 0.2204, 0.2548, 0.2029, 0.0212, 0.2744, 0.1279], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0099, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:51:42,245 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1403, 1.8962, 1.6632, 1.7419, 1.7702, 1.8354, 1.8242, 2.6195], + device='cuda:0'), covar=tensor([0.4192, 0.5065, 0.3773, 0.4452, 0.4635, 0.2606, 0.4501, 0.1781], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0259, 0.0223, 0.0278, 0.0246, 0.0212, 0.0247, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:51:52,569 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-03-26 15:51:55,370 INFO [finetune.py:976] (0/7) Epoch 13, batch 2650, loss[loss=0.2172, simple_loss=0.2886, pruned_loss=0.07292, over 4788.00 frames. ], tot_loss[loss=0.193, simple_loss=0.261, pruned_loss=0.06248, over 952170.20 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:51:58,296 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:52:05,699 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1725, 1.9417, 2.7551, 4.2221, 3.1094, 2.8670, 1.3508, 3.4557], + device='cuda:0'), covar=tensor([0.1771, 0.1479, 0.1253, 0.0606, 0.0667, 0.1578, 0.1775, 0.0439], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:52:29,325 INFO [finetune.py:976] (0/7) Epoch 13, batch 2700, loss[loss=0.1783, simple_loss=0.2551, pruned_loss=0.05074, over 4822.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2603, pruned_loss=0.06261, over 953772.10 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:52:33,597 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 15:52:34,537 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.578e+02 1.884e+02 2.307e+02 4.300e+02, threshold=3.769e+02, percent-clipped=2.0 +2023-03-26 15:53:02,923 INFO [finetune.py:976] (0/7) Epoch 13, batch 2750, loss[loss=0.2012, simple_loss=0.261, pruned_loss=0.07075, over 4911.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2564, pruned_loss=0.06132, over 953109.18 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:53:36,642 INFO [finetune.py:976] (0/7) Epoch 13, batch 2800, loss[loss=0.205, simple_loss=0.2697, pruned_loss=0.07013, over 4909.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2536, pruned_loss=0.06044, over 953197.33 frames. ], batch size: 32, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:53:40,882 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.564e+02 1.863e+02 2.304e+02 3.302e+02, threshold=3.726e+02, percent-clipped=0.0 +2023-03-26 15:53:50,402 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5525, 1.5189, 2.0295, 3.0319, 2.0821, 2.2991, 0.9440, 2.4492], + device='cuda:0'), covar=tensor([0.1620, 0.1357, 0.1131, 0.0568, 0.0833, 0.1482, 0.1778, 0.0573], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 15:54:23,057 INFO [finetune.py:976] (0/7) Epoch 13, batch 2850, loss[loss=0.1741, simple_loss=0.2307, pruned_loss=0.05877, over 4716.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.252, pruned_loss=0.05965, over 953615.59 frames. ], batch size: 23, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:54:42,096 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6998, 1.5039, 1.4371, 1.7360, 1.9873, 1.7562, 1.4164, 1.4463], + device='cuda:0'), covar=tensor([0.2023, 0.2089, 0.1779, 0.1612, 0.1564, 0.1166, 0.2274, 0.1764], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0206, 0.0209, 0.0189, 0.0239, 0.0182, 0.0213, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:54:52,349 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:55:05,836 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6371, 1.5709, 1.5900, 1.6217, 1.0497, 3.3256, 1.2415, 1.6315], + device='cuda:0'), covar=tensor([0.3388, 0.2572, 0.2093, 0.2396, 0.1982, 0.0215, 0.2716, 0.1341], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0116, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:55:06,323 INFO [finetune.py:976] (0/7) Epoch 13, batch 2900, loss[loss=0.1839, simple_loss=0.2568, pruned_loss=0.05553, over 4889.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2549, pruned_loss=0.06105, over 952197.83 frames. ], batch size: 32, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:55:13,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8111, 1.6893, 1.4432, 1.4166, 1.7769, 1.5427, 1.8696, 1.7946], + device='cuda:0'), covar=tensor([0.1459, 0.2123, 0.3184, 0.2649, 0.3012, 0.1802, 0.3041, 0.1920], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0188, 0.0234, 0.0255, 0.0245, 0.0199, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:55:15,492 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.635e+01 1.661e+02 1.944e+02 2.530e+02 6.475e+02, threshold=3.888e+02, percent-clipped=5.0 +2023-03-26 15:55:24,624 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71648.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:55:49,134 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:55:58,581 INFO [finetune.py:976] (0/7) Epoch 13, batch 2950, loss[loss=0.2209, simple_loss=0.2994, pruned_loss=0.07121, over 4901.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2599, pruned_loss=0.06277, over 953837.74 frames. ], batch size: 43, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:56:00,470 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:11,118 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:36,990 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1525, 2.0012, 1.6306, 1.9460, 2.0564, 1.8299, 2.3341, 2.1634], + device='cuda:0'), covar=tensor([0.1253, 0.1985, 0.2795, 0.2611, 0.2385, 0.1503, 0.2919, 0.1637], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0188, 0.0234, 0.0254, 0.0244, 0.0199, 0.0213, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 15:56:39,795 INFO [finetune.py:976] (0/7) Epoch 13, batch 3000, loss[loss=0.2264, simple_loss=0.2873, pruned_loss=0.08271, over 4910.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2606, pruned_loss=0.06296, over 955040.53 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:56:39,796 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 15:56:50,411 INFO [finetune.py:1010] (0/7) Epoch 13, validation: loss=0.1572, simple_loss=0.2278, pruned_loss=0.04333, over 2265189.00 frames. +2023-03-26 15:56:50,412 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 15:56:51,090 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71734.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:51,164 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5813, 1.6771, 1.6319, 1.0425, 1.6746, 1.9309, 1.8500, 1.4555], + device='cuda:0'), covar=tensor([0.0983, 0.0632, 0.0457, 0.0522, 0.0425, 0.0540, 0.0326, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0121, 0.0129, 0.0130, 0.0126, 0.0142, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.2757e-05, 1.1073e-04, 8.7206e-05, 9.2573e-05, 9.1978e-05, 9.1617e-05, + 1.0302e-04, 1.0493e-04], device='cuda:0') +2023-03-26 15:56:55,665 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.624e+02 1.953e+02 2.376e+02 4.887e+02, threshold=3.907e+02, percent-clipped=1.0 +2023-03-26 15:57:22,729 INFO [finetune.py:976] (0/7) Epoch 13, batch 3050, loss[loss=0.2113, simple_loss=0.2685, pruned_loss=0.07701, over 4895.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2602, pruned_loss=0.06257, over 954167.55 frames. ], batch size: 32, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:57:40,675 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1040, 4.1276, 3.9701, 2.3128, 4.1855, 3.1523, 0.7949, 3.2281], + device='cuda:0'), covar=tensor([0.1895, 0.1752, 0.1322, 0.2966, 0.0990, 0.1062, 0.4779, 0.1407], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0129, 0.0157, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 15:57:55,478 INFO [finetune.py:976] (0/7) Epoch 13, batch 3100, loss[loss=0.177, simple_loss=0.2449, pruned_loss=0.05459, over 4894.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2581, pruned_loss=0.06168, over 953654.42 frames. ], batch size: 43, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:58:01,083 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.560e+02 1.843e+02 2.215e+02 5.565e+02, threshold=3.687e+02, percent-clipped=1.0 +2023-03-26 15:58:21,871 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8247, 1.8286, 1.6443, 1.8905, 1.4789, 4.3099, 1.6883, 2.1247], + device='cuda:0'), covar=tensor([0.3353, 0.2402, 0.2103, 0.2298, 0.1681, 0.0138, 0.2402, 0.1201], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 15:58:26,283 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 15:58:29,153 INFO [finetune.py:976] (0/7) Epoch 13, batch 3150, loss[loss=0.1356, simple_loss=0.2062, pruned_loss=0.03249, over 4814.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2548, pruned_loss=0.06057, over 953594.28 frames. ], batch size: 51, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:58:53,738 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71919.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:03,052 INFO [finetune.py:976] (0/7) Epoch 13, batch 3200, loss[loss=0.1496, simple_loss=0.2198, pruned_loss=0.03971, over 4767.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2518, pruned_loss=0.06006, over 954932.97 frames. ], batch size: 26, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:59:07,311 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.561e+02 1.912e+02 2.265e+02 3.518e+02, threshold=3.824e+02, percent-clipped=0.0 +2023-03-26 15:59:36,749 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 15:59:38,342 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7852, 3.5868, 3.4296, 1.6817, 3.7029, 2.7722, 0.7518, 2.4555], + device='cuda:0'), covar=tensor([0.2384, 0.2114, 0.1513, 0.3470, 0.1035, 0.1040, 0.4850, 0.1530], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0158, 0.0129, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 15:59:40,761 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:54,359 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:56,074 INFO [finetune.py:976] (0/7) Epoch 13, batch 3250, loss[loss=0.1834, simple_loss=0.2502, pruned_loss=0.05832, over 4764.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.254, pruned_loss=0.06054, over 955923.13 frames. ], batch size: 26, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 16:00:07,026 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-72000.pt +2023-03-26 16:00:18,476 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 16:00:22,751 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-26 16:00:39,651 INFO [finetune.py:976] (0/7) Epoch 13, batch 3300, loss[loss=0.1825, simple_loss=0.2588, pruned_loss=0.05313, over 4906.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2586, pruned_loss=0.06192, over 957405.25 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:00:44,477 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.593e+02 1.995e+02 2.341e+02 5.205e+02, threshold=3.991e+02, percent-clipped=4.0 +2023-03-26 16:01:02,421 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9316, 1.4624, 1.1073, 1.7788, 2.0758, 1.5979, 1.6430, 1.7410], + device='cuda:0'), covar=tensor([0.1265, 0.1794, 0.1875, 0.1051, 0.1845, 0.2056, 0.1278, 0.1647], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0120, 0.0094, 0.0099, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:01:29,182 INFO [finetune.py:976] (0/7) Epoch 13, batch 3350, loss[loss=0.1945, simple_loss=0.2629, pruned_loss=0.06301, over 4835.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2602, pruned_loss=0.06247, over 955819.33 frames. ], batch size: 49, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:01:29,322 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4419, 2.4974, 2.6260, 1.1338, 2.9601, 3.1557, 2.6198, 2.2914], + device='cuda:0'), covar=tensor([0.0998, 0.0898, 0.0482, 0.0861, 0.0572, 0.0578, 0.0448, 0.0867], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0123, 0.0129, 0.0131, 0.0127, 0.0143, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.3715e-05, 1.1160e-04, 8.8271e-05, 9.3198e-05, 9.2648e-05, 9.2259e-05, + 1.0411e-04, 1.0582e-04], device='cuda:0') +2023-03-26 16:01:45,249 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4852, 2.3205, 1.9171, 0.9824, 2.0727, 1.9436, 1.7880, 2.0458], + device='cuda:0'), covar=tensor([0.1012, 0.0798, 0.1691, 0.2045, 0.1454, 0.2383, 0.2161, 0.1025], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0196, 0.0200, 0.0185, 0.0213, 0.0207, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:02:00,575 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7301, 1.4496, 0.9464, 0.2388, 1.1657, 1.4627, 1.3013, 1.3336], + device='cuda:0'), covar=tensor([0.0829, 0.1045, 0.1507, 0.2249, 0.1533, 0.2457, 0.2701, 0.1032], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0196, 0.0199, 0.0184, 0.0212, 0.0206, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:02:07,943 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:02:11,041 INFO [finetune.py:976] (0/7) Epoch 13, batch 3400, loss[loss=0.2258, simple_loss=0.2837, pruned_loss=0.08393, over 4826.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2618, pruned_loss=0.06352, over 951359.87 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:02:16,764 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.708e+02 2.008e+02 2.371e+02 4.954e+02, threshold=4.015e+02, percent-clipped=4.0 +2023-03-26 16:02:26,716 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4891, 2.3219, 1.8399, 2.4812, 2.4252, 2.0275, 2.9180, 2.4021], + device='cuda:0'), covar=tensor([0.1350, 0.2474, 0.3282, 0.2917, 0.2701, 0.1767, 0.3166, 0.1796], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0188, 0.0235, 0.0256, 0.0246, 0.0200, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:02:49,868 INFO [finetune.py:976] (0/7) Epoch 13, batch 3450, loss[loss=0.1724, simple_loss=0.2462, pruned_loss=0.04932, over 4819.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2606, pruned_loss=0.06263, over 951014.74 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:02:55,186 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:01,406 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-03-26 16:03:01,940 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6261, 1.4392, 1.4620, 1.5588, 1.9077, 1.7405, 1.5924, 1.3755], + device='cuda:0'), covar=tensor([0.0322, 0.0354, 0.0502, 0.0295, 0.0199, 0.0449, 0.0312, 0.0404], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0110, 0.0140, 0.0113, 0.0101, 0.0105, 0.0095, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2818e-05, 8.4934e-05, 1.1119e-04, 8.7724e-05, 7.9167e-05, 7.7898e-05, + 7.2153e-05, 8.3588e-05], device='cuda:0') +2023-03-26 16:03:23,344 INFO [finetune.py:976] (0/7) Epoch 13, batch 3500, loss[loss=0.1459, simple_loss=0.2129, pruned_loss=0.03942, over 4788.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2564, pruned_loss=0.06119, over 952661.89 frames. ], batch size: 45, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:03:29,064 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.641e+02 1.993e+02 2.438e+02 4.377e+02, threshold=3.986e+02, percent-clipped=2.0 +2023-03-26 16:03:49,432 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 16:03:49,884 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:51,644 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72275.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:54,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4007, 1.6734, 2.0132, 1.7840, 1.7211, 3.9030, 1.4444, 1.8066], + device='cuda:0'), covar=tensor([0.0942, 0.1636, 0.1254, 0.0925, 0.1414, 0.0189, 0.1392, 0.1584], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:03:56,424 INFO [finetune.py:976] (0/7) Epoch 13, batch 3550, loss[loss=0.1618, simple_loss=0.2317, pruned_loss=0.04594, over 4930.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2534, pruned_loss=0.06062, over 952473.76 frames. ], batch size: 38, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:04:37,444 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:04:47,534 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:04:51,694 INFO [finetune.py:976] (0/7) Epoch 13, batch 3600, loss[loss=0.2202, simple_loss=0.2819, pruned_loss=0.07923, over 4904.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2515, pruned_loss=0.05969, over 954327.04 frames. ], batch size: 37, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:04:55,485 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 16:04:58,281 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.845e+01 1.525e+02 1.754e+02 2.048e+02 3.586e+02, threshold=3.507e+02, percent-clipped=0.0 +2023-03-26 16:05:42,445 INFO [finetune.py:976] (0/7) Epoch 13, batch 3650, loss[loss=0.1994, simple_loss=0.2743, pruned_loss=0.06221, over 4847.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2551, pruned_loss=0.06124, over 955161.80 frames. ], batch size: 49, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:05:50,452 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72387.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:05:51,783 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-26 16:06:21,557 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6631, 1.4123, 1.8647, 1.2588, 1.7380, 1.8152, 1.4103, 2.0156], + device='cuda:0'), covar=tensor([0.1220, 0.2141, 0.1351, 0.1816, 0.0911, 0.1364, 0.2897, 0.0828], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0204, 0.0192, 0.0190, 0.0177, 0.0212, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:06:54,367 INFO [finetune.py:976] (0/7) Epoch 13, batch 3700, loss[loss=0.1699, simple_loss=0.2533, pruned_loss=0.04326, over 4781.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2579, pruned_loss=0.06152, over 955448.16 frames. ], batch size: 26, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:07:04,384 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.616e+02 1.915e+02 2.308e+02 4.437e+02, threshold=3.829e+02, percent-clipped=1.0 +2023-03-26 16:07:52,820 INFO [finetune.py:976] (0/7) Epoch 13, batch 3750, loss[loss=0.2235, simple_loss=0.2973, pruned_loss=0.07487, over 4740.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2592, pruned_loss=0.06226, over 954838.02 frames. ], batch size: 54, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:07:54,132 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72485.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:08:13,374 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4918, 2.1500, 2.7508, 1.6227, 2.4549, 2.6272, 1.9400, 2.8471], + device='cuda:0'), covar=tensor([0.1407, 0.1996, 0.1767, 0.2495, 0.0968, 0.1722, 0.2633, 0.0911], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0203, 0.0192, 0.0190, 0.0177, 0.0212, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:08:29,282 INFO [finetune.py:976] (0/7) Epoch 13, batch 3800, loss[loss=0.1646, simple_loss=0.2368, pruned_loss=0.04622, over 4831.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2607, pruned_loss=0.06248, over 955878.14 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:08:34,664 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.445e+01 1.578e+02 1.803e+02 2.155e+02 3.901e+02, threshold=3.607e+02, percent-clipped=1.0 +2023-03-26 16:08:55,206 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 16:08:56,807 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72575.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:09:02,584 INFO [finetune.py:976] (0/7) Epoch 13, batch 3850, loss[loss=0.2101, simple_loss=0.275, pruned_loss=0.07258, over 4839.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.259, pruned_loss=0.06169, over 955112.63 frames. ], batch size: 49, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:09:30,586 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72623.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:09:36,081 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2262, 2.7578, 2.5783, 1.3019, 2.7032, 2.3003, 2.0494, 2.4853], + device='cuda:0'), covar=tensor([0.0925, 0.0924, 0.1922, 0.2461, 0.1659, 0.1981, 0.2270, 0.1278], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0197, 0.0201, 0.0186, 0.0213, 0.0208, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:09:36,093 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2201, 2.0795, 2.1870, 0.9202, 2.3880, 2.5141, 2.2128, 1.9367], + device='cuda:0'), covar=tensor([0.0927, 0.0671, 0.0454, 0.0701, 0.0440, 0.0547, 0.0473, 0.0630], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0153, 0.0122, 0.0129, 0.0131, 0.0126, 0.0142, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.3367e-05, 1.1136e-04, 8.8047e-05, 9.2905e-05, 9.3056e-05, 9.1281e-05, + 1.0363e-04, 1.0550e-04], device='cuda:0') +2023-03-26 16:09:46,306 INFO [finetune.py:976] (0/7) Epoch 13, batch 3900, loss[loss=0.1573, simple_loss=0.2191, pruned_loss=0.04772, over 4818.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2567, pruned_loss=0.06124, over 956306.14 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:09:51,186 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.402e+01 1.493e+02 1.751e+02 2.217e+02 3.590e+02, threshold=3.501e+02, percent-clipped=0.0 +2023-03-26 16:10:10,399 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0258, 2.7055, 3.1634, 2.2267, 2.8546, 3.2152, 2.6981, 3.3518], + device='cuda:0'), covar=tensor([0.1168, 0.1596, 0.1466, 0.1957, 0.0865, 0.1323, 0.1991, 0.0734], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0204, 0.0193, 0.0190, 0.0177, 0.0213, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:10:18,066 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72682.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:10:18,645 INFO [finetune.py:976] (0/7) Epoch 13, batch 3950, loss[loss=0.1652, simple_loss=0.2487, pruned_loss=0.04082, over 4870.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2527, pruned_loss=0.05931, over 956741.59 frames. ], batch size: 34, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:10:51,433 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 16:11:10,484 INFO [finetune.py:976] (0/7) Epoch 13, batch 4000, loss[loss=0.2059, simple_loss=0.2734, pruned_loss=0.06921, over 4770.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2513, pruned_loss=0.05892, over 955867.75 frames. ], batch size: 26, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:11:16,799 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.596e+02 1.921e+02 2.181e+02 4.609e+02, threshold=3.842e+02, percent-clipped=3.0 +2023-03-26 16:11:38,453 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4027, 1.2805, 1.5139, 0.7460, 1.4866, 1.4118, 1.3923, 1.2306], + device='cuda:0'), covar=tensor([0.0611, 0.0739, 0.0635, 0.0985, 0.0825, 0.0746, 0.0628, 0.1206], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0132, 0.0141, 0.0122, 0.0123, 0.0141, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:11:44,631 INFO [finetune.py:976] (0/7) Epoch 13, batch 4050, loss[loss=0.2052, simple_loss=0.2705, pruned_loss=0.06994, over 4832.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2549, pruned_loss=0.06038, over 954497.76 frames. ], batch size: 33, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:11:46,019 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72785.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:11:47,160 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0733, 1.8513, 1.7333, 1.7809, 1.7718, 1.7722, 1.7677, 2.4387], + device='cuda:0'), covar=tensor([0.3540, 0.4235, 0.3044, 0.3685, 0.3845, 0.2406, 0.3835, 0.1579], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0259, 0.0222, 0.0275, 0.0245, 0.0211, 0.0246, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:12:37,535 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4939, 1.5257, 1.6518, 1.8317, 1.5950, 3.2230, 1.3656, 1.5631], + device='cuda:0'), covar=tensor([0.0919, 0.1728, 0.1133, 0.0911, 0.1604, 0.0266, 0.1524, 0.1759], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:12:39,437 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0606, 0.8989, 0.8943, 1.0929, 1.2235, 1.1275, 0.9959, 0.9097], + device='cuda:0'), covar=tensor([0.0362, 0.0334, 0.0646, 0.0332, 0.0298, 0.0495, 0.0346, 0.0450], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0139, 0.0112, 0.0101, 0.0105, 0.0095, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2845e-05, 8.4626e-05, 1.1037e-04, 8.7319e-05, 7.8874e-05, 7.7784e-05, + 7.1565e-05, 8.3128e-05], device='cuda:0') +2023-03-26 16:12:39,938 INFO [finetune.py:976] (0/7) Epoch 13, batch 4100, loss[loss=0.1816, simple_loss=0.2622, pruned_loss=0.05045, over 4789.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2584, pruned_loss=0.06192, over 954529.69 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:12:40,000 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72833.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:12:45,292 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.592e+02 1.875e+02 2.230e+02 3.624e+02, threshold=3.749e+02, percent-clipped=0.0 +2023-03-26 16:13:11,603 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5110, 2.3286, 1.8713, 2.5230, 2.3495, 2.0045, 2.8854, 2.4365], + device='cuda:0'), covar=tensor([0.1302, 0.2333, 0.3181, 0.2634, 0.2582, 0.1702, 0.3641, 0.1859], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0188, 0.0235, 0.0255, 0.0245, 0.0200, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:13:13,288 INFO [finetune.py:976] (0/7) Epoch 13, batch 4150, loss[loss=0.2256, simple_loss=0.2886, pruned_loss=0.0813, over 4765.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2603, pruned_loss=0.0627, over 950789.28 frames. ], batch size: 54, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:13:26,364 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72894.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:13:33,979 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7490, 1.5282, 2.2061, 3.4175, 2.3413, 2.4212, 1.0980, 2.7526], + device='cuda:0'), covar=tensor([0.1708, 0.1428, 0.1214, 0.0506, 0.0747, 0.1324, 0.1756, 0.0509], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0114, 0.0132, 0.0163, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:13:52,638 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0321, 4.7771, 4.5751, 2.9003, 4.9053, 3.7059, 0.9754, 3.3001], + device='cuda:0'), covar=tensor([0.2215, 0.1526, 0.1147, 0.2503, 0.0666, 0.0785, 0.4482, 0.1223], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 16:14:03,790 INFO [finetune.py:976] (0/7) Epoch 13, batch 4200, loss[loss=0.2058, simple_loss=0.2735, pruned_loss=0.06902, over 4800.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2593, pruned_loss=0.06182, over 949589.99 frames. ], batch size: 25, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:14:08,711 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.693e+01 1.496e+02 1.812e+02 2.169e+02 4.504e+02, threshold=3.624e+02, percent-clipped=2.0 +2023-03-26 16:14:18,708 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:14:30,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72964.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:14:52,055 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72982.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:14:53,055 INFO [finetune.py:976] (0/7) Epoch 13, batch 4250, loss[loss=0.1899, simple_loss=0.2519, pruned_loss=0.06401, over 4897.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2576, pruned_loss=0.06184, over 951313.05 frames. ], batch size: 32, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:14:56,859 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1156, 2.0454, 2.2220, 1.5394, 2.1833, 2.3169, 2.2805, 1.8008], + device='cuda:0'), covar=tensor([0.0587, 0.0635, 0.0626, 0.0862, 0.0646, 0.0653, 0.0562, 0.1099], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0131, 0.0140, 0.0122, 0.0122, 0.0140, 0.0139, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:14:58,106 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0169, 1.6477, 2.1471, 1.5642, 1.9597, 2.1284, 1.6618, 2.2578], + device='cuda:0'), covar=tensor([0.0762, 0.1632, 0.1053, 0.1391, 0.0585, 0.0969, 0.2008, 0.0572], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0205, 0.0192, 0.0191, 0.0178, 0.0213, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:15:10,521 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8905, 1.7041, 1.5992, 1.8744, 2.3353, 2.0069, 1.4955, 1.5475], + device='cuda:0'), covar=tensor([0.2147, 0.2044, 0.1949, 0.1718, 0.1687, 0.1132, 0.2457, 0.1860], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0209, 0.0211, 0.0192, 0.0242, 0.0184, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:15:21,363 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7021, 1.5669, 1.4976, 1.6182, 0.9644, 3.2472, 1.1908, 1.6043], + device='cuda:0'), covar=tensor([0.3405, 0.2434, 0.2132, 0.2350, 0.1944, 0.0209, 0.2689, 0.1382], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0123, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:15:21,389 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73025.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 16:15:24,375 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:15:26,658 INFO [finetune.py:976] (0/7) Epoch 13, batch 4300, loss[loss=0.1919, simple_loss=0.2609, pruned_loss=0.06148, over 4737.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.255, pruned_loss=0.06085, over 950963.88 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:15:26,776 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8273, 1.6974, 1.5032, 1.8931, 2.2119, 1.9752, 1.4954, 1.5069], + device='cuda:0'), covar=tensor([0.2127, 0.2007, 0.1960, 0.1616, 0.1608, 0.1127, 0.2520, 0.1895], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0208, 0.0210, 0.0191, 0.0241, 0.0183, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:15:27,942 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5737, 1.5096, 1.7008, 1.8149, 1.5955, 3.1653, 1.3588, 1.6602], + device='cuda:0'), covar=tensor([0.0863, 0.1746, 0.1106, 0.0878, 0.1477, 0.0257, 0.1397, 0.1625], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:15:29,132 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1598, 1.7045, 2.1791, 2.0969, 1.8615, 1.8683, 2.1037, 1.9529], + device='cuda:0'), covar=tensor([0.4411, 0.4659, 0.3471, 0.4246, 0.5222, 0.4262, 0.5002, 0.3572], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0237, 0.0255, 0.0264, 0.0261, 0.0235, 0.0275, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:15:31,990 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.498e+02 1.782e+02 2.254e+02 4.055e+02, threshold=3.563e+02, percent-clipped=2.0 +2023-03-26 16:15:40,036 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-26 16:15:54,781 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9216, 1.2945, 1.9097, 1.8324, 1.6626, 1.6125, 1.7881, 1.7349], + device='cuda:0'), covar=tensor([0.3696, 0.4095, 0.3342, 0.3613, 0.4599, 0.3746, 0.4446, 0.3183], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0237, 0.0255, 0.0263, 0.0261, 0.0235, 0.0275, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:15:59,433 INFO [finetune.py:976] (0/7) Epoch 13, batch 4350, loss[loss=0.1854, simple_loss=0.2468, pruned_loss=0.06199, over 4743.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2522, pruned_loss=0.06018, over 952004.92 frames. ], batch size: 54, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:16:34,969 INFO [finetune.py:976] (0/7) Epoch 13, batch 4400, loss[loss=0.2251, simple_loss=0.288, pruned_loss=0.08115, over 4768.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2519, pruned_loss=0.06009, over 950369.98 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:16:40,310 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.328e+01 1.433e+02 1.829e+02 2.142e+02 3.915e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-03-26 16:16:55,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0884, 1.6803, 2.1298, 2.0408, 1.8057, 1.8405, 1.9720, 1.9016], + device='cuda:0'), covar=tensor([0.4326, 0.4728, 0.3533, 0.4457, 0.5498, 0.4161, 0.5351, 0.3543], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0235, 0.0254, 0.0262, 0.0259, 0.0234, 0.0274, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:17:06,970 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5391, 1.4224, 1.3679, 1.4237, 0.9081, 2.9898, 1.1751, 1.6246], + device='cuda:0'), covar=tensor([0.3633, 0.2686, 0.2385, 0.2608, 0.2169, 0.0285, 0.2763, 0.1415], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0098, 0.0098, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:17:08,723 INFO [finetune.py:976] (0/7) Epoch 13, batch 4450, loss[loss=0.1874, simple_loss=0.2607, pruned_loss=0.057, over 4899.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2545, pruned_loss=0.06037, over 951973.27 frames. ], batch size: 35, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:17:53,150 INFO [finetune.py:976] (0/7) Epoch 13, batch 4500, loss[loss=0.1955, simple_loss=0.2666, pruned_loss=0.06222, over 4866.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2565, pruned_loss=0.06104, over 952994.70 frames. ], batch size: 31, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:17:58,021 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.732e+02 2.105e+02 2.505e+02 4.470e+02, threshold=4.210e+02, percent-clipped=3.0 +2023-03-26 16:18:04,471 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:18:26,891 INFO [finetune.py:976] (0/7) Epoch 13, batch 4550, loss[loss=0.1611, simple_loss=0.2398, pruned_loss=0.04115, over 4862.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2585, pruned_loss=0.06161, over 953561.92 frames. ], batch size: 34, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:18:34,961 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:19:07,528 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73320.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:19:08,262 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 16:19:09,330 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2435, 2.1465, 1.7256, 2.1868, 2.0407, 2.0279, 1.9811, 3.0454], + device='cuda:0'), covar=tensor([0.4256, 0.5734, 0.3957, 0.4992, 0.5054, 0.2676, 0.5329, 0.1721], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0261, 0.0224, 0.0279, 0.0247, 0.0213, 0.0249, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:19:19,978 INFO [finetune.py:976] (0/7) Epoch 13, batch 4600, loss[loss=0.2242, simple_loss=0.2832, pruned_loss=0.08257, over 4820.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2583, pruned_loss=0.06158, over 954621.98 frames. ], batch size: 38, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:19:24,888 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.809e+01 1.600e+02 1.901e+02 2.194e+02 3.702e+02, threshold=3.803e+02, percent-clipped=0.0 +2023-03-26 16:19:42,011 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73354.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:20:11,744 INFO [finetune.py:976] (0/7) Epoch 13, batch 4650, loss[loss=0.1404, simple_loss=0.2147, pruned_loss=0.03308, over 4756.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2565, pruned_loss=0.06103, over 956273.10 frames. ], batch size: 27, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:20:20,296 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1252, 1.9591, 1.7855, 1.6115, 1.8882, 1.9231, 1.8872, 2.4979], + device='cuda:0'), covar=tensor([0.3977, 0.3673, 0.3481, 0.3884, 0.3647, 0.2578, 0.3687, 0.1847], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0223, 0.0278, 0.0247, 0.0213, 0.0248, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:20:45,672 INFO [finetune.py:976] (0/7) Epoch 13, batch 4700, loss[loss=0.1813, simple_loss=0.2496, pruned_loss=0.05651, over 4778.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2536, pruned_loss=0.06037, over 955541.12 frames. ], batch size: 26, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:20:50,436 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.614e+02 1.909e+02 2.257e+02 3.771e+02, threshold=3.817e+02, percent-clipped=0.0 +2023-03-26 16:21:03,505 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6333, 1.6043, 1.3988, 1.4735, 1.8926, 1.8338, 1.6648, 1.3946], + device='cuda:0'), covar=tensor([0.0318, 0.0304, 0.0512, 0.0315, 0.0235, 0.0490, 0.0299, 0.0461], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0112, 0.0100, 0.0105, 0.0095, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2662e-05, 8.3979e-05, 1.0983e-04, 8.7061e-05, 7.8451e-05, 7.7465e-05, + 7.1472e-05, 8.3260e-05], device='cuda:0') +2023-03-26 16:21:15,595 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-03-26 16:21:18,783 INFO [finetune.py:976] (0/7) Epoch 13, batch 4750, loss[loss=0.1834, simple_loss=0.247, pruned_loss=0.0599, over 4709.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2517, pruned_loss=0.05963, over 954279.56 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:21:27,238 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 16:21:51,903 INFO [finetune.py:976] (0/7) Epoch 13, batch 4800, loss[loss=0.2166, simple_loss=0.2892, pruned_loss=0.07197, over 4753.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2533, pruned_loss=0.06038, over 952209.25 frames. ], batch size: 59, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:21:55,513 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7575, 2.4276, 1.9757, 1.2128, 2.2645, 2.2871, 1.8503, 2.1312], + device='cuda:0'), covar=tensor([0.0765, 0.0734, 0.1376, 0.1725, 0.1116, 0.1392, 0.1937, 0.0840], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0195, 0.0199, 0.0185, 0.0211, 0.0206, 0.0221, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:21:57,193 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.637e+02 2.007e+02 2.318e+02 3.852e+02, threshold=4.014e+02, percent-clipped=1.0 +2023-03-26 16:22:03,326 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:17,475 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 16:22:24,805 INFO [finetune.py:976] (0/7) Epoch 13, batch 4850, loss[loss=0.1917, simple_loss=0.2768, pruned_loss=0.05324, over 4802.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2573, pruned_loss=0.06188, over 951899.51 frames. ], batch size: 45, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:22:30,077 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73590.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:30,722 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1188, 1.9870, 1.6726, 1.9936, 1.9505, 1.8925, 1.9237, 2.6765], + device='cuda:0'), covar=tensor([0.4000, 0.4721, 0.3441, 0.4098, 0.4203, 0.2424, 0.3997, 0.1625], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0224, 0.0278, 0.0247, 0.0213, 0.0248, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:22:37,204 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73598.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:48,776 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-26 16:22:56,816 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2864, 1.9770, 1.4496, 0.5721, 1.7167, 1.8807, 1.6329, 1.8427], + device='cuda:0'), covar=tensor([0.0648, 0.0698, 0.1265, 0.1697, 0.1074, 0.1815, 0.2047, 0.0757], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0195, 0.0199, 0.0185, 0.0211, 0.0206, 0.0221, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:23:00,189 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73620.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:23:08,515 INFO [finetune.py:976] (0/7) Epoch 13, batch 4900, loss[loss=0.1786, simple_loss=0.2466, pruned_loss=0.05529, over 4761.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2589, pruned_loss=0.06217, over 953149.67 frames. ], batch size: 26, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:23:14,277 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.751e+02 2.108e+02 2.596e+02 5.059e+02, threshold=4.217e+02, percent-clipped=3.0 +2023-03-26 16:23:19,729 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:23:21,031 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:23:27,110 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-26 16:23:28,873 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1306, 2.0818, 1.6539, 1.9620, 2.0873, 1.8141, 2.4327, 2.1888], + device='cuda:0'), covar=tensor([0.1345, 0.2221, 0.3109, 0.2682, 0.2495, 0.1741, 0.3005, 0.1725], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0187, 0.0233, 0.0253, 0.0242, 0.0198, 0.0211, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:23:31,878 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73668.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:23:37,830 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6953, 1.5842, 1.8924, 2.9599, 2.1064, 2.2319, 1.0587, 2.4370], + device='cuda:0'), covar=tensor([0.1655, 0.1346, 0.1234, 0.0623, 0.0799, 0.1151, 0.1704, 0.0606], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0115, 0.0133, 0.0165, 0.0101, 0.0138, 0.0126, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:23:41,382 INFO [finetune.py:976] (0/7) Epoch 13, batch 4950, loss[loss=0.1839, simple_loss=0.2639, pruned_loss=0.05197, over 4775.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2599, pruned_loss=0.06219, over 953843.13 frames. ], batch size: 26, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:23:45,768 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 16:23:57,354 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 16:24:03,084 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0062, 1.9264, 2.1090, 1.4077, 2.0402, 2.1175, 2.1281, 1.6897], + device='cuda:0'), covar=tensor([0.0560, 0.0650, 0.0635, 0.0945, 0.0691, 0.0688, 0.0544, 0.1113], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0132, 0.0140, 0.0122, 0.0122, 0.0140, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:24:24,438 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:24:24,937 INFO [finetune.py:976] (0/7) Epoch 13, batch 5000, loss[loss=0.1572, simple_loss=0.2344, pruned_loss=0.03996, over 4900.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2591, pruned_loss=0.06176, over 954433.43 frames. ], batch size: 43, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:24:33,742 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.586e+02 1.888e+02 2.371e+02 3.310e+02, threshold=3.776e+02, percent-clipped=1.0 +2023-03-26 16:25:17,330 INFO [finetune.py:976] (0/7) Epoch 13, batch 5050, loss[loss=0.1322, simple_loss=0.1986, pruned_loss=0.03286, over 4811.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.257, pruned_loss=0.06159, over 953390.90 frames. ], batch size: 25, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:25:26,622 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:25:53,899 INFO [finetune.py:976] (0/7) Epoch 13, batch 5100, loss[loss=0.1448, simple_loss=0.217, pruned_loss=0.03631, over 4746.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2533, pruned_loss=0.05999, over 954615.44 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:25:55,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1646, 2.1483, 2.0439, 2.2781, 2.5627, 2.1900, 2.0330, 1.6625], + device='cuda:0'), covar=tensor([0.2142, 0.1884, 0.1754, 0.1615, 0.1818, 0.1139, 0.2073, 0.1987], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0207, 0.0210, 0.0190, 0.0240, 0.0183, 0.0213, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:25:59,158 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.804e+01 1.435e+02 1.752e+02 2.086e+02 3.868e+02, threshold=3.504e+02, percent-clipped=1.0 +2023-03-26 16:26:11,093 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 16:26:27,646 INFO [finetune.py:976] (0/7) Epoch 13, batch 5150, loss[loss=0.2394, simple_loss=0.2934, pruned_loss=0.09274, over 4896.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2533, pruned_loss=0.06033, over 953555.83 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:26:36,028 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3396, 1.3353, 1.4817, 1.5201, 1.4434, 2.9830, 1.1726, 1.4463], + device='cuda:0'), covar=tensor([0.1375, 0.2447, 0.1357, 0.1278, 0.2049, 0.0419, 0.2207, 0.2316], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0080, 0.0073, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:26:48,498 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 16:27:01,331 INFO [finetune.py:976] (0/7) Epoch 13, batch 5200, loss[loss=0.2002, simple_loss=0.2688, pruned_loss=0.06581, over 4901.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2577, pruned_loss=0.0621, over 953383.78 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:27:06,219 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.675e+02 1.952e+02 2.217e+02 3.649e+02, threshold=3.904e+02, percent-clipped=2.0 +2023-03-26 16:27:09,798 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73946.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:11,698 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:23,649 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 16:27:34,704 INFO [finetune.py:976] (0/7) Epoch 13, batch 5250, loss[loss=0.1459, simple_loss=0.2162, pruned_loss=0.03784, over 4736.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2574, pruned_loss=0.06109, over 953447.92 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:27:43,821 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0376, 1.8761, 1.6733, 1.7937, 1.8275, 1.8337, 1.8387, 2.5422], + device='cuda:0'), covar=tensor([0.3950, 0.4667, 0.3425, 0.4402, 0.4305, 0.2453, 0.4292, 0.1800], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0224, 0.0277, 0.0246, 0.0213, 0.0247, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:27:44,361 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:44,402 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6060, 1.2988, 0.8937, 1.4274, 1.9948, 1.0303, 1.3900, 1.6039], + device='cuda:0'), covar=tensor([0.1538, 0.2080, 0.2009, 0.1264, 0.2150, 0.2047, 0.1546, 0.1831], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0092, 0.0120, 0.0094, 0.0099, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:27:46,254 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-74000.pt +2023-03-26 16:28:10,312 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 16:28:11,317 INFO [finetune.py:976] (0/7) Epoch 13, batch 5300, loss[loss=0.237, simple_loss=0.2929, pruned_loss=0.09053, over 4870.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2603, pruned_loss=0.06202, over 955615.56 frames. ], batch size: 34, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:28:17,129 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.638e+02 2.059e+02 2.515e+02 4.122e+02, threshold=4.117e+02, percent-clipped=3.0 +2023-03-26 16:28:24,429 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6380, 0.6874, 1.6221, 1.5483, 1.4801, 1.3703, 1.4656, 1.5672], + device='cuda:0'), covar=tensor([0.3694, 0.4131, 0.3367, 0.3619, 0.4646, 0.3772, 0.4100, 0.3271], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0237, 0.0255, 0.0264, 0.0261, 0.0235, 0.0275, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:28:33,154 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:28:45,016 INFO [finetune.py:976] (0/7) Epoch 13, batch 5350, loss[loss=0.2177, simple_loss=0.28, pruned_loss=0.07766, over 4878.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2608, pruned_loss=0.06235, over 954884.44 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:28:48,551 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:28:55,629 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6058, 1.5141, 2.0450, 1.8081, 1.7387, 4.1928, 1.3618, 1.7927], + device='cuda:0'), covar=tensor([0.0908, 0.1824, 0.1263, 0.0982, 0.1556, 0.0229, 0.1557, 0.1698], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:29:12,972 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74125.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:29:18,157 INFO [finetune.py:976] (0/7) Epoch 13, batch 5400, loss[loss=0.1551, simple_loss=0.2176, pruned_loss=0.04634, over 4384.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2571, pruned_loss=0.06086, over 953678.00 frames. ], batch size: 19, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:29:27,916 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.534e+02 1.853e+02 2.261e+02 4.254e+02, threshold=3.706e+02, percent-clipped=1.0 +2023-03-26 16:30:11,887 INFO [finetune.py:976] (0/7) Epoch 13, batch 5450, loss[loss=0.1437, simple_loss=0.2166, pruned_loss=0.03538, over 4798.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.255, pruned_loss=0.05993, over 955305.48 frames. ], batch size: 29, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:30:56,458 INFO [finetune.py:976] (0/7) Epoch 13, batch 5500, loss[loss=0.2132, simple_loss=0.2799, pruned_loss=0.07323, over 4803.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2518, pruned_loss=0.05918, over 957450.43 frames. ], batch size: 45, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:31:01,348 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:01,848 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.610e+01 1.534e+02 1.911e+02 2.187e+02 5.924e+02, threshold=3.822e+02, percent-clipped=2.0 +2023-03-26 16:31:04,994 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:30,481 INFO [finetune.py:976] (0/7) Epoch 13, batch 5550, loss[loss=0.2643, simple_loss=0.3257, pruned_loss=0.1014, over 4838.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2539, pruned_loss=0.06037, over 957409.35 frames. ], batch size: 47, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:31:37,728 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:42,487 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74301.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:32:02,277 INFO [finetune.py:976] (0/7) Epoch 13, batch 5600, loss[loss=0.2058, simple_loss=0.2747, pruned_loss=0.06843, over 4754.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2573, pruned_loss=0.06093, over 956510.82 frames. ], batch size: 59, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:32:06,848 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.439e+01 1.528e+02 1.833e+02 2.251e+02 4.644e+02, threshold=3.666e+02, percent-clipped=1.0 +2023-03-26 16:32:31,596 INFO [finetune.py:976] (0/7) Epoch 13, batch 5650, loss[loss=0.2256, simple_loss=0.2961, pruned_loss=0.07751, over 4825.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2598, pruned_loss=0.06182, over 956078.94 frames. ], batch size: 40, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:32:35,019 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:32:54,213 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:32:57,284 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 16:33:00,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1424, 4.2297, 3.9495, 2.6822, 4.2898, 3.5290, 1.4149, 3.1058], + device='cuda:0'), covar=tensor([0.2298, 0.2251, 0.1664, 0.3045, 0.1039, 0.0940, 0.4665, 0.1560], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0161, 0.0128, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 16:33:01,854 INFO [finetune.py:976] (0/7) Epoch 13, batch 5700, loss[loss=0.1504, simple_loss=0.2069, pruned_loss=0.04699, over 4062.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2561, pruned_loss=0.06122, over 937969.92 frames. ], batch size: 17, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:33:03,647 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:33:06,487 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.627e+01 1.487e+02 1.916e+02 2.529e+02 4.839e+02, threshold=3.833e+02, percent-clipped=5.0 +2023-03-26 16:33:18,403 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-13.pt +2023-03-26 16:33:31,119 INFO [finetune.py:976] (0/7) Epoch 14, batch 0, loss[loss=0.1801, simple_loss=0.2601, pruned_loss=0.0501, over 4897.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2601, pruned_loss=0.0501, over 4897.00 frames. ], batch size: 43, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:33:31,120 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 16:33:41,692 INFO [finetune.py:1010] (0/7) Epoch 14, validation: loss=0.1582, simple_loss=0.2295, pruned_loss=0.04344, over 2265189.00 frames. +2023-03-26 16:33:41,692 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 16:33:53,063 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5269, 1.3554, 1.7182, 1.7469, 1.5042, 3.3827, 1.3329, 1.5015], + device='cuda:0'), covar=tensor([0.0954, 0.1848, 0.1088, 0.0994, 0.1637, 0.0258, 0.1600, 0.1804], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:34:14,520 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 16:34:14,919 INFO [finetune.py:976] (0/7) Epoch 14, batch 50, loss[loss=0.1468, simple_loss=0.2241, pruned_loss=0.03471, over 4758.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2564, pruned_loss=0.0608, over 216581.28 frames. ], batch size: 26, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:34:42,636 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.553e+01 1.578e+02 1.920e+02 2.248e+02 3.729e+02, threshold=3.841e+02, percent-clipped=1.0 +2023-03-26 16:34:45,167 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.07 vs. limit=5.0 +2023-03-26 16:35:04,285 INFO [finetune.py:976] (0/7) Epoch 14, batch 100, loss[loss=0.1619, simple_loss=0.2206, pruned_loss=0.05158, over 4886.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2515, pruned_loss=0.05989, over 380999.04 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:35:04,420 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2022, 2.1804, 1.9287, 2.3394, 2.1308, 2.0754, 2.0230, 2.8552], + device='cuda:0'), covar=tensor([0.3471, 0.5228, 0.3323, 0.4502, 0.4544, 0.2445, 0.4862, 0.1439], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0225, 0.0277, 0.0247, 0.0213, 0.0248, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:35:31,041 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:35:31,622 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74596.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:35:49,899 INFO [finetune.py:976] (0/7) Epoch 14, batch 150, loss[loss=0.1523, simple_loss=0.2182, pruned_loss=0.04321, over 4899.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2482, pruned_loss=0.05928, over 508238.62 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:36:22,702 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.552e+02 1.795e+02 2.139e+02 3.747e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-26 16:36:32,532 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:36:35,947 INFO [finetune.py:976] (0/7) Epoch 14, batch 200, loss[loss=0.1536, simple_loss=0.2233, pruned_loss=0.04198, over 4905.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2459, pruned_loss=0.05822, over 606427.17 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:37:09,842 INFO [finetune.py:976] (0/7) Epoch 14, batch 250, loss[loss=0.1275, simple_loss=0.2015, pruned_loss=0.02672, over 4752.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2501, pruned_loss=0.06002, over 679786.35 frames. ], batch size: 27, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:37:15,969 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:37:30,103 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.719e+02 2.081e+02 2.468e+02 4.342e+02, threshold=4.162e+02, percent-clipped=2.0 +2023-03-26 16:37:41,608 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7836, 1.5894, 1.3474, 1.2413, 1.5480, 1.5043, 1.5058, 2.1186], + device='cuda:0'), covar=tensor([0.3683, 0.3910, 0.3227, 0.3743, 0.3732, 0.2334, 0.3762, 0.1716], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0224, 0.0275, 0.0246, 0.0213, 0.0247, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:37:42,696 INFO [finetune.py:976] (0/7) Epoch 14, batch 300, loss[loss=0.13, simple_loss=0.1956, pruned_loss=0.03223, over 4722.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2553, pruned_loss=0.0619, over 741017.48 frames. ], batch size: 23, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:37:48,012 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74768.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:38:16,372 INFO [finetune.py:976] (0/7) Epoch 14, batch 350, loss[loss=0.2092, simple_loss=0.2788, pruned_loss=0.06975, over 4906.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2581, pruned_loss=0.06246, over 789617.92 frames. ], batch size: 37, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:38:36,813 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.618e+01 1.648e+02 1.967e+02 2.475e+02 5.107e+02, threshold=3.933e+02, percent-clipped=3.0 +2023-03-26 16:38:39,429 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 16:38:49,809 INFO [finetune.py:976] (0/7) Epoch 14, batch 400, loss[loss=0.2308, simple_loss=0.3103, pruned_loss=0.07562, over 4877.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2584, pruned_loss=0.06165, over 827906.95 frames. ], batch size: 43, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:39:13,571 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74896.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:22,419 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:23,515 INFO [finetune.py:976] (0/7) Epoch 14, batch 450, loss[loss=0.2022, simple_loss=0.2686, pruned_loss=0.06792, over 4896.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2558, pruned_loss=0.05968, over 855440.49 frames. ], batch size: 43, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:39:35,424 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0935, 2.0132, 1.7998, 2.0400, 1.8045, 4.5693, 1.7257, 2.4729], + device='cuda:0'), covar=tensor([0.3091, 0.2242, 0.2086, 0.2124, 0.1354, 0.0093, 0.2335, 0.1112], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:39:43,596 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.587e+02 1.868e+02 2.260e+02 4.285e+02, threshold=3.737e+02, percent-clipped=2.0 +2023-03-26 16:39:45,922 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:50,146 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:58,628 INFO [finetune.py:976] (0/7) Epoch 14, batch 500, loss[loss=0.1458, simple_loss=0.2163, pruned_loss=0.03768, over 4781.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2532, pruned_loss=0.05871, over 877529.18 frames. ], batch size: 26, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:40:00,570 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:40:08,973 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:40:45,474 INFO [finetune.py:976] (0/7) Epoch 14, batch 550, loss[loss=0.1929, simple_loss=0.2749, pruned_loss=0.05542, over 4835.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2516, pruned_loss=0.05863, over 894949.52 frames. ], batch size: 49, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:40:53,731 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1613, 2.2626, 1.8759, 2.2874, 2.1786, 2.1162, 2.1086, 2.9097], + device='cuda:0'), covar=tensor([0.4405, 0.4990, 0.3723, 0.4739, 0.4550, 0.2715, 0.4730, 0.1865], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0225, 0.0276, 0.0247, 0.0214, 0.0249, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:40:57,867 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75025.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:41:09,612 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.564e+02 1.846e+02 2.204e+02 7.411e+02, threshold=3.691e+02, percent-clipped=3.0 +2023-03-26 16:41:32,954 INFO [finetune.py:976] (0/7) Epoch 14, batch 600, loss[loss=0.1839, simple_loss=0.2597, pruned_loss=0.05403, over 4875.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2524, pruned_loss=0.05938, over 908472.86 frames. ], batch size: 34, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:41:34,810 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4889, 1.3852, 1.3613, 1.4242, 0.8541, 2.8883, 0.9073, 1.3426], + device='cuda:0'), covar=tensor([0.3312, 0.2511, 0.2248, 0.2452, 0.2046, 0.0266, 0.2912, 0.1444], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0098, 0.0098, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:41:56,823 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 16:41:57,400 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-03-26 16:42:00,290 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7042, 1.7229, 1.3984, 1.5410, 2.0649, 1.9722, 1.7510, 1.4650], + device='cuda:0'), covar=tensor([0.0279, 0.0314, 0.0608, 0.0350, 0.0191, 0.0484, 0.0335, 0.0480], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0140, 0.0113, 0.0100, 0.0105, 0.0095, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2165e-05, 8.3839e-05, 1.1070e-04, 8.7597e-05, 7.8024e-05, 7.8068e-05, + 7.1515e-05, 8.2861e-05], device='cuda:0') +2023-03-26 16:42:05,138 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 16:42:10,221 INFO [finetune.py:976] (0/7) Epoch 14, batch 650, loss[loss=0.1813, simple_loss=0.2535, pruned_loss=0.05458, over 4924.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2537, pruned_loss=0.05951, over 917525.44 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:42:30,914 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.620e+02 1.922e+02 2.248e+02 3.855e+02, threshold=3.845e+02, percent-clipped=1.0 +2023-03-26 16:42:43,796 INFO [finetune.py:976] (0/7) Epoch 14, batch 700, loss[loss=0.138, simple_loss=0.193, pruned_loss=0.04154, over 3872.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2569, pruned_loss=0.06133, over 925206.71 frames. ], batch size: 16, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:43:16,885 INFO [finetune.py:976] (0/7) Epoch 14, batch 750, loss[loss=0.1482, simple_loss=0.2275, pruned_loss=0.03451, over 4745.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2578, pruned_loss=0.06114, over 933430.34 frames. ], batch size: 26, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:43:28,279 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75228.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:37,703 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.931e+01 1.583e+02 1.834e+02 2.163e+02 4.783e+02, threshold=3.668e+02, percent-clipped=1.0 +2023-03-26 16:43:44,241 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:50,665 INFO [finetune.py:976] (0/7) Epoch 14, batch 800, loss[loss=0.205, simple_loss=0.2607, pruned_loss=0.07465, over 4813.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2572, pruned_loss=0.06058, over 935641.95 frames. ], batch size: 25, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:43:52,032 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2772, 1.9371, 1.9020, 2.2407, 2.5451, 2.1462, 2.0531, 1.7024], + device='cuda:0'), covar=tensor([0.1991, 0.2100, 0.1776, 0.1601, 0.1767, 0.1112, 0.2059, 0.1809], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0206, 0.0208, 0.0189, 0.0238, 0.0183, 0.0212, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:43:53,645 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:56,675 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:09,562 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:16,557 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:24,292 INFO [finetune.py:976] (0/7) Epoch 14, batch 850, loss[loss=0.1747, simple_loss=0.2385, pruned_loss=0.05546, over 4047.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2563, pruned_loss=0.06111, over 939043.54 frames. ], batch size: 17, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:44:30,261 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:37,473 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75331.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:44:44,944 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.704e+01 1.586e+02 1.985e+02 2.275e+02 3.825e+02, threshold=3.970e+02, percent-clipped=2.0 +2023-03-26 16:44:57,420 INFO [finetune.py:976] (0/7) Epoch 14, batch 900, loss[loss=0.1795, simple_loss=0.2505, pruned_loss=0.05426, over 4768.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2532, pruned_loss=0.05969, over 942721.91 frames. ], batch size: 27, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:45:44,929 INFO [finetune.py:976] (0/7) Epoch 14, batch 950, loss[loss=0.235, simple_loss=0.2929, pruned_loss=0.08853, over 4913.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2525, pruned_loss=0.0601, over 947204.66 frames. ], batch size: 37, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:45:54,647 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0382, 2.6644, 2.5094, 1.3809, 2.6519, 2.2045, 2.0281, 2.3119], + device='cuda:0'), covar=tensor([0.1203, 0.0863, 0.1812, 0.2046, 0.1894, 0.2136, 0.2232, 0.1491], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0194, 0.0197, 0.0183, 0.0211, 0.0206, 0.0221, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:46:05,787 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.942e+01 1.566e+02 1.871e+02 2.243e+02 4.539e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-26 16:46:18,859 INFO [finetune.py:976] (0/7) Epoch 14, batch 1000, loss[loss=0.1695, simple_loss=0.2509, pruned_loss=0.04403, over 4752.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2556, pruned_loss=0.061, over 949702.94 frames. ], batch size: 28, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:46:27,930 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7667, 1.1327, 1.7722, 1.7275, 1.5378, 1.4879, 1.6102, 1.6053], + device='cuda:0'), covar=tensor([0.3810, 0.4191, 0.3346, 0.3557, 0.4701, 0.3915, 0.4629, 0.3380], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0237, 0.0255, 0.0264, 0.0262, 0.0236, 0.0276, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:47:07,212 INFO [finetune.py:976] (0/7) Epoch 14, batch 1050, loss[loss=0.2148, simple_loss=0.2655, pruned_loss=0.08204, over 4932.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2576, pruned_loss=0.06053, over 952853.02 frames. ], batch size: 33, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:47:31,085 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.109e+01 1.606e+02 2.003e+02 2.356e+02 8.983e+02, threshold=4.007e+02, percent-clipped=2.0 +2023-03-26 16:47:44,013 INFO [finetune.py:976] (0/7) Epoch 14, batch 1100, loss[loss=0.2419, simple_loss=0.3065, pruned_loss=0.08868, over 4924.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2587, pruned_loss=0.06063, over 954674.25 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:47:45,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6466, 1.5363, 1.5143, 1.5491, 1.0389, 3.3633, 1.1680, 1.6622], + device='cuda:0'), covar=tensor([0.3383, 0.2498, 0.2157, 0.2560, 0.1896, 0.0204, 0.2760, 0.1358], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:47:47,090 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:47:59,612 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:09,084 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7717, 3.3631, 3.5183, 3.6702, 3.5515, 3.3909, 3.8900, 1.2031], + device='cuda:0'), covar=tensor([0.1001, 0.0872, 0.0860, 0.1011, 0.1425, 0.1716, 0.0826, 0.5536], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0277, 0.0290, 0.0332, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:48:14,630 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7259, 1.5852, 1.9235, 1.9844, 1.7244, 4.3139, 1.4437, 1.8097], + device='cuda:0'), covar=tensor([0.0956, 0.1776, 0.1394, 0.0913, 0.1641, 0.0196, 0.1481, 0.1690], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:48:17,573 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4655, 1.4354, 1.3420, 1.5139, 1.8240, 1.7119, 1.5938, 1.3020], + device='cuda:0'), covar=tensor([0.0370, 0.0268, 0.0582, 0.0318, 0.0207, 0.0409, 0.0275, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0112, 0.0100, 0.0105, 0.0095, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.2544e-05, 8.3703e-05, 1.1027e-04, 8.7462e-05, 7.7687e-05, 7.7753e-05, + 7.1512e-05, 8.2270e-05], device='cuda:0') +2023-03-26 16:48:18,069 INFO [finetune.py:976] (0/7) Epoch 14, batch 1150, loss[loss=0.1638, simple_loss=0.2506, pruned_loss=0.03852, over 4896.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2602, pruned_loss=0.06148, over 953908.59 frames. ], batch size: 36, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:48:19,335 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:24,055 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75620.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:28,202 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75626.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:48:38,760 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.806e+01 1.586e+02 1.986e+02 2.337e+02 5.787e+02, threshold=3.972e+02, percent-clipped=2.0 +2023-03-26 16:48:51,183 INFO [finetune.py:976] (0/7) Epoch 14, batch 1200, loss[loss=0.1946, simple_loss=0.2475, pruned_loss=0.07089, over 4936.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2578, pruned_loss=0.06045, over 954577.95 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:48:56,394 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:49:24,700 INFO [finetune.py:976] (0/7) Epoch 14, batch 1250, loss[loss=0.2006, simple_loss=0.2574, pruned_loss=0.07193, over 4274.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2546, pruned_loss=0.0594, over 954536.31 frames. ], batch size: 65, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:49:31,368 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 16:49:45,240 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.497e+02 1.829e+02 2.293e+02 4.240e+02, threshold=3.659e+02, percent-clipped=2.0 +2023-03-26 16:49:55,179 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 16:49:57,807 INFO [finetune.py:976] (0/7) Epoch 14, batch 1300, loss[loss=0.1706, simple_loss=0.2392, pruned_loss=0.05102, over 4826.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2527, pruned_loss=0.05908, over 955411.56 frames. ], batch size: 33, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:50:31,734 INFO [finetune.py:976] (0/7) Epoch 14, batch 1350, loss[loss=0.2115, simple_loss=0.2822, pruned_loss=0.0704, over 4834.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2525, pruned_loss=0.05929, over 953733.91 frames. ], batch size: 47, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:51:00,988 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 16:51:07,733 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.659e+01 1.636e+02 1.953e+02 2.256e+02 6.748e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 16:51:09,777 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 16:51:19,719 INFO [finetune.py:976] (0/7) Epoch 14, batch 1400, loss[loss=0.1842, simple_loss=0.2651, pruned_loss=0.05158, over 4803.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2574, pruned_loss=0.06092, over 954768.07 frames. ], batch size: 41, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:51:35,804 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:51:53,578 INFO [finetune.py:976] (0/7) Epoch 14, batch 1450, loss[loss=0.1578, simple_loss=0.2329, pruned_loss=0.04132, over 4752.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2589, pruned_loss=0.06088, over 954921.33 frames. ], batch size: 59, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:52:08,070 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75926.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:52:12,660 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75932.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:52:27,769 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.664e+02 1.939e+02 2.609e+02 1.085e+03, threshold=3.877e+02, percent-clipped=5.0 +2023-03-26 16:52:44,458 INFO [finetune.py:976] (0/7) Epoch 14, batch 1500, loss[loss=0.1936, simple_loss=0.2768, pruned_loss=0.05523, over 4911.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.26, pruned_loss=0.06125, over 955650.01 frames. ], batch size: 33, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:52:52,970 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75974.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:53:11,248 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-76000.pt +2023-03-26 16:53:14,325 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-03-26 16:53:19,428 INFO [finetune.py:976] (0/7) Epoch 14, batch 1550, loss[loss=0.17, simple_loss=0.2468, pruned_loss=0.04662, over 4887.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2599, pruned_loss=0.06108, over 956971.85 frames. ], batch size: 32, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:53:32,708 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5720, 2.2026, 2.8281, 1.7024, 2.5762, 2.6902, 2.0808, 2.9974], + device='cuda:0'), covar=tensor([0.1410, 0.1919, 0.1632, 0.2411, 0.0888, 0.1493, 0.2543, 0.0896], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0204, 0.0192, 0.0189, 0.0176, 0.0213, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:53:40,206 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.569e+01 1.489e+02 1.761e+02 2.263e+02 3.823e+02, threshold=3.522e+02, percent-clipped=0.0 +2023-03-26 16:53:42,226 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-26 16:53:53,247 INFO [finetune.py:976] (0/7) Epoch 14, batch 1600, loss[loss=0.1701, simple_loss=0.2376, pruned_loss=0.05132, over 4915.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2573, pruned_loss=0.06059, over 957829.38 frames. ], batch size: 46, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:54:26,637 INFO [finetune.py:976] (0/7) Epoch 14, batch 1650, loss[loss=0.1605, simple_loss=0.2394, pruned_loss=0.04085, over 4791.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2538, pruned_loss=0.05909, over 958702.95 frames. ], batch size: 29, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:54:47,811 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.271e+01 1.580e+02 1.872e+02 2.187e+02 4.946e+02, threshold=3.744e+02, percent-clipped=3.0 +2023-03-26 16:54:50,409 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8912, 1.5378, 2.0038, 1.8145, 1.6873, 1.6147, 1.8169, 1.8547], + device='cuda:0'), covar=tensor([0.3398, 0.3579, 0.2876, 0.3784, 0.4526, 0.3742, 0.4196, 0.2749], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0238, 0.0256, 0.0265, 0.0262, 0.0237, 0.0277, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:55:00,263 INFO [finetune.py:976] (0/7) Epoch 14, batch 1700, loss[loss=0.1881, simple_loss=0.2588, pruned_loss=0.0587, over 4874.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2514, pruned_loss=0.05864, over 957748.96 frames. ], batch size: 34, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:55:34,231 INFO [finetune.py:976] (0/7) Epoch 14, batch 1750, loss[loss=0.2132, simple_loss=0.2683, pruned_loss=0.07902, over 4760.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2534, pruned_loss=0.05926, over 956789.59 frames. ], batch size: 27, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:55:55,253 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.620e+02 1.973e+02 2.349e+02 4.562e+02, threshold=3.945e+02, percent-clipped=4.0 +2023-03-26 16:56:17,742 INFO [finetune.py:976] (0/7) Epoch 14, batch 1800, loss[loss=0.1585, simple_loss=0.2302, pruned_loss=0.04342, over 4751.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2564, pruned_loss=0.0598, over 956584.86 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:56:17,852 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3239, 1.4343, 1.4881, 1.5239, 1.5403, 2.9903, 1.4353, 1.5111], + device='cuda:0'), covar=tensor([0.1051, 0.1810, 0.1134, 0.1036, 0.1643, 0.0298, 0.1471, 0.1743], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0080, 0.0073, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 16:56:27,646 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76269.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:56:47,385 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:56:58,601 INFO [finetune.py:976] (0/7) Epoch 14, batch 1850, loss[loss=0.1564, simple_loss=0.2165, pruned_loss=0.04817, over 4716.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2591, pruned_loss=0.06061, over 957858.25 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:57:06,638 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76323.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:57:11,378 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76330.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 16:57:19,099 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.514e+02 1.907e+02 2.299e+02 3.483e+02, threshold=3.815e+02, percent-clipped=0.0 +2023-03-26 16:57:35,298 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:57:38,872 INFO [finetune.py:976] (0/7) Epoch 14, batch 1900, loss[loss=0.1488, simple_loss=0.2126, pruned_loss=0.04251, over 4372.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.26, pruned_loss=0.06068, over 957314.48 frames. ], batch size: 19, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:57:52,878 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3009, 1.9449, 1.5021, 0.7301, 1.9232, 1.7011, 1.3119, 1.7771], + device='cuda:0'), covar=tensor([0.0804, 0.1171, 0.2041, 0.2539, 0.1634, 0.2582, 0.3129, 0.1306], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0196, 0.0198, 0.0185, 0.0213, 0.0208, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:57:57,163 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76384.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:58:15,527 INFO [finetune.py:976] (0/7) Epoch 14, batch 1950, loss[loss=0.1558, simple_loss=0.23, pruned_loss=0.0408, over 4823.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2593, pruned_loss=0.06031, over 957014.19 frames. ], batch size: 39, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:58:35,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.938e+01 1.459e+02 1.786e+02 2.082e+02 3.715e+02, threshold=3.572e+02, percent-clipped=0.0 +2023-03-26 16:58:38,809 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8179, 1.7818, 1.7096, 1.9950, 2.2168, 1.9632, 1.6249, 1.4836], + device='cuda:0'), covar=tensor([0.2101, 0.1907, 0.1699, 0.1507, 0.1681, 0.1155, 0.2239, 0.1884], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0206, 0.0209, 0.0189, 0.0239, 0.0184, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 16:58:49,143 INFO [finetune.py:976] (0/7) Epoch 14, batch 2000, loss[loss=0.1729, simple_loss=0.2381, pruned_loss=0.0538, over 4817.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2558, pruned_loss=0.05979, over 956498.84 frames. ], batch size: 39, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:59:22,662 INFO [finetune.py:976] (0/7) Epoch 14, batch 2050, loss[loss=0.1682, simple_loss=0.2342, pruned_loss=0.05112, over 4765.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2519, pruned_loss=0.05891, over 954625.70 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:59:28,804 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6011, 1.1618, 2.0068, 3.1772, 2.1165, 2.4066, 0.8601, 2.7094], + device='cuda:0'), covar=tensor([0.1939, 0.2174, 0.1516, 0.0903, 0.0967, 0.1527, 0.2316, 0.0629], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0164, 0.0100, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:59:42,964 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.833e+01 1.460e+02 1.798e+02 2.157e+02 5.136e+02, threshold=3.595e+02, percent-clipped=3.0 +2023-03-26 16:59:46,691 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6420, 1.4461, 2.2151, 3.2383, 2.2453, 2.3288, 1.0021, 2.5855], + device='cuda:0'), covar=tensor([0.1669, 0.1403, 0.1178, 0.0616, 0.0779, 0.1848, 0.1727, 0.0593], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0164, 0.0100, 0.0138, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 16:59:56,036 INFO [finetune.py:976] (0/7) Epoch 14, batch 2100, loss[loss=0.173, simple_loss=0.2463, pruned_loss=0.04981, over 4916.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2513, pruned_loss=0.05897, over 955419.92 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:00:29,580 INFO [finetune.py:976] (0/7) Epoch 14, batch 2150, loss[loss=0.1678, simple_loss=0.2491, pruned_loss=0.04326, over 4801.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2552, pruned_loss=0.06063, over 953442.02 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:00:38,762 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76625.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:00:50,408 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.680e+02 1.855e+02 2.274e+02 3.771e+02, threshold=3.710e+02, percent-clipped=2.0 +2023-03-26 17:00:55,365 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:01:02,484 INFO [finetune.py:976] (0/7) Epoch 14, batch 2200, loss[loss=0.1889, simple_loss=0.2503, pruned_loss=0.06375, over 4743.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2572, pruned_loss=0.06111, over 954593.56 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:01:21,648 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76679.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:01:26,363 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0587, 0.9763, 0.9219, 1.1877, 1.2616, 1.1688, 0.9966, 0.9247], + device='cuda:0'), covar=tensor([0.0345, 0.0328, 0.0596, 0.0297, 0.0290, 0.0483, 0.0342, 0.0440], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0110, 0.0142, 0.0114, 0.0101, 0.0106, 0.0097, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3330e-05, 8.4926e-05, 1.1228e-04, 8.8466e-05, 7.8720e-05, 7.8672e-05, + 7.2835e-05, 8.3036e-05], device='cuda:0') +2023-03-26 17:01:36,046 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5795, 2.6038, 2.3335, 2.6969, 2.3492, 5.0108, 2.4504, 2.9041], + device='cuda:0'), covar=tensor([0.2824, 0.1936, 0.1751, 0.1851, 0.1241, 0.0094, 0.1941, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:01:57,531 INFO [finetune.py:976] (0/7) Epoch 14, batch 2250, loss[loss=0.1787, simple_loss=0.2327, pruned_loss=0.06238, over 4717.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2575, pruned_loss=0.06126, over 953178.52 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:02:03,743 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1186, 3.5910, 3.7464, 4.0062, 3.8601, 3.6493, 4.2410, 1.4644], + device='cuda:0'), covar=tensor([0.0871, 0.0906, 0.1065, 0.1074, 0.1376, 0.1645, 0.0743, 0.5383], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0277, 0.0292, 0.0331, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:02:04,497 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 17:02:11,962 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 17:02:18,736 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.580e+02 1.848e+02 2.151e+02 3.368e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 17:02:31,272 INFO [finetune.py:976] (0/7) Epoch 14, batch 2300, loss[loss=0.171, simple_loss=0.2382, pruned_loss=0.05185, over 4898.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2582, pruned_loss=0.06118, over 951462.81 frames. ], batch size: 36, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:02:43,801 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1584, 1.9983, 1.8517, 2.0372, 1.5606, 4.6807, 1.8438, 2.2768], + device='cuda:0'), covar=tensor([0.3034, 0.2214, 0.1908, 0.2165, 0.1556, 0.0116, 0.2207, 0.1172], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0116, 0.0121, 0.0124, 0.0116, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:03:06,756 INFO [finetune.py:976] (0/7) Epoch 14, batch 2350, loss[loss=0.1912, simple_loss=0.2655, pruned_loss=0.05847, over 4737.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2556, pruned_loss=0.06012, over 952396.27 frames. ], batch size: 54, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:03:24,102 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6206, 3.9452, 4.1925, 4.4215, 4.3858, 4.0203, 4.6607, 1.5221], + device='cuda:0'), covar=tensor([0.0678, 0.0870, 0.0863, 0.0927, 0.1157, 0.1522, 0.0617, 0.5469], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0277, 0.0293, 0.0332, 0.0282, 0.0302, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:03:25,968 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76839.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:03:28,175 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.531e+02 1.863e+02 2.217e+02 4.521e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-03-26 17:03:40,657 INFO [finetune.py:976] (0/7) Epoch 14, batch 2400, loss[loss=0.1797, simple_loss=0.2444, pruned_loss=0.05754, over 4831.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2525, pruned_loss=0.05877, over 954283.32 frames. ], batch size: 40, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:04:06,927 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:04:13,502 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8969, 1.3506, 0.8028, 1.6454, 2.1218, 1.5485, 1.6083, 1.6582], + device='cuda:0'), covar=tensor([0.1320, 0.1964, 0.2127, 0.1173, 0.1935, 0.1954, 0.1384, 0.1835], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0112, 0.0091, 0.0118, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:04:14,026 INFO [finetune.py:976] (0/7) Epoch 14, batch 2450, loss[loss=0.1721, simple_loss=0.238, pruned_loss=0.05304, over 4822.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2515, pruned_loss=0.05906, over 954542.11 frames. ], batch size: 39, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:04:23,091 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76925.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:04:34,661 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.650e+01 1.627e+02 1.914e+02 2.448e+02 4.488e+02, threshold=3.829e+02, percent-clipped=3.0 +2023-03-26 17:04:40,093 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:04:45,427 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-03-26 17:04:47,675 INFO [finetune.py:976] (0/7) Epoch 14, batch 2500, loss[loss=0.2343, simple_loss=0.3001, pruned_loss=0.08427, over 4858.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.254, pruned_loss=0.06052, over 955192.25 frames. ], batch size: 44, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:04:55,523 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76973.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:04:59,663 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76979.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:05:12,263 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76998.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:05:21,746 INFO [finetune.py:976] (0/7) Epoch 14, batch 2550, loss[loss=0.205, simple_loss=0.2724, pruned_loss=0.06879, over 4806.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2573, pruned_loss=0.06111, over 954446.48 frames. ], batch size: 41, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:05:31,461 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9776, 1.6194, 2.3365, 1.5416, 2.0784, 2.1747, 1.6291, 2.4395], + device='cuda:0'), covar=tensor([0.1333, 0.2086, 0.1332, 0.2061, 0.0794, 0.1442, 0.2680, 0.0722], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0205, 0.0192, 0.0190, 0.0177, 0.0214, 0.0217, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:05:32,004 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77027.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:05:42,423 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.599e+02 1.863e+02 2.531e+02 5.028e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-03-26 17:05:55,388 INFO [finetune.py:976] (0/7) Epoch 14, batch 2600, loss[loss=0.1974, simple_loss=0.2685, pruned_loss=0.06318, over 4816.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2581, pruned_loss=0.06104, over 955153.66 frames. ], batch size: 25, lr: 3.55e-03, grad_scale: 32.0 +2023-03-26 17:06:18,018 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77095.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:06:30,487 INFO [finetune.py:976] (0/7) Epoch 14, batch 2650, loss[loss=0.2015, simple_loss=0.2694, pruned_loss=0.06675, over 4853.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2587, pruned_loss=0.06109, over 954449.65 frames. ], batch size: 31, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:07:08,844 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.606e+02 1.948e+02 2.371e+02 3.624e+02, threshold=3.895e+02, percent-clipped=0.0 +2023-03-26 17:07:17,034 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 17:07:22,909 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77156.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:07:24,146 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9331, 1.3010, 1.8922, 1.8027, 1.6210, 1.5538, 1.7799, 1.7417], + device='cuda:0'), covar=tensor([0.3757, 0.3910, 0.3364, 0.3754, 0.4861, 0.3816, 0.4315, 0.3274], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0239, 0.0257, 0.0266, 0.0264, 0.0237, 0.0278, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:07:26,328 INFO [finetune.py:976] (0/7) Epoch 14, batch 2700, loss[loss=0.1714, simple_loss=0.2331, pruned_loss=0.05485, over 4689.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2566, pruned_loss=0.05989, over 954434.53 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:07:50,444 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 17:07:55,982 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8589, 2.4831, 2.0021, 1.0540, 2.3062, 2.3413, 1.9490, 2.2179], + device='cuda:0'), covar=tensor([0.0653, 0.0750, 0.1227, 0.1725, 0.1209, 0.1578, 0.1827, 0.0875], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0197, 0.0201, 0.0186, 0.0215, 0.0210, 0.0225, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:07:57,703 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:07,956 INFO [finetune.py:976] (0/7) Epoch 14, batch 2750, loss[loss=0.1454, simple_loss=0.2166, pruned_loss=0.03705, over 4703.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.254, pruned_loss=0.05929, over 953174.09 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:08:08,062 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:28,394 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.514e+02 1.871e+02 2.163e+02 3.576e+02, threshold=3.742e+02, percent-clipped=0.0 +2023-03-26 17:08:40,953 INFO [finetune.py:976] (0/7) Epoch 14, batch 2800, loss[loss=0.2435, simple_loss=0.2747, pruned_loss=0.1061, over 4017.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2503, pruned_loss=0.05784, over 952344.50 frames. ], batch size: 17, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:08:48,257 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:48,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:52,186 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77277.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:14,627 INFO [finetune.py:976] (0/7) Epoch 14, batch 2850, loss[loss=0.1345, simple_loss=0.2062, pruned_loss=0.03136, over 4750.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2492, pruned_loss=0.05779, over 954010.94 frames. ], batch size: 26, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:09:19,616 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3934, 1.4496, 1.6816, 1.7610, 1.5597, 3.3037, 1.2861, 1.5527], + device='cuda:0'), covar=tensor([0.1027, 0.1814, 0.1123, 0.0966, 0.1646, 0.0238, 0.1527, 0.1737], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0080, 0.0073, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:09:26,882 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 17:09:29,753 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77334.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:32,166 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77338.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:34,912 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.563e+02 1.884e+02 2.320e+02 5.201e+02, threshold=3.768e+02, percent-clipped=2.0 +2023-03-26 17:09:47,966 INFO [finetune.py:976] (0/7) Epoch 14, batch 2900, loss[loss=0.1833, simple_loss=0.2478, pruned_loss=0.05934, over 4758.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2527, pruned_loss=0.05949, over 953101.22 frames. ], batch size: 26, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:14,874 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3135, 2.3312, 1.9339, 2.4672, 2.1901, 2.1776, 2.1916, 3.1097], + device='cuda:0'), covar=tensor([0.3913, 0.5094, 0.3482, 0.4426, 0.4272, 0.2573, 0.4330, 0.1566], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0226, 0.0279, 0.0248, 0.0215, 0.0250, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:10:21,779 INFO [finetune.py:976] (0/7) Epoch 14, batch 2950, loss[loss=0.2186, simple_loss=0.2798, pruned_loss=0.0787, over 4918.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2562, pruned_loss=0.06085, over 950425.23 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:23,124 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1925, 2.2114, 1.6509, 2.2228, 2.1327, 1.7708, 2.5210, 2.2303], + device='cuda:0'), covar=tensor([0.1350, 0.2114, 0.3157, 0.2701, 0.2667, 0.1776, 0.3191, 0.1772], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0187, 0.0232, 0.0252, 0.0243, 0.0198, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:10:41,985 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.681e+01 1.651e+02 1.924e+02 2.203e+02 4.754e+02, threshold=3.848e+02, percent-clipped=2.0 +2023-03-26 17:10:48,021 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77451.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:10:50,520 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9895, 1.4080, 2.0068, 2.0045, 1.8014, 1.6936, 1.8862, 1.8100], + device='cuda:0'), covar=tensor([0.3758, 0.4202, 0.3434, 0.3803, 0.4816, 0.3678, 0.4688, 0.3340], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0238, 0.0256, 0.0265, 0.0263, 0.0236, 0.0277, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:10:54,978 INFO [finetune.py:976] (0/7) Epoch 14, batch 3000, loss[loss=0.2013, simple_loss=0.2756, pruned_loss=0.06352, over 4865.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2571, pruned_loss=0.06095, over 952108.69 frames. ], batch size: 34, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:54,979 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 17:11:07,752 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8499, 3.3981, 3.5126, 3.7431, 3.5978, 3.4467, 3.9383, 1.2840], + device='cuda:0'), covar=tensor([0.0878, 0.1036, 0.0971, 0.0941, 0.1528, 0.1747, 0.0861, 0.5219], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0243, 0.0276, 0.0292, 0.0332, 0.0283, 0.0299, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:11:09,361 INFO [finetune.py:1010] (0/7) Epoch 14, validation: loss=0.1563, simple_loss=0.2268, pruned_loss=0.04293, over 2265189.00 frames. +2023-03-26 17:11:09,361 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 17:11:34,064 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:11:44,250 INFO [finetune.py:976] (0/7) Epoch 14, batch 3050, loss[loss=0.1984, simple_loss=0.2822, pruned_loss=0.05729, over 4879.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2572, pruned_loss=0.06052, over 953390.75 frames. ], batch size: 32, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:12:13,224 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.567e+02 1.800e+02 2.244e+02 5.193e+02, threshold=3.600e+02, percent-clipped=2.0 +2023-03-26 17:12:13,924 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:12:35,684 INFO [finetune.py:976] (0/7) Epoch 14, batch 3100, loss[loss=0.1734, simple_loss=0.2426, pruned_loss=0.05208, over 4927.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2542, pruned_loss=0.05883, over 954339.39 frames. ], batch size: 33, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:12:39,896 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:16,737 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-26 17:13:22,406 INFO [finetune.py:976] (0/7) Epoch 14, batch 3150, loss[loss=0.1782, simple_loss=0.2454, pruned_loss=0.05551, over 4833.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2526, pruned_loss=0.05879, over 955759.73 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:13:25,562 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:35,443 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:36,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4250, 1.4780, 1.6126, 1.7129, 1.5569, 3.3344, 1.3347, 1.5988], + device='cuda:0'), covar=tensor([0.0985, 0.1813, 0.1010, 0.0957, 0.1556, 0.0218, 0.1472, 0.1669], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:13:37,875 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77633.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:43,292 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.643e+02 1.968e+02 2.398e+02 4.679e+02, threshold=3.936e+02, percent-clipped=3.0 +2023-03-26 17:13:56,370 INFO [finetune.py:976] (0/7) Epoch 14, batch 3200, loss[loss=0.1777, simple_loss=0.2416, pruned_loss=0.05686, over 4795.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2502, pruned_loss=0.05823, over 955917.20 frames. ], batch size: 51, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:14:06,641 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:06,652 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:29,511 INFO [finetune.py:976] (0/7) Epoch 14, batch 3250, loss[loss=0.2077, simple_loss=0.2701, pruned_loss=0.07269, over 4929.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2527, pruned_loss=0.05981, over 958072.54 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:14:47,325 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:49,600 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.623e+02 1.885e+02 2.287e+02 7.301e+02, threshold=3.769e+02, percent-clipped=4.0 +2023-03-26 17:14:53,335 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9592, 1.8771, 1.9640, 1.4051, 1.9191, 2.1575, 2.0703, 1.6488], + device='cuda:0'), covar=tensor([0.0633, 0.0687, 0.0707, 0.0895, 0.0691, 0.0662, 0.0646, 0.1194], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0132, 0.0141, 0.0123, 0.0123, 0.0140, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:14:55,598 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77751.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:15:02,067 INFO [finetune.py:976] (0/7) Epoch 14, batch 3300, loss[loss=0.1808, simple_loss=0.2469, pruned_loss=0.05732, over 4822.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2547, pruned_loss=0.06032, over 954062.14 frames. ], batch size: 30, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:15:27,677 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:15:35,621 INFO [finetune.py:976] (0/7) Epoch 14, batch 3350, loss[loss=0.2213, simple_loss=0.2731, pruned_loss=0.08473, over 4796.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2575, pruned_loss=0.06125, over 955146.75 frames. ], batch size: 45, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:15:40,892 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6253, 3.6944, 3.5009, 1.7857, 3.6606, 2.8900, 0.7316, 2.5100], + device='cuda:0'), covar=tensor([0.2617, 0.1502, 0.1311, 0.2925, 0.0997, 0.0920, 0.4224, 0.1338], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0172, 0.0159, 0.0127, 0.0156, 0.0121, 0.0144, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 17:15:57,262 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.683e+02 1.958e+02 2.277e+02 5.309e+02, threshold=3.915e+02, percent-clipped=1.0 +2023-03-26 17:16:09,330 INFO [finetune.py:976] (0/7) Epoch 14, batch 3400, loss[loss=0.2159, simple_loss=0.2878, pruned_loss=0.07199, over 4927.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2593, pruned_loss=0.06218, over 955703.58 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:16:18,543 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77867.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:16:34,989 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5565, 1.3208, 1.8256, 3.2484, 2.1677, 2.0417, 0.7367, 2.5804], + device='cuda:0'), covar=tensor([0.1910, 0.1676, 0.1496, 0.0579, 0.0822, 0.1555, 0.2193, 0.0601], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0164, 0.0101, 0.0137, 0.0125, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:16:40,651 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7538, 1.2907, 0.8368, 1.5965, 2.0428, 1.2246, 1.4312, 1.5234], + device='cuda:0'), covar=tensor([0.1465, 0.2075, 0.2058, 0.1285, 0.1958, 0.2000, 0.1530, 0.2016], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:16:43,712 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5703, 1.7281, 2.1387, 1.8675, 1.8656, 4.4243, 1.7469, 2.0025], + device='cuda:0'), covar=tensor([0.0980, 0.1642, 0.1218, 0.1038, 0.1475, 0.0148, 0.1308, 0.1547], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:16:51,414 INFO [finetune.py:976] (0/7) Epoch 14, batch 3450, loss[loss=0.1628, simple_loss=0.2334, pruned_loss=0.04607, over 4864.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2591, pruned_loss=0.0616, over 955441.10 frames. ], batch size: 31, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:16:53,842 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:16:59,131 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8551, 1.6980, 1.4780, 1.9012, 2.2128, 1.9083, 1.4738, 1.4690], + device='cuda:0'), covar=tensor([0.1989, 0.1891, 0.1809, 0.1433, 0.1560, 0.1120, 0.2341, 0.1847], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0207, 0.0209, 0.0190, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:17:03,226 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:05,731 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:10,907 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0804, 1.9521, 2.3844, 1.6966, 2.1767, 2.5150, 1.9036, 2.6060], + device='cuda:0'), covar=tensor([0.1375, 0.1963, 0.1515, 0.1862, 0.1019, 0.1303, 0.2386, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0204, 0.0192, 0.0190, 0.0176, 0.0214, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:17:11,407 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.160e+01 1.485e+02 1.825e+02 2.093e+02 4.049e+02, threshold=3.650e+02, percent-clipped=1.0 +2023-03-26 17:17:33,221 INFO [finetune.py:976] (0/7) Epoch 14, batch 3500, loss[loss=0.217, simple_loss=0.2716, pruned_loss=0.08123, over 4213.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2573, pruned_loss=0.06125, over 955484.18 frames. ], batch size: 65, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:17:40,922 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:44,501 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:46,351 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4456, 3.5033, 3.2436, 1.5029, 3.4918, 2.5514, 0.8378, 2.2751], + device='cuda:0'), covar=tensor([0.2460, 0.2213, 0.1828, 0.3725, 0.1289, 0.1205, 0.4522, 0.1860], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0173, 0.0160, 0.0127, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 17:17:46,954 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:18:06,487 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-78000.pt +2023-03-26 17:18:20,965 INFO [finetune.py:976] (0/7) Epoch 14, batch 3550, loss[loss=0.1884, simple_loss=0.2481, pruned_loss=0.0643, over 4937.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2547, pruned_loss=0.06015, over 956743.32 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:18:28,308 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2272, 2.0147, 2.0947, 0.9737, 2.2841, 2.5482, 2.2370, 1.8961], + device='cuda:0'), covar=tensor([0.0842, 0.0720, 0.0551, 0.0698, 0.0557, 0.0526, 0.0398, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0154, 0.0123, 0.0131, 0.0131, 0.0128, 0.0143, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.4085e-05, 1.1219e-04, 8.8343e-05, 9.3823e-05, 9.3014e-05, 9.2521e-05, + 1.0405e-04, 1.0646e-04], device='cuda:0') +2023-03-26 17:18:29,454 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5863, 2.1629, 2.8194, 4.3559, 3.1055, 2.7793, 0.9197, 3.6436], + device='cuda:0'), covar=tensor([0.1427, 0.1271, 0.1258, 0.0476, 0.0647, 0.1561, 0.2167, 0.0480], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:18:31,781 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 17:18:34,114 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3942, 2.2321, 1.9524, 2.1955, 2.1568, 2.1476, 2.1363, 2.8016], + device='cuda:0'), covar=tensor([0.4263, 0.4651, 0.3827, 0.4114, 0.3955, 0.2787, 0.4224, 0.2147], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0226, 0.0277, 0.0247, 0.0213, 0.0249, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:18:35,856 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78033.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:18:41,165 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.505e+02 1.960e+02 2.472e+02 4.194e+02, threshold=3.920e+02, percent-clipped=4.0 +2023-03-26 17:18:42,560 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8297, 1.6915, 1.4430, 1.4183, 1.8700, 1.5688, 1.8656, 1.8338], + device='cuda:0'), covar=tensor([0.1434, 0.2013, 0.3269, 0.2428, 0.2702, 0.1806, 0.2893, 0.1796], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0188, 0.0234, 0.0253, 0.0244, 0.0199, 0.0212, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:18:54,340 INFO [finetune.py:976] (0/7) Epoch 14, batch 3600, loss[loss=0.1983, simple_loss=0.2655, pruned_loss=0.06554, over 4910.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.253, pruned_loss=0.05979, over 956696.45 frames. ], batch size: 37, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:19:28,413 INFO [finetune.py:976] (0/7) Epoch 14, batch 3650, loss[loss=0.166, simple_loss=0.2396, pruned_loss=0.04617, over 4769.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2539, pruned_loss=0.05978, over 956649.92 frames. ], batch size: 28, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:19:48,729 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.711e+02 2.033e+02 2.406e+02 8.151e+02, threshold=4.067e+02, percent-clipped=4.0 +2023-03-26 17:20:00,505 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7519, 1.7137, 2.2221, 1.4207, 1.8720, 2.1591, 1.6427, 2.3201], + device='cuda:0'), covar=tensor([0.1710, 0.2328, 0.1558, 0.2216, 0.1058, 0.1701, 0.2937, 0.0914], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0206, 0.0193, 0.0192, 0.0178, 0.0215, 0.0218, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:20:02,236 INFO [finetune.py:976] (0/7) Epoch 14, batch 3700, loss[loss=0.2493, simple_loss=0.3169, pruned_loss=0.09085, over 4840.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.258, pruned_loss=0.06128, over 955798.14 frames. ], batch size: 49, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:20:35,984 INFO [finetune.py:976] (0/7) Epoch 14, batch 3750, loss[loss=0.192, simple_loss=0.2693, pruned_loss=0.05734, over 4812.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2591, pruned_loss=0.06134, over 953915.01 frames. ], batch size: 39, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:20:49,443 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:20:55,808 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.631e+02 1.949e+02 2.239e+02 4.423e+02, threshold=3.899e+02, percent-clipped=1.0 +2023-03-26 17:21:08,215 INFO [finetune.py:976] (0/7) Epoch 14, batch 3800, loss[loss=0.1915, simple_loss=0.2635, pruned_loss=0.05975, over 4827.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2599, pruned_loss=0.06172, over 951234.29 frames. ], batch size: 39, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:21:12,925 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 17:21:15,892 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:21:21,440 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 17:21:31,057 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78293.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:21:41,080 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5043, 2.2639, 1.6942, 0.8128, 2.0223, 2.0213, 1.8552, 1.9615], + device='cuda:0'), covar=tensor([0.0793, 0.0681, 0.1306, 0.1888, 0.1132, 0.2049, 0.1834, 0.0807], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0195, 0.0199, 0.0183, 0.0213, 0.0207, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:21:42,318 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8686, 1.2252, 1.8351, 1.8428, 1.6502, 1.5644, 1.7839, 1.6943], + device='cuda:0'), covar=tensor([0.3655, 0.3850, 0.3251, 0.3478, 0.4456, 0.3504, 0.4215, 0.3001], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0239, 0.0257, 0.0266, 0.0263, 0.0237, 0.0278, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:21:49,333 INFO [finetune.py:976] (0/7) Epoch 14, batch 3850, loss[loss=0.1894, simple_loss=0.2634, pruned_loss=0.05774, over 4824.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2597, pruned_loss=0.06167, over 952067.90 frames. ], batch size: 47, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:21:55,862 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:22:02,008 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0568, 1.9767, 2.3798, 1.5575, 2.0842, 2.3933, 1.8759, 2.5724], + device='cuda:0'), covar=tensor([0.1269, 0.1813, 0.1363, 0.2216, 0.1012, 0.1437, 0.2364, 0.0756], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0205, 0.0190, 0.0190, 0.0176, 0.0213, 0.0216, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:22:03,794 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:22:04,449 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6456, 1.4090, 1.1596, 1.2537, 1.9087, 1.8876, 1.5923, 1.3641], + device='cuda:0'), covar=tensor([0.0287, 0.0355, 0.0799, 0.0414, 0.0234, 0.0417, 0.0329, 0.0415], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0110, 0.0143, 0.0115, 0.0102, 0.0108, 0.0097, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.3385e-05, 8.5397e-05, 1.1327e-04, 8.9195e-05, 7.9247e-05, 7.9678e-05, + 7.3248e-05, 8.3766e-05], device='cuda:0') +2023-03-26 17:22:10,190 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.518e+02 1.784e+02 2.158e+02 4.566e+02, threshold=3.568e+02, percent-clipped=2.0 +2023-03-26 17:22:22,702 INFO [finetune.py:976] (0/7) Epoch 14, batch 3900, loss[loss=0.1801, simple_loss=0.2546, pruned_loss=0.05284, over 4827.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2562, pruned_loss=0.06017, over 953815.52 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:22:26,611 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-03-26 17:22:45,628 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:23:09,952 INFO [finetune.py:976] (0/7) Epoch 14, batch 3950, loss[loss=0.1917, simple_loss=0.2586, pruned_loss=0.06236, over 4906.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2541, pruned_loss=0.05993, over 954935.24 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:23:37,906 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.587e+02 1.893e+02 2.259e+02 3.905e+02, threshold=3.786e+02, percent-clipped=1.0 +2023-03-26 17:23:50,387 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78460.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:23:50,892 INFO [finetune.py:976] (0/7) Epoch 14, batch 4000, loss[loss=0.2212, simple_loss=0.2822, pruned_loss=0.08008, over 4775.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2531, pruned_loss=0.05968, over 954613.28 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:24:07,753 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-26 17:24:24,848 INFO [finetune.py:976] (0/7) Epoch 14, batch 4050, loss[loss=0.2477, simple_loss=0.2956, pruned_loss=0.09985, over 4931.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2574, pruned_loss=0.06197, over 955823.55 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:24:31,536 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:24:45,460 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.285e+01 1.581e+02 1.919e+02 2.315e+02 3.488e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-03-26 17:24:54,867 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:24:57,450 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 17:24:57,772 INFO [finetune.py:976] (0/7) Epoch 14, batch 4100, loss[loss=0.1832, simple_loss=0.2524, pruned_loss=0.05699, over 4902.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2582, pruned_loss=0.06179, over 955179.78 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:24:59,690 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2895, 1.9932, 2.0560, 0.9944, 2.3172, 2.4825, 2.1392, 1.9508], + device='cuda:0'), covar=tensor([0.0802, 0.0689, 0.0428, 0.0718, 0.0516, 0.0519, 0.0416, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0122, 0.0129, 0.0129, 0.0126, 0.0141, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.2497e-05, 1.1054e-04, 8.7660e-05, 9.2549e-05, 9.1358e-05, 9.0978e-05, + 1.0252e-04, 1.0445e-04], device='cuda:0') +2023-03-26 17:25:16,497 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78588.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:25:29,390 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 17:25:31,551 INFO [finetune.py:976] (0/7) Epoch 14, batch 4150, loss[loss=0.2147, simple_loss=0.2723, pruned_loss=0.07856, over 4827.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2594, pruned_loss=0.06214, over 955204.36 frames. ], batch size: 30, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:25:35,311 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78617.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:25:52,377 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.587e+02 1.869e+02 2.218e+02 3.242e+02, threshold=3.739e+02, percent-clipped=0.0 +2023-03-26 17:26:03,280 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4771, 1.4731, 1.7273, 1.7072, 1.6042, 3.3040, 1.4050, 1.5290], + device='cuda:0'), covar=tensor([0.1243, 0.2359, 0.1271, 0.1189, 0.1956, 0.0324, 0.1942, 0.2318], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0080, 0.0073, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:26:04,875 INFO [finetune.py:976] (0/7) Epoch 14, batch 4200, loss[loss=0.1798, simple_loss=0.2653, pruned_loss=0.04714, over 4738.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2605, pruned_loss=0.06192, over 954900.64 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:26:37,997 INFO [finetune.py:976] (0/7) Epoch 14, batch 4250, loss[loss=0.1788, simple_loss=0.2381, pruned_loss=0.05977, over 4904.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2589, pruned_loss=0.06158, over 954525.56 frames. ], batch size: 36, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:27:05,885 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.501e+02 1.722e+02 2.085e+02 5.543e+02, threshold=3.444e+02, percent-clipped=2.0 +2023-03-26 17:27:21,259 INFO [finetune.py:976] (0/7) Epoch 14, batch 4300, loss[loss=0.1747, simple_loss=0.2355, pruned_loss=0.05694, over 4915.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2546, pruned_loss=0.0594, over 954692.46 frames. ], batch size: 43, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:27:59,986 INFO [finetune.py:976] (0/7) Epoch 14, batch 4350, loss[loss=0.187, simple_loss=0.2544, pruned_loss=0.0598, over 4850.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2518, pruned_loss=0.05898, over 954400.78 frames. ], batch size: 47, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:28:06,791 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:28:34,427 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.179e+02 1.620e+02 1.919e+02 2.276e+02 4.946e+02, threshold=3.838e+02, percent-clipped=3.0 +2023-03-26 17:28:54,136 INFO [finetune.py:976] (0/7) Epoch 14, batch 4400, loss[loss=0.1875, simple_loss=0.2649, pruned_loss=0.05509, over 4904.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2521, pruned_loss=0.05901, over 954687.12 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:28:59,037 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 17:29:11,521 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 17:29:12,591 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:27,939 INFO [finetune.py:976] (0/7) Epoch 14, batch 4450, loss[loss=0.2045, simple_loss=0.2736, pruned_loss=0.0677, over 4767.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2548, pruned_loss=0.05946, over 955551.74 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:29:28,596 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78912.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:34,555 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8458, 1.5750, 2.2883, 3.7996, 2.6028, 2.5154, 0.8067, 3.0503], + device='cuda:0'), covar=tensor([0.1871, 0.1620, 0.1424, 0.0507, 0.0805, 0.1905, 0.2158, 0.0516], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:29:44,121 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:48,694 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.623e+02 2.075e+02 2.454e+02 4.700e+02, threshold=4.150e+02, percent-clipped=3.0 +2023-03-26 17:30:01,641 INFO [finetune.py:976] (0/7) Epoch 14, batch 4500, loss[loss=0.2165, simple_loss=0.2876, pruned_loss=0.07274, over 4817.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2572, pruned_loss=0.06073, over 955301.09 frames. ], batch size: 47, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:30:04,117 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78965.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:30:34,871 INFO [finetune.py:976] (0/7) Epoch 14, batch 4550, loss[loss=0.1804, simple_loss=0.255, pruned_loss=0.05293, over 4738.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2589, pruned_loss=0.06133, over 954658.67 frames. ], batch size: 54, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:30:44,021 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79026.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:30:54,526 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.627e+02 1.835e+02 2.182e+02 4.419e+02, threshold=3.671e+02, percent-clipped=1.0 +2023-03-26 17:31:04,493 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5968, 1.4371, 1.3705, 1.4936, 1.8595, 1.7270, 1.5610, 1.3278], + device='cuda:0'), covar=tensor([0.0290, 0.0273, 0.0594, 0.0295, 0.0193, 0.0405, 0.0289, 0.0390], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0109, 0.0142, 0.0114, 0.0101, 0.0107, 0.0097, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3440e-05, 8.4511e-05, 1.1280e-04, 8.8825e-05, 7.8897e-05, 7.9441e-05, + 7.2740e-05, 8.3393e-05], device='cuda:0') +2023-03-26 17:31:08,635 INFO [finetune.py:976] (0/7) Epoch 14, batch 4600, loss[loss=0.1708, simple_loss=0.234, pruned_loss=0.05385, over 4923.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2575, pruned_loss=0.06019, over 954812.30 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 64.0 +2023-03-26 17:31:42,445 INFO [finetune.py:976] (0/7) Epoch 14, batch 4650, loss[loss=0.1328, simple_loss=0.214, pruned_loss=0.02576, over 4781.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2552, pruned_loss=0.06005, over 953658.63 frames. ], batch size: 29, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:31:45,572 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:32:02,947 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.520e+02 1.886e+02 2.415e+02 4.094e+02, threshold=3.771e+02, percent-clipped=1.0 +2023-03-26 17:32:24,492 INFO [finetune.py:976] (0/7) Epoch 14, batch 4700, loss[loss=0.1796, simple_loss=0.2301, pruned_loss=0.06456, over 4834.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2515, pruned_loss=0.05887, over 955025.38 frames. ], batch size: 40, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:32:26,305 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:32:34,962 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0882, 1.9983, 1.6815, 1.6512, 2.0703, 1.7997, 2.2044, 2.1488], + device='cuda:0'), covar=tensor([0.1306, 0.1956, 0.2979, 0.2702, 0.2625, 0.1566, 0.3467, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0186, 0.0233, 0.0252, 0.0242, 0.0198, 0.0211, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:32:58,004 INFO [finetune.py:976] (0/7) Epoch 14, batch 4750, loss[loss=0.2266, simple_loss=0.2873, pruned_loss=0.08299, over 4873.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2495, pruned_loss=0.05828, over 953808.33 frames. ], batch size: 34, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:32:59,202 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79212.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:33:04,259 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 17:33:32,021 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.558e+01 1.654e+02 1.996e+02 2.359e+02 6.861e+02, threshold=3.993e+02, percent-clipped=2.0 +2023-03-26 17:33:50,652 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:33:51,675 INFO [finetune.py:976] (0/7) Epoch 14, batch 4800, loss[loss=0.1618, simple_loss=0.2396, pruned_loss=0.04201, over 4875.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2521, pruned_loss=0.05958, over 952686.59 frames. ], batch size: 31, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:33:54,333 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-26 17:34:10,286 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79285.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:34:27,364 INFO [finetune.py:976] (0/7) Epoch 14, batch 4850, loss[loss=0.1472, simple_loss=0.2321, pruned_loss=0.03116, over 4895.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2556, pruned_loss=0.06019, over 953226.97 frames. ], batch size: 43, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:34:35,464 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:34:42,723 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:34:49,131 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.639e+02 1.937e+02 2.312e+02 4.640e+02, threshold=3.873e+02, percent-clipped=1.0 +2023-03-26 17:34:51,081 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79346.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:34:58,149 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8054, 3.6167, 3.4083, 1.5850, 3.7392, 2.8330, 0.9615, 2.5213], + device='cuda:0'), covar=tensor([0.2606, 0.1538, 0.1609, 0.3293, 0.0908, 0.0948, 0.4054, 0.1392], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0172, 0.0159, 0.0126, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 17:35:00,513 INFO [finetune.py:976] (0/7) Epoch 14, batch 4900, loss[loss=0.1632, simple_loss=0.2324, pruned_loss=0.04696, over 4205.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2578, pruned_loss=0.06106, over 951167.83 frames. ], batch size: 66, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:35:03,433 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79364.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:11,516 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79375.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:35:11,625 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 17:35:22,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0358, 2.0583, 2.2030, 1.4963, 2.0705, 2.2774, 2.2014, 1.7093], + device='cuda:0'), covar=tensor([0.0569, 0.0633, 0.0581, 0.0880, 0.0709, 0.0656, 0.0573, 0.1131], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0131, 0.0140, 0.0121, 0.0122, 0.0139, 0.0139, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:35:23,208 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:30,239 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6128, 1.3438, 1.8600, 3.3387, 2.2024, 2.3088, 0.8054, 2.7266], + device='cuda:0'), covar=tensor([0.1786, 0.1745, 0.1553, 0.0725, 0.0905, 0.1590, 0.2196, 0.0617], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0124, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:35:34,335 INFO [finetune.py:976] (0/7) Epoch 14, batch 4950, loss[loss=0.2083, simple_loss=0.2709, pruned_loss=0.07281, over 4844.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2591, pruned_loss=0.0612, over 951260.11 frames. ], batch size: 44, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:35:45,283 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:51,932 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79436.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:35:55,994 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.541e+02 1.877e+02 2.434e+02 3.585e+02, threshold=3.755e+02, percent-clipped=0.0 +2023-03-26 17:36:07,907 INFO [finetune.py:976] (0/7) Epoch 14, batch 5000, loss[loss=0.154, simple_loss=0.2276, pruned_loss=0.04019, over 4798.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2573, pruned_loss=0.06048, over 952554.50 frames. ], batch size: 25, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:36:35,550 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1603, 2.9161, 2.7533, 1.2382, 2.9629, 2.1314, 0.8915, 1.8223], + device='cuda:0'), covar=tensor([0.2524, 0.2229, 0.1866, 0.3979, 0.1402, 0.1361, 0.4250, 0.1988], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0173, 0.0160, 0.0128, 0.0157, 0.0122, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 17:36:41,414 INFO [finetune.py:976] (0/7) Epoch 14, batch 5050, loss[loss=0.2019, simple_loss=0.2694, pruned_loss=0.06725, over 4810.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2543, pruned_loss=0.05994, over 954236.19 frames. ], batch size: 39, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:36:44,570 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-26 17:36:47,568 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 17:36:56,165 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1815, 2.0755, 1.7362, 2.1226, 2.1924, 1.8749, 2.5224, 2.1813], + device='cuda:0'), covar=tensor([0.1565, 0.2099, 0.3259, 0.2501, 0.2737, 0.1896, 0.2582, 0.1985], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0254, 0.0243, 0.0199, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:37:02,363 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.19 vs. limit=5.0 +2023-03-26 17:37:02,697 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.527e+02 1.776e+02 2.127e+02 3.568e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-03-26 17:37:04,648 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9799, 3.2162, 2.9766, 2.2490, 2.8736, 3.4551, 3.1858, 2.8021], + device='cuda:0'), covar=tensor([0.0532, 0.0478, 0.0613, 0.0801, 0.0871, 0.0580, 0.0530, 0.0800], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0130, 0.0139, 0.0121, 0.0122, 0.0139, 0.0139, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:37:10,043 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3722, 1.9180, 2.3656, 2.2338, 1.9757, 1.9844, 2.1549, 2.1105], + device='cuda:0'), covar=tensor([0.3920, 0.4486, 0.3516, 0.4264, 0.5679, 0.4105, 0.5613, 0.3346], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0240, 0.0258, 0.0268, 0.0266, 0.0239, 0.0280, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:37:14,551 INFO [finetune.py:976] (0/7) Epoch 14, batch 5100, loss[loss=0.2204, simple_loss=0.2692, pruned_loss=0.08586, over 4834.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2514, pruned_loss=0.05917, over 953684.06 frames. ], batch size: 33, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:37:15,412 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 17:37:38,097 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:37:52,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4748, 1.5813, 1.6894, 0.9323, 1.6666, 1.9256, 1.8949, 1.4533], + device='cuda:0'), covar=tensor([0.1023, 0.0668, 0.0488, 0.0641, 0.0544, 0.0653, 0.0395, 0.0727], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0151, 0.0122, 0.0129, 0.0130, 0.0126, 0.0141, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.2795e-05, 1.0997e-04, 8.7774e-05, 9.2296e-05, 9.1851e-05, 9.1140e-05, + 1.0191e-04, 1.0476e-04], device='cuda:0') +2023-03-26 17:37:57,945 INFO [finetune.py:976] (0/7) Epoch 14, batch 5150, loss[loss=0.1658, simple_loss=0.2429, pruned_loss=0.0444, over 4894.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2507, pruned_loss=0.05873, over 953362.59 frames. ], batch size: 35, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:38:04,628 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79621.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:20,362 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79641.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:38:21,527 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.601e+02 1.924e+02 2.369e+02 3.228e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-26 17:38:23,480 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:34,358 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:41,522 INFO [finetune.py:976] (0/7) Epoch 14, batch 5200, loss[loss=0.1766, simple_loss=0.2535, pruned_loss=0.04988, over 4903.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2522, pruned_loss=0.05902, over 952259.03 frames. ], batch size: 37, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:38:50,974 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:02,922 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2449, 1.4416, 0.7781, 2.0551, 2.4867, 1.7790, 1.9745, 2.0234], + device='cuda:0'), covar=tensor([0.1391, 0.2023, 0.2125, 0.1163, 0.1806, 0.1906, 0.1301, 0.1855], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:39:10,035 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79689.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:27,297 INFO [finetune.py:976] (0/7) Epoch 14, batch 5250, loss[loss=0.1987, simple_loss=0.2721, pruned_loss=0.06267, over 4811.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2552, pruned_loss=0.06007, over 951333.67 frames. ], batch size: 45, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:39:32,141 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:33,304 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:40,418 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79731.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:39:49,074 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.683e+02 1.987e+02 2.478e+02 3.642e+02, threshold=3.974e+02, percent-clipped=0.0 +2023-03-26 17:39:59,969 INFO [finetune.py:976] (0/7) Epoch 14, batch 5300, loss[loss=0.1784, simple_loss=0.2481, pruned_loss=0.05435, over 4854.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2575, pruned_loss=0.06042, over 951252.69 frames. ], batch size: 31, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:40:00,694 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79762.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:40:02,495 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2346, 1.3968, 0.7998, 2.0352, 2.4552, 1.9338, 1.9750, 2.0958], + device='cuda:0'), covar=tensor([0.1349, 0.1924, 0.2017, 0.1105, 0.1728, 0.1727, 0.1270, 0.1705], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:40:14,302 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6882, 1.5457, 1.9255, 1.2555, 1.7458, 1.8509, 1.4942, 2.0864], + device='cuda:0'), covar=tensor([0.1184, 0.2006, 0.1392, 0.1760, 0.0833, 0.1317, 0.2609, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0204, 0.0192, 0.0190, 0.0175, 0.0213, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:40:33,373 INFO [finetune.py:976] (0/7) Epoch 14, batch 5350, loss[loss=0.1365, simple_loss=0.2076, pruned_loss=0.03267, over 4755.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.258, pruned_loss=0.06071, over 951284.68 frames. ], batch size: 26, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:40:35,322 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3440, 2.0536, 1.6713, 0.7701, 1.8713, 1.8467, 1.7382, 1.9121], + device='cuda:0'), covar=tensor([0.0916, 0.0840, 0.1718, 0.1894, 0.1391, 0.2316, 0.2109, 0.0960], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0195, 0.0200, 0.0183, 0.0214, 0.0208, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:40:41,369 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:40:55,394 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.503e+02 1.776e+02 2.291e+02 4.117e+02, threshold=3.553e+02, percent-clipped=3.0 +2023-03-26 17:41:06,813 INFO [finetune.py:976] (0/7) Epoch 14, batch 5400, loss[loss=0.2151, simple_loss=0.2679, pruned_loss=0.08114, over 4842.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2559, pruned_loss=0.05983, over 952067.17 frames. ], batch size: 49, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:41:09,374 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9433, 1.6933, 1.5080, 1.2820, 1.6577, 1.6658, 1.6414, 2.2386], + device='cuda:0'), covar=tensor([0.3901, 0.4616, 0.3482, 0.4210, 0.4040, 0.2536, 0.3988, 0.1985], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0226, 0.0278, 0.0248, 0.0215, 0.0250, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:41:40,249 INFO [finetune.py:976] (0/7) Epoch 14, batch 5450, loss[loss=0.1809, simple_loss=0.2436, pruned_loss=0.05909, over 4912.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2526, pruned_loss=0.05846, over 955003.89 frames. ], batch size: 37, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:41:43,530 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-03-26 17:41:44,677 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7020, 0.7240, 1.7863, 1.6088, 1.5624, 1.4893, 1.5556, 1.7246], + device='cuda:0'), covar=tensor([0.3883, 0.4029, 0.3394, 0.3649, 0.4637, 0.3465, 0.4457, 0.3239], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0240, 0.0258, 0.0268, 0.0267, 0.0239, 0.0280, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:41:59,674 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:41:59,717 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79941.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:42:00,807 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.485e+02 1.773e+02 2.175e+02 4.141e+02, threshold=3.546e+02, percent-clipped=3.0 +2023-03-26 17:42:14,248 INFO [finetune.py:976] (0/7) Epoch 14, batch 5500, loss[loss=0.1907, simple_loss=0.2465, pruned_loss=0.0675, over 4821.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2491, pruned_loss=0.05748, over 955132.74 frames. ], batch size: 30, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:42:21,240 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-26 17:42:31,230 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3380, 1.4037, 1.4888, 0.7195, 1.4167, 1.7177, 1.7067, 1.3545], + device='cuda:0'), covar=tensor([0.1050, 0.0664, 0.0453, 0.0644, 0.0469, 0.0601, 0.0331, 0.0723], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0122, 0.0128, 0.0129, 0.0126, 0.0140, 0.0144], + device='cuda:0'), out_proj_covar=tensor([9.1991e-05, 1.0966e-04, 8.7425e-05, 9.2032e-05, 9.1494e-05, 9.1333e-05, + 1.0124e-04, 1.0386e-04], device='cuda:0') +2023-03-26 17:42:32,370 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79989.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:42:32,388 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:42:39,626 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-80000.pt +2023-03-26 17:42:42,509 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7556, 1.1560, 0.8178, 1.6184, 2.0915, 1.2386, 1.5361, 1.5535], + device='cuda:0'), covar=tensor([0.1605, 0.2251, 0.2179, 0.1327, 0.1973, 0.2177, 0.1566, 0.2153], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:42:54,668 INFO [finetune.py:976] (0/7) Epoch 14, batch 5550, loss[loss=0.1653, simple_loss=0.2453, pruned_loss=0.04266, over 4901.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2502, pruned_loss=0.05801, over 954521.28 frames. ], batch size: 43, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:42:55,965 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80013.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:00,248 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:02,117 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2202, 1.9075, 2.0292, 0.9561, 2.3011, 2.5000, 2.1620, 1.9426], + device='cuda:0'), covar=tensor([0.0945, 0.0914, 0.0528, 0.0736, 0.0498, 0.0695, 0.0446, 0.0718], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0152, 0.0123, 0.0129, 0.0130, 0.0128, 0.0141, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.2964e-05, 1.1072e-04, 8.8256e-05, 9.2766e-05, 9.2329e-05, 9.2356e-05, + 1.0205e-04, 1.0486e-04], device='cuda:0') +2023-03-26 17:43:07,432 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80031.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:43:11,545 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:15,065 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.609e+02 1.862e+02 2.441e+02 4.163e+02, threshold=3.724e+02, percent-clipped=2.0 +2023-03-26 17:43:25,569 INFO [finetune.py:976] (0/7) Epoch 14, batch 5600, loss[loss=0.1757, simple_loss=0.2531, pruned_loss=0.04909, over 4813.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2552, pruned_loss=0.05918, over 955780.81 frames. ], batch size: 40, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:43:29,672 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80068.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:36,034 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80079.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:44:12,673 INFO [finetune.py:976] (0/7) Epoch 14, batch 5650, loss[loss=0.2112, simple_loss=0.2685, pruned_loss=0.07696, over 4822.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2577, pruned_loss=0.06, over 954831.44 frames. ], batch size: 25, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:44:21,790 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:44:40,040 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6343, 0.7185, 1.6811, 1.5799, 1.5039, 1.4644, 1.4986, 1.6166], + device='cuda:0'), covar=tensor([0.3186, 0.3674, 0.3259, 0.3306, 0.4150, 0.3206, 0.3854, 0.2976], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0240, 0.0258, 0.0267, 0.0267, 0.0239, 0.0280, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:44:41,676 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.565e+02 1.838e+02 2.201e+02 4.652e+02, threshold=3.676e+02, percent-clipped=2.0 +2023-03-26 17:44:56,306 INFO [finetune.py:976] (0/7) Epoch 14, batch 5700, loss[loss=0.1965, simple_loss=0.2477, pruned_loss=0.07263, over 4279.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2545, pruned_loss=0.05975, over 938044.94 frames. ], batch size: 18, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:45:12,613 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-14.pt +2023-03-26 17:45:28,231 INFO [finetune.py:976] (0/7) Epoch 15, batch 0, loss[loss=0.2241, simple_loss=0.2915, pruned_loss=0.07836, over 4788.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.2915, pruned_loss=0.07836, over 4788.00 frames. ], batch size: 51, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:45:28,232 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 17:45:30,416 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2972, 2.0563, 1.4728, 0.5899, 1.8483, 1.8942, 1.7657, 1.9340], + device='cuda:0'), covar=tensor([0.0851, 0.0828, 0.1564, 0.2115, 0.1314, 0.2792, 0.2281, 0.0861], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0195, 0.0200, 0.0183, 0.0213, 0.0207, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:45:42,543 INFO [finetune.py:1010] (0/7) Epoch 15, validation: loss=0.1586, simple_loss=0.2288, pruned_loss=0.0442, over 2265189.00 frames. +2023-03-26 17:45:42,543 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 17:45:42,691 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3613, 2.3126, 1.7479, 2.5987, 2.3752, 2.0151, 2.8527, 2.4204], + device='cuda:0'), covar=tensor([0.1402, 0.2460, 0.3268, 0.2750, 0.2635, 0.1817, 0.3386, 0.1882], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0255, 0.0245, 0.0200, 0.0213, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:46:12,249 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2308, 1.3801, 0.7558, 2.0614, 2.4023, 1.7804, 1.7782, 2.0376], + device='cuda:0'), covar=tensor([0.1349, 0.2032, 0.2151, 0.1096, 0.1826, 0.1919, 0.1350, 0.1837], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 17:46:15,203 INFO [finetune.py:976] (0/7) Epoch 15, batch 50, loss[loss=0.2171, simple_loss=0.2806, pruned_loss=0.07678, over 4807.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2638, pruned_loss=0.0648, over 217768.69 frames. ], batch size: 40, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:46:17,641 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:46:17,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8857, 1.8266, 1.5611, 2.0016, 2.5164, 2.0453, 1.6665, 1.4769], + device='cuda:0'), covar=tensor([0.2202, 0.2115, 0.1978, 0.1633, 0.1614, 0.1177, 0.2418, 0.2007], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0207, 0.0211, 0.0191, 0.0241, 0.0185, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:46:18,753 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.470e+02 1.896e+02 2.201e+02 3.299e+02, threshold=3.792e+02, percent-clipped=0.0 +2023-03-26 17:46:21,712 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4845, 1.5856, 2.1855, 1.7856, 1.7621, 4.0346, 1.4431, 1.7243], + device='cuda:0'), covar=tensor([0.0976, 0.1757, 0.1114, 0.1049, 0.1486, 0.0222, 0.1545, 0.1667], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0080, 0.0073, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:46:37,383 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 17:46:45,379 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80283.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:46:48,316 INFO [finetune.py:976] (0/7) Epoch 15, batch 100, loss[loss=0.1691, simple_loss=0.2399, pruned_loss=0.0492, over 4717.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2554, pruned_loss=0.06189, over 380954.28 frames. ], batch size: 23, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:46:48,982 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:03,370 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 17:47:05,084 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80313.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:21,598 INFO [finetune.py:976] (0/7) Epoch 15, batch 150, loss[loss=0.1451, simple_loss=0.212, pruned_loss=0.03911, over 4818.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2502, pruned_loss=0.06007, over 507388.84 frames. ], batch size: 25, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:47:22,342 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1144, 1.8902, 1.4054, 0.5399, 1.7409, 1.7783, 1.5886, 1.7674], + device='cuda:0'), covar=tensor([0.0759, 0.0742, 0.1348, 0.1787, 0.1109, 0.2094, 0.2095, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0193, 0.0199, 0.0182, 0.0212, 0.0205, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:47:25,138 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.875e+01 1.589e+02 1.860e+02 2.189e+02 4.694e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 17:47:25,893 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80344.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:36,664 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80361.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:42,667 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4355, 1.5212, 1.2700, 1.5504, 1.8510, 1.6245, 1.5334, 1.3430], + device='cuda:0'), covar=tensor([0.0380, 0.0303, 0.0578, 0.0274, 0.0187, 0.0563, 0.0328, 0.0394], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0141, 0.0113, 0.0100, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2362e-05, 8.3734e-05, 1.1152e-04, 8.7637e-05, 7.8303e-05, 7.9081e-05, + 7.2584e-05, 8.2632e-05], device='cuda:0') +2023-03-26 17:47:54,529 INFO [finetune.py:976] (0/7) Epoch 15, batch 200, loss[loss=0.2072, simple_loss=0.273, pruned_loss=0.07072, over 4805.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.248, pruned_loss=0.05885, over 607189.33 frames. ], batch size: 51, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:47:57,733 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:16,692 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:21,072 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 17:48:34,169 INFO [finetune.py:976] (0/7) Epoch 15, batch 250, loss[loss=0.2727, simple_loss=0.3294, pruned_loss=0.108, over 4806.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2524, pruned_loss=0.06016, over 685833.49 frames. ], batch size: 45, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:48:37,160 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.638e+02 2.049e+02 2.410e+02 5.367e+02, threshold=4.098e+02, percent-clipped=2.0 +2023-03-26 17:48:48,492 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:58,619 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:49:00,355 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:49:20,044 INFO [finetune.py:976] (0/7) Epoch 15, batch 300, loss[loss=0.1574, simple_loss=0.2392, pruned_loss=0.03782, over 4878.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.256, pruned_loss=0.06083, over 746275.46 frames. ], batch size: 32, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:49:24,816 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80494.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:03,909 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:12,917 INFO [finetune.py:976] (0/7) Epoch 15, batch 350, loss[loss=0.2005, simple_loss=0.2657, pruned_loss=0.06769, over 4806.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2599, pruned_loss=0.06212, over 794955.61 frames. ], batch size: 39, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:50:14,888 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3185, 1.4978, 1.6245, 0.8086, 1.5370, 1.7793, 1.8628, 1.4322], + device='cuda:0'), covar=tensor([0.0973, 0.0620, 0.0484, 0.0593, 0.0470, 0.0633, 0.0327, 0.0750], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0152, 0.0123, 0.0130, 0.0130, 0.0128, 0.0142, 0.0145], + device='cuda:0'), out_proj_covar=tensor([9.3312e-05, 1.1110e-04, 8.8617e-05, 9.3165e-05, 9.1907e-05, 9.2593e-05, + 1.0288e-04, 1.0513e-04], device='cuda:0') +2023-03-26 17:50:16,427 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.511e+02 1.809e+02 2.185e+02 3.892e+02, threshold=3.618e+02, percent-clipped=0.0 +2023-03-26 17:50:20,706 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0062, 1.9359, 1.7680, 2.1066, 2.5181, 2.1146, 1.6749, 1.6520], + device='cuda:0'), covar=tensor([0.2177, 0.2053, 0.2027, 0.1769, 0.1584, 0.1207, 0.2452, 0.2004], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0207, 0.0210, 0.0191, 0.0240, 0.0185, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:50:24,335 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:47,435 INFO [finetune.py:976] (0/7) Epoch 15, batch 400, loss[loss=0.1719, simple_loss=0.2404, pruned_loss=0.05175, over 4770.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2588, pruned_loss=0.06105, over 830408.01 frames. ], batch size: 26, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:51:29,119 INFO [finetune.py:976] (0/7) Epoch 15, batch 450, loss[loss=0.1804, simple_loss=0.2534, pruned_loss=0.05371, over 4898.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2571, pruned_loss=0.06054, over 858146.94 frames. ], batch size: 36, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:51:29,777 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:51:32,682 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.580e+02 1.854e+02 2.177e+02 4.594e+02, threshold=3.707e+02, percent-clipped=2.0 +2023-03-26 17:52:03,114 INFO [finetune.py:976] (0/7) Epoch 15, batch 500, loss[loss=0.179, simple_loss=0.2401, pruned_loss=0.05894, over 4862.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2547, pruned_loss=0.05963, over 879977.23 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:52:23,154 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9493, 3.8895, 3.7212, 1.9653, 3.9675, 3.0994, 0.9525, 2.9606], + device='cuda:0'), covar=tensor([0.2279, 0.1679, 0.1372, 0.3014, 0.0979, 0.0943, 0.4321, 0.1298], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0174, 0.0160, 0.0128, 0.0157, 0.0123, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 17:52:36,001 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80736.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:52:37,122 INFO [finetune.py:976] (0/7) Epoch 15, batch 550, loss[loss=0.1355, simple_loss=0.2045, pruned_loss=0.03328, over 4723.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2508, pruned_loss=0.05809, over 896616.35 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:52:40,201 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.496e+02 1.725e+02 2.011e+02 3.976e+02, threshold=3.451e+02, percent-clipped=1.0 +2023-03-26 17:52:44,375 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80749.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:01,299 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:10,746 INFO [finetune.py:976] (0/7) Epoch 15, batch 600, loss[loss=0.177, simple_loss=0.2508, pruned_loss=0.05162, over 4932.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2526, pruned_loss=0.05863, over 910762.26 frames. ], batch size: 42, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:53:16,333 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80797.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:32,683 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:36,027 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 17:53:44,648 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:45,362 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 17:53:47,023 INFO [finetune.py:976] (0/7) Epoch 15, batch 650, loss[loss=0.1976, simple_loss=0.2584, pruned_loss=0.06846, over 4147.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2557, pruned_loss=0.06012, over 920380.23 frames. ], batch size: 65, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:53:50,575 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.643e+02 1.965e+02 2.358e+02 6.399e+02, threshold=3.929e+02, percent-clipped=5.0 +2023-03-26 17:53:54,881 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:54:08,887 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5717, 1.4486, 1.4508, 1.3951, 0.8393, 2.3313, 0.8359, 1.2803], + device='cuda:0'), covar=tensor([0.3425, 0.2581, 0.2233, 0.2659, 0.2079, 0.0343, 0.2806, 0.1441], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:54:10,833 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 17:54:16,853 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8439, 1.5562, 2.3962, 1.4203, 2.1663, 2.1718, 1.4948, 2.2918], + device='cuda:0'), covar=tensor([0.1491, 0.2513, 0.1452, 0.2167, 0.1120, 0.1770, 0.3140, 0.1213], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0203, 0.0192, 0.0190, 0.0175, 0.0214, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:54:29,242 INFO [finetune.py:976] (0/7) Epoch 15, batch 700, loss[loss=0.2212, simple_loss=0.2795, pruned_loss=0.08141, over 4772.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2575, pruned_loss=0.06072, over 927717.68 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:55:23,120 INFO [finetune.py:976] (0/7) Epoch 15, batch 750, loss[loss=0.1806, simple_loss=0.252, pruned_loss=0.05463, over 4878.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2583, pruned_loss=0.06082, over 934640.40 frames. ], batch size: 43, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:55:23,801 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:55:26,163 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.628e+02 1.856e+02 2.303e+02 3.612e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-26 17:55:38,150 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0303, 2.0506, 2.1013, 1.6383, 2.1142, 2.1780, 2.3057, 1.8560], + device='cuda:0'), covar=tensor([0.0592, 0.0611, 0.0628, 0.0831, 0.0618, 0.0701, 0.0520, 0.1004], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0132, 0.0140, 0.0122, 0.0123, 0.0140, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 17:55:56,364 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:55:56,905 INFO [finetune.py:976] (0/7) Epoch 15, batch 800, loss[loss=0.2153, simple_loss=0.2797, pruned_loss=0.07549, over 4170.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2575, pruned_loss=0.05993, over 938754.98 frames. ], batch size: 65, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:56:35,927 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6296, 1.5238, 1.8709, 1.8327, 1.6918, 3.4536, 1.3297, 1.5830], + device='cuda:0'), covar=tensor([0.0912, 0.1790, 0.1113, 0.0966, 0.1602, 0.0270, 0.1611, 0.1748], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 17:56:36,697 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.42 vs. limit=5.0 +2023-03-26 17:56:38,263 INFO [finetune.py:976] (0/7) Epoch 15, batch 850, loss[loss=0.192, simple_loss=0.2654, pruned_loss=0.05927, over 4815.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2552, pruned_loss=0.05894, over 943557.71 frames. ], batch size: 41, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:56:41,040 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 17:56:41,290 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.679e+02 1.976e+02 2.340e+02 3.768e+02, threshold=3.952e+02, percent-clipped=2.0 +2023-03-26 17:56:44,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:06,262 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 17:57:11,961 INFO [finetune.py:976] (0/7) Epoch 15, batch 900, loss[loss=0.1425, simple_loss=0.2271, pruned_loss=0.02889, over 4901.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.253, pruned_loss=0.0584, over 946240.02 frames. ], batch size: 35, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:57:14,454 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:17,454 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81097.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:29,372 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:31,657 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:38,661 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:45,615 INFO [finetune.py:976] (0/7) Epoch 15, batch 950, loss[loss=0.1739, simple_loss=0.2374, pruned_loss=0.05513, over 4745.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2519, pruned_loss=0.05878, over 946613.66 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:57:48,668 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.129e+01 1.456e+02 1.848e+02 2.216e+02 5.430e+02, threshold=3.695e+02, percent-clipped=2.0 +2023-03-26 17:57:52,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:03,765 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:11,298 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81177.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:19,387 INFO [finetune.py:976] (0/7) Epoch 15, batch 1000, loss[loss=0.1905, simple_loss=0.2676, pruned_loss=0.05668, over 4923.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2545, pruned_loss=0.05962, over 950495.41 frames. ], batch size: 36, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:58:25,511 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81198.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:48,523 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-03-26 17:58:52,893 INFO [finetune.py:976] (0/7) Epoch 15, batch 1050, loss[loss=0.1868, simple_loss=0.2563, pruned_loss=0.05865, over 4741.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2565, pruned_loss=0.06004, over 952727.52 frames. ], batch size: 27, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:58:56,385 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.183e+02 1.566e+02 1.800e+02 2.282e+02 3.514e+02, threshold=3.601e+02, percent-clipped=0.0 +2023-03-26 17:59:31,695 INFO [finetune.py:976] (0/7) Epoch 15, batch 1100, loss[loss=0.1728, simple_loss=0.2459, pruned_loss=0.04984, over 4914.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2577, pruned_loss=0.05973, over 954940.47 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:59:43,370 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:59:49,373 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81301.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:59:53,679 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2840, 2.2089, 1.7316, 2.2896, 2.1782, 1.8776, 2.5869, 2.2554], + device='cuda:0'), covar=tensor([0.1379, 0.2291, 0.3117, 0.2696, 0.2698, 0.1755, 0.3013, 0.1863], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0187, 0.0234, 0.0255, 0.0245, 0.0200, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:00:01,195 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 18:00:16,529 INFO [finetune.py:976] (0/7) Epoch 15, batch 1150, loss[loss=0.1745, simple_loss=0.2386, pruned_loss=0.05519, over 4777.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2569, pruned_loss=0.05933, over 954273.54 frames. ], batch size: 26, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:00:23,240 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.635e+02 2.084e+02 2.407e+02 3.907e+02, threshold=4.168e+02, percent-clipped=1.0 +2023-03-26 18:00:31,933 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1557, 2.0669, 1.6356, 2.1439, 2.0908, 1.7861, 2.3988, 2.1074], + device='cuda:0'), covar=tensor([0.1321, 0.2317, 0.3291, 0.2634, 0.2706, 0.1700, 0.3670, 0.1841], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0187, 0.0234, 0.0254, 0.0245, 0.0199, 0.0213, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:00:41,619 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:00:42,876 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:00:46,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 18:00:59,899 INFO [finetune.py:976] (0/7) Epoch 15, batch 1200, loss[loss=0.2031, simple_loss=0.264, pruned_loss=0.07112, over 4925.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2553, pruned_loss=0.05875, over 954202.16 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:01:02,911 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81392.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:27,560 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:35,102 INFO [finetune.py:976] (0/7) Epoch 15, batch 1250, loss[loss=0.1796, simple_loss=0.2502, pruned_loss=0.05454, over 4742.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2544, pruned_loss=0.0593, over 953249.11 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:01:35,269 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 18:01:40,100 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:42,327 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.103e+01 1.546e+02 1.830e+02 2.259e+02 3.665e+02, threshold=3.660e+02, percent-clipped=0.0 +2023-03-26 18:02:05,448 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:02:08,448 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:02:15,564 INFO [finetune.py:976] (0/7) Epoch 15, batch 1300, loss[loss=0.1912, simple_loss=0.246, pruned_loss=0.06819, over 4750.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2508, pruned_loss=0.05804, over 953037.89 frames. ], batch size: 59, lr: 3.50e-03, grad_scale: 64.0 +2023-03-26 18:02:25,419 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9105, 1.3612, 0.9024, 1.8439, 2.1713, 1.5282, 1.5545, 1.6269], + device='cuda:0'), covar=tensor([0.1593, 0.2199, 0.2138, 0.1260, 0.1963, 0.2088, 0.1608, 0.2210], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0096, 0.0112, 0.0094, 0.0121, 0.0095, 0.0100, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 18:02:42,505 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 18:02:47,729 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6840, 1.6303, 1.5323, 1.6251, 1.1664, 4.0112, 1.6483, 1.8658], + device='cuda:0'), covar=tensor([0.3310, 0.2305, 0.2115, 0.2406, 0.1843, 0.0150, 0.2600, 0.1342], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:02:49,394 INFO [finetune.py:976] (0/7) Epoch 15, batch 1350, loss[loss=0.216, simple_loss=0.2862, pruned_loss=0.07295, over 4925.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2506, pruned_loss=0.05819, over 954188.03 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:02:53,035 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1762, 2.0222, 1.7392, 1.9065, 2.1776, 1.8461, 2.2443, 2.1266], + device='cuda:0'), covar=tensor([0.1424, 0.2222, 0.3288, 0.2627, 0.2624, 0.1813, 0.2998, 0.1978], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0188, 0.0235, 0.0256, 0.0246, 0.0200, 0.0214, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:02:53,475 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.608e+02 1.859e+02 2.257e+02 3.880e+02, threshold=3.719e+02, percent-clipped=1.0 +2023-03-26 18:03:06,621 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8934, 1.8498, 1.8809, 1.4271, 1.9998, 2.1069, 2.0751, 1.6503], + device='cuda:0'), covar=tensor([0.0561, 0.0609, 0.0680, 0.0807, 0.0610, 0.0638, 0.0506, 0.1050], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0132, 0.0140, 0.0121, 0.0122, 0.0139, 0.0139, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:03:10,257 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0831, 1.8130, 2.3501, 1.4930, 2.0496, 2.3662, 1.7585, 2.4335], + device='cuda:0'), covar=tensor([0.1263, 0.2022, 0.1310, 0.1862, 0.0932, 0.1291, 0.2622, 0.0777], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0206, 0.0193, 0.0192, 0.0179, 0.0215, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:03:22,722 INFO [finetune.py:976] (0/7) Epoch 15, batch 1400, loss[loss=0.1638, simple_loss=0.2449, pruned_loss=0.04139, over 4822.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2541, pruned_loss=0.05923, over 954789.78 frames. ], batch size: 33, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:03:33,296 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6031, 2.7985, 2.5811, 2.0094, 2.8521, 3.0009, 2.9061, 2.5529], + device='cuda:0'), covar=tensor([0.0702, 0.0547, 0.0748, 0.0915, 0.0493, 0.0678, 0.0590, 0.0911], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0132, 0.0140, 0.0122, 0.0122, 0.0140, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:03:56,014 INFO [finetune.py:976] (0/7) Epoch 15, batch 1450, loss[loss=0.1914, simple_loss=0.2572, pruned_loss=0.06277, over 4821.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2554, pruned_loss=0.05916, over 955689.16 frames. ], batch size: 39, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:04:00,098 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.620e+02 1.887e+02 2.237e+02 3.719e+02, threshold=3.774e+02, percent-clipped=1.0 +2023-03-26 18:04:07,940 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:09,554 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:18,096 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81670.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:29,491 INFO [finetune.py:976] (0/7) Epoch 15, batch 1500, loss[loss=0.1781, simple_loss=0.2542, pruned_loss=0.05102, over 4898.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2577, pruned_loss=0.06008, over 955554.25 frames. ], batch size: 36, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:05:16,701 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81731.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:05:20,259 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6175, 3.7635, 3.5511, 1.9529, 3.8344, 2.9729, 0.8383, 2.6851], + device='cuda:0'), covar=tensor([0.2538, 0.1696, 0.1487, 0.3118, 0.1073, 0.0994, 0.4598, 0.1460], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0156, 0.0122, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 18:05:20,803 INFO [finetune.py:976] (0/7) Epoch 15, batch 1550, loss[loss=0.1963, simple_loss=0.2632, pruned_loss=0.06475, over 4907.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2577, pruned_loss=0.06022, over 955740.84 frames. ], batch size: 37, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:05:24,955 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.530e+02 1.898e+02 2.293e+02 4.636e+02, threshold=3.795e+02, percent-clipped=1.0 +2023-03-26 18:05:50,065 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81772.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:06:03,756 INFO [finetune.py:976] (0/7) Epoch 15, batch 1600, loss[loss=0.2029, simple_loss=0.2661, pruned_loss=0.06981, over 4759.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2553, pruned_loss=0.05959, over 954583.56 frames. ], batch size: 27, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:06:12,831 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6787, 3.7723, 3.5967, 1.7606, 3.8545, 2.9256, 0.7934, 2.6845], + device='cuda:0'), covar=tensor([0.2377, 0.1883, 0.1453, 0.3354, 0.1086, 0.1002, 0.4491, 0.1457], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0172, 0.0157, 0.0127, 0.0155, 0.0121, 0.0144, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 18:06:25,743 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81820.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:06:34,865 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-26 18:06:37,140 INFO [finetune.py:976] (0/7) Epoch 15, batch 1650, loss[loss=0.1508, simple_loss=0.2218, pruned_loss=0.03995, over 4819.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2532, pruned_loss=0.05863, over 955617.21 frames. ], batch size: 39, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:06:40,761 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.362e+01 1.564e+02 1.826e+02 2.251e+02 4.924e+02, threshold=3.651e+02, percent-clipped=3.0 +2023-03-26 18:06:53,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4229, 3.8420, 4.0668, 4.2165, 4.1649, 3.9445, 4.4786, 1.5327], + device='cuda:0'), covar=tensor([0.0700, 0.0928, 0.0838, 0.1027, 0.1267, 0.1368, 0.0712, 0.5043], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0240, 0.0270, 0.0289, 0.0326, 0.0279, 0.0295, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:07:18,076 INFO [finetune.py:976] (0/7) Epoch 15, batch 1700, loss[loss=0.2343, simple_loss=0.2871, pruned_loss=0.09073, over 4242.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2506, pruned_loss=0.05782, over 955328.10 frames. ], batch size: 65, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:07:40,802 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 18:07:51,480 INFO [finetune.py:976] (0/7) Epoch 15, batch 1750, loss[loss=0.1989, simple_loss=0.2661, pruned_loss=0.06582, over 4940.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.253, pruned_loss=0.05835, over 955631.64 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:07:55,583 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.510e+02 1.914e+02 2.293e+02 4.004e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-03-26 18:08:02,970 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:04,185 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:24,300 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7856, 1.5634, 2.0154, 1.3714, 1.8793, 2.0167, 1.5300, 2.1195], + device='cuda:0'), covar=tensor([0.1353, 0.2136, 0.1478, 0.1805, 0.0851, 0.1350, 0.2767, 0.0861], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0203, 0.0192, 0.0189, 0.0176, 0.0212, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:08:24,904 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:25,414 INFO [finetune.py:976] (0/7) Epoch 15, batch 1800, loss[loss=0.1732, simple_loss=0.2533, pruned_loss=0.04657, over 4789.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2559, pruned_loss=0.05935, over 952357.08 frames. ], batch size: 45, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:08:33,315 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-82000.pt +2023-03-26 18:08:36,195 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82003.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:37,849 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82005.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:52,497 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82026.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:09:00,025 INFO [finetune.py:976] (0/7) Epoch 15, batch 1850, loss[loss=0.2329, simple_loss=0.2871, pruned_loss=0.08935, over 4831.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2572, pruned_loss=0.05984, over 953240.31 frames. ], batch size: 49, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:09:03,672 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.664e+02 1.894e+02 2.440e+02 3.763e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-03-26 18:09:06,721 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:09:11,007 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:09:33,282 INFO [finetune.py:976] (0/7) Epoch 15, batch 1900, loss[loss=0.1736, simple_loss=0.2428, pruned_loss=0.05219, over 4875.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2566, pruned_loss=0.05873, over 954835.27 frames. ], batch size: 32, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:09:51,830 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:10:16,117 INFO [finetune.py:976] (0/7) Epoch 15, batch 1950, loss[loss=0.2088, simple_loss=0.2724, pruned_loss=0.07261, over 4885.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2543, pruned_loss=0.05776, over 952165.37 frames. ], batch size: 35, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:10:24,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.525e+02 1.906e+02 2.226e+02 4.434e+02, threshold=3.812e+02, percent-clipped=2.0 +2023-03-26 18:11:01,528 INFO [finetune.py:976] (0/7) Epoch 15, batch 2000, loss[loss=0.1775, simple_loss=0.2482, pruned_loss=0.05344, over 4936.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2519, pruned_loss=0.0578, over 948435.33 frames. ], batch size: 33, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:11:10,712 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.4227, 3.0193, 2.6755, 1.4485, 2.8732, 2.4910, 2.3293, 2.5158], + device='cuda:0'), covar=tensor([0.0850, 0.0846, 0.2029, 0.2277, 0.1803, 0.1948, 0.2149, 0.1214], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0201, 0.0184, 0.0214, 0.0208, 0.0226, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:11:38,380 INFO [finetune.py:976] (0/7) Epoch 15, batch 2050, loss[loss=0.1849, simple_loss=0.2602, pruned_loss=0.05478, over 4829.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2501, pruned_loss=0.05768, over 951436.75 frames. ], batch size: 30, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:11:42,512 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.417e+02 1.727e+02 2.274e+02 4.171e+02, threshold=3.454e+02, percent-clipped=1.0 +2023-03-26 18:11:55,954 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1135, 1.2457, 1.5748, 1.0365, 1.1307, 1.4020, 1.2477, 1.5571], + device='cuda:0'), covar=tensor([0.1322, 0.2264, 0.1254, 0.1478, 0.1010, 0.1192, 0.2842, 0.0891], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0206, 0.0195, 0.0192, 0.0178, 0.0214, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:12:14,304 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4544, 1.4174, 1.7752, 1.2478, 1.5521, 1.7595, 1.3364, 1.9338], + device='cuda:0'), covar=tensor([0.1525, 0.2252, 0.1415, 0.1757, 0.0989, 0.1396, 0.3191, 0.0900], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0206, 0.0195, 0.0192, 0.0178, 0.0215, 0.0219, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:12:24,922 INFO [finetune.py:976] (0/7) Epoch 15, batch 2100, loss[loss=0.1745, simple_loss=0.2372, pruned_loss=0.05594, over 4778.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2498, pruned_loss=0.0579, over 952723.75 frames. ], batch size: 26, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:12:39,862 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1598, 1.9230, 1.4776, 0.5925, 1.6612, 1.8374, 1.6453, 1.7389], + device='cuda:0'), covar=tensor([0.1013, 0.0832, 0.1695, 0.2100, 0.1459, 0.2315, 0.2326, 0.0903], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0193, 0.0199, 0.0183, 0.0213, 0.0207, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:12:54,206 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:12:57,984 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9395, 1.8637, 1.5354, 1.7464, 1.6721, 1.6636, 1.7832, 2.4349], + device='cuda:0'), covar=tensor([0.3743, 0.4068, 0.3199, 0.3769, 0.4025, 0.2427, 0.3648, 0.1622], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0226, 0.0275, 0.0247, 0.0214, 0.0249, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:13:00,930 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9301, 1.9425, 1.6127, 1.8170, 1.9030, 1.6508, 2.1673, 1.9588], + device='cuda:0'), covar=tensor([0.1181, 0.1894, 0.2619, 0.2299, 0.2386, 0.1489, 0.2546, 0.1492], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0187, 0.0234, 0.0253, 0.0245, 0.0200, 0.0212, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:13:02,490 INFO [finetune.py:976] (0/7) Epoch 15, batch 2150, loss[loss=0.1849, simple_loss=0.2297, pruned_loss=0.07008, over 4182.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2538, pruned_loss=0.05904, over 952406.26 frames. ], batch size: 18, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:13:06,114 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:06,660 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.137e+02 1.626e+02 1.861e+02 2.291e+02 4.001e+02, threshold=3.721e+02, percent-clipped=2.0 +2023-03-26 18:13:07,989 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:25,988 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:35,325 INFO [finetune.py:976] (0/7) Epoch 15, batch 2200, loss[loss=0.2132, simple_loss=0.2864, pruned_loss=0.06997, over 4874.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2557, pruned_loss=0.0594, over 950612.98 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:13:48,064 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:50,415 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82411.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:14:03,492 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1260, 2.0981, 1.9879, 2.1333, 1.8299, 3.9470, 1.9845, 2.5499], + device='cuda:0'), covar=tensor([0.3035, 0.2155, 0.1737, 0.2043, 0.1475, 0.0197, 0.2122, 0.0887], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:14:08,110 INFO [finetune.py:976] (0/7) Epoch 15, batch 2250, loss[loss=0.1465, simple_loss=0.2235, pruned_loss=0.03476, over 4740.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2573, pruned_loss=0.06002, over 951216.36 frames. ], batch size: 27, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:14:12,182 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.505e+02 1.711e+02 2.071e+02 3.892e+02, threshold=3.421e+02, percent-clipped=2.0 +2023-03-26 18:14:41,720 INFO [finetune.py:976] (0/7) Epoch 15, batch 2300, loss[loss=0.1667, simple_loss=0.2394, pruned_loss=0.04699, over 4860.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2561, pruned_loss=0.05905, over 950926.95 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:14:45,292 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82493.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 18:14:51,037 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9328, 2.6430, 2.3491, 3.1607, 2.8485, 2.6124, 3.3637, 2.9011], + device='cuda:0'), covar=tensor([0.1286, 0.2315, 0.3291, 0.2610, 0.2544, 0.1615, 0.2576, 0.1870], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0188, 0.0235, 0.0254, 0.0246, 0.0201, 0.0214, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:15:17,448 INFO [finetune.py:976] (0/7) Epoch 15, batch 2350, loss[loss=0.1963, simple_loss=0.2396, pruned_loss=0.0765, over 4870.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.255, pruned_loss=0.05973, over 950615.56 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:15:21,101 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.644e+02 1.984e+02 2.390e+02 4.799e+02, threshold=3.967e+02, percent-clipped=3.0 +2023-03-26 18:15:28,731 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82554.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:15:33,657 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 18:15:36,151 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 18:15:44,892 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:16:00,854 INFO [finetune.py:976] (0/7) Epoch 15, batch 2400, loss[loss=0.1956, simple_loss=0.2566, pruned_loss=0.06735, over 4869.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2522, pruned_loss=0.05884, over 951612.51 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:16:34,394 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1779, 1.9718, 1.4968, 0.6141, 1.7072, 1.8388, 1.6427, 1.7478], + device='cuda:0'), covar=tensor([0.0951, 0.0695, 0.1380, 0.1864, 0.1205, 0.2241, 0.2122, 0.0868], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0195, 0.0201, 0.0185, 0.0214, 0.0208, 0.0226, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:16:37,327 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:16:42,551 INFO [finetune.py:976] (0/7) Epoch 15, batch 2450, loss[loss=0.2109, simple_loss=0.2605, pruned_loss=0.08064, over 4820.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2494, pruned_loss=0.05779, over 953924.62 frames. ], batch size: 33, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:16:45,671 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:16:46,164 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.277e+01 1.596e+02 1.937e+02 2.264e+02 4.235e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-26 18:17:12,243 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7236, 1.5781, 2.1750, 3.5010, 2.4226, 2.4684, 1.2408, 2.9258], + device='cuda:0'), covar=tensor([0.1667, 0.1424, 0.1293, 0.0490, 0.0721, 0.1486, 0.1649, 0.0452], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0163, 0.0099, 0.0137, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:17:18,146 INFO [finetune.py:976] (0/7) Epoch 15, batch 2500, loss[loss=0.1634, simple_loss=0.2327, pruned_loss=0.04711, over 4722.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2525, pruned_loss=0.05948, over 954030.24 frames. ], batch size: 23, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:17:20,060 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82691.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:17:35,742 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82702.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:17:46,362 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82711.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:17:57,253 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6959, 1.5184, 1.0488, 0.2777, 1.3163, 1.4648, 1.4348, 1.4786], + device='cuda:0'), covar=tensor([0.0978, 0.0957, 0.1548, 0.2209, 0.1590, 0.2667, 0.2510, 0.0950], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0194, 0.0200, 0.0184, 0.0213, 0.0207, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:18:03,677 INFO [finetune.py:976] (0/7) Epoch 15, batch 2550, loss[loss=0.1601, simple_loss=0.2303, pruned_loss=0.04492, over 4776.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2552, pruned_loss=0.05973, over 953273.95 frames. ], batch size: 26, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:18:06,086 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6387, 1.5408, 1.4209, 1.5794, 1.2487, 4.3118, 1.4462, 1.9241], + device='cuda:0'), covar=tensor([0.3450, 0.2619, 0.2330, 0.2452, 0.1790, 0.0123, 0.2744, 0.1371], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0114, 0.0097, 0.0097, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:18:07,766 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.574e+02 1.850e+02 2.269e+02 4.152e+02, threshold=3.700e+02, percent-clipped=3.0 +2023-03-26 18:18:16,095 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2509, 2.0898, 1.6716, 2.0383, 1.9740, 1.9272, 2.0270, 2.7702], + device='cuda:0'), covar=tensor([0.3933, 0.4618, 0.3540, 0.4268, 0.3841, 0.2635, 0.4129, 0.1690], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0225, 0.0274, 0.0246, 0.0213, 0.0247, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:18:17,863 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:18:36,861 INFO [finetune.py:976] (0/7) Epoch 15, batch 2600, loss[loss=0.219, simple_loss=0.2847, pruned_loss=0.07664, over 4910.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2568, pruned_loss=0.0603, over 953426.70 frames. ], batch size: 42, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:19:06,662 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1052, 1.7030, 2.0563, 2.0645, 1.7951, 1.8021, 1.9941, 1.9005], + device='cuda:0'), covar=tensor([0.3973, 0.4304, 0.3306, 0.3908, 0.4989, 0.3974, 0.4863, 0.3351], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0238, 0.0256, 0.0268, 0.0266, 0.0239, 0.0280, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:19:10,656 INFO [finetune.py:976] (0/7) Epoch 15, batch 2650, loss[loss=0.1776, simple_loss=0.2553, pruned_loss=0.04991, over 4831.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2588, pruned_loss=0.06119, over 952309.02 frames. ], batch size: 47, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:19:14,266 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.579e+02 1.879e+02 2.251e+02 6.929e+02, threshold=3.759e+02, percent-clipped=2.0 +2023-03-26 18:19:17,847 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82849.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:19:18,473 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5240, 1.4266, 1.9083, 2.9523, 1.9666, 2.1061, 1.0417, 2.5447], + device='cuda:0'), covar=tensor([0.1777, 0.1457, 0.1222, 0.0559, 0.0859, 0.1412, 0.1679, 0.0505], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0133, 0.0164, 0.0100, 0.0138, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:19:36,772 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 18:19:43,179 INFO [finetune.py:976] (0/7) Epoch 15, batch 2700, loss[loss=0.1911, simple_loss=0.2591, pruned_loss=0.06148, over 4870.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2579, pruned_loss=0.06061, over 952437.83 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:20:08,050 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:20:16,388 INFO [finetune.py:976] (0/7) Epoch 15, batch 2750, loss[loss=0.1599, simple_loss=0.2318, pruned_loss=0.04402, over 4755.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2537, pruned_loss=0.05875, over 952373.43 frames. ], batch size: 26, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:20:20,499 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.574e+02 1.758e+02 2.107e+02 4.076e+02, threshold=3.515e+02, percent-clipped=2.0 +2023-03-26 18:20:33,248 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82963.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:20:39,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 18:20:49,686 INFO [finetune.py:976] (0/7) Epoch 15, batch 2800, loss[loss=0.162, simple_loss=0.2189, pruned_loss=0.05251, over 4050.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2502, pruned_loss=0.05743, over 952427.03 frames. ], batch size: 17, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:21:07,056 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:21:15,929 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6477, 1.5340, 1.3342, 1.5532, 1.9123, 1.7699, 1.6086, 1.3382], + device='cuda:0'), covar=tensor([0.0265, 0.0293, 0.0612, 0.0286, 0.0194, 0.0413, 0.0294, 0.0357], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0143, 0.0113, 0.0100, 0.0107, 0.0098, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.3033e-05, 8.4486e-05, 1.1287e-04, 8.7577e-05, 7.8102e-05, 7.9378e-05, + 7.3458e-05, 8.3628e-05], device='cuda:0') +2023-03-26 18:21:21,940 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:21:37,765 INFO [finetune.py:976] (0/7) Epoch 15, batch 2850, loss[loss=0.1594, simple_loss=0.2349, pruned_loss=0.04197, over 4872.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2502, pruned_loss=0.05794, over 953836.84 frames. ], batch size: 34, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:21:41,397 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.604e+02 1.866e+02 2.227e+02 4.125e+02, threshold=3.733e+02, percent-clipped=3.0 +2023-03-26 18:21:48,544 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:22:05,104 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7389, 1.7174, 1.6836, 1.7027, 1.3719, 3.3410, 1.5942, 1.9883], + device='cuda:0'), covar=tensor([0.3006, 0.2167, 0.1857, 0.2107, 0.1531, 0.0214, 0.2659, 0.1107], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:22:15,027 INFO [finetune.py:976] (0/7) Epoch 15, batch 2900, loss[loss=0.1729, simple_loss=0.2501, pruned_loss=0.04785, over 4844.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2528, pruned_loss=0.05869, over 954340.33 frames. ], batch size: 49, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:22:57,705 INFO [finetune.py:976] (0/7) Epoch 15, batch 2950, loss[loss=0.1698, simple_loss=0.2379, pruned_loss=0.0508, over 4323.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2572, pruned_loss=0.06009, over 955099.00 frames. ], batch size: 19, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:23:01,328 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.748e+02 2.030e+02 2.368e+02 3.585e+02, threshold=4.059e+02, percent-clipped=0.0 +2023-03-26 18:23:08,897 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83149.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:23:11,900 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8124, 1.2802, 1.0109, 1.6392, 2.0405, 1.5197, 1.5114, 1.6904], + device='cuda:0'), covar=tensor([0.1326, 0.1972, 0.1879, 0.1141, 0.1895, 0.1928, 0.1315, 0.1743], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:23:29,405 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6159, 2.4290, 2.9858, 1.7750, 2.6180, 2.9501, 2.3408, 3.2175], + device='cuda:0'), covar=tensor([0.1473, 0.2018, 0.1830, 0.2442, 0.1059, 0.1611, 0.2452, 0.0947], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0204, 0.0193, 0.0190, 0.0175, 0.0213, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:23:35,419 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2773, 2.1102, 1.7577, 2.2506, 2.1969, 1.9026, 2.4818, 2.2292], + device='cuda:0'), covar=tensor([0.1390, 0.2277, 0.3222, 0.2507, 0.2739, 0.1756, 0.2920, 0.1937], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0187, 0.0235, 0.0252, 0.0244, 0.0200, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:23:37,597 INFO [finetune.py:976] (0/7) Epoch 15, batch 3000, loss[loss=0.205, simple_loss=0.2773, pruned_loss=0.06638, over 4899.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2569, pruned_loss=0.05976, over 954215.94 frames. ], batch size: 36, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:23:37,598 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 18:23:46,409 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2686, 2.1048, 1.5205, 0.6148, 1.7458, 1.9322, 1.7974, 1.9376], + device='cuda:0'), covar=tensor([0.0846, 0.0788, 0.1479, 0.2035, 0.1441, 0.2283, 0.2120, 0.0787], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0194, 0.0199, 0.0183, 0.0212, 0.0206, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:23:48,369 INFO [finetune.py:1010] (0/7) Epoch 15, validation: loss=0.1564, simple_loss=0.2269, pruned_loss=0.04296, over 2265189.00 frames. +2023-03-26 18:23:48,370 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 18:23:59,741 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:00,266 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83197.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:24:21,548 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:30,317 INFO [finetune.py:976] (0/7) Epoch 15, batch 3050, loss[loss=0.1909, simple_loss=0.2685, pruned_loss=0.05661, over 4907.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2564, pruned_loss=0.05928, over 953182.27 frames. ], batch size: 46, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:24:34,916 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.542e+02 1.763e+02 2.135e+02 3.801e+02, threshold=3.526e+02, percent-clipped=0.0 +2023-03-26 18:24:41,662 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1953, 2.1474, 1.6966, 2.2351, 2.1261, 1.8537, 2.4662, 2.2270], + device='cuda:0'), covar=tensor([0.1312, 0.2236, 0.3009, 0.2650, 0.2478, 0.1675, 0.3190, 0.1794], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0187, 0.0235, 0.0254, 0.0245, 0.0201, 0.0212, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:24:41,759 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 18:24:44,039 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83257.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:54,194 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:01,282 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:04,096 INFO [finetune.py:976] (0/7) Epoch 15, batch 3100, loss[loss=0.1581, simple_loss=0.2202, pruned_loss=0.04801, over 4733.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2549, pruned_loss=0.05893, over 950387.53 frames. ], batch size: 23, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:25:12,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7646, 1.2138, 0.8947, 1.6763, 2.2178, 1.2745, 1.5283, 1.6166], + device='cuda:0'), covar=tensor([0.1402, 0.2076, 0.1930, 0.1191, 0.1759, 0.1886, 0.1381, 0.1836], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0093, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:25:24,779 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83319.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:37,273 INFO [finetune.py:976] (0/7) Epoch 15, batch 3150, loss[loss=0.1766, simple_loss=0.2526, pruned_loss=0.05029, over 4912.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2515, pruned_loss=0.05755, over 947772.02 frames. ], batch size: 36, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:25:41,381 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.524e+02 1.821e+02 2.258e+02 3.585e+02, threshold=3.643e+02, percent-clipped=2.0 +2023-03-26 18:25:42,587 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:26:12,519 INFO [finetune.py:976] (0/7) Epoch 15, batch 3200, loss[loss=0.1811, simple_loss=0.2461, pruned_loss=0.05805, over 4815.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2496, pruned_loss=0.05759, over 948987.04 frames. ], batch size: 33, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:26:25,182 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-26 18:26:47,751 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-26 18:26:55,876 INFO [finetune.py:976] (0/7) Epoch 15, batch 3250, loss[loss=0.1719, simple_loss=0.2405, pruned_loss=0.05167, over 4891.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2501, pruned_loss=0.05806, over 951008.86 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:27:00,086 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.648e+01 1.539e+02 1.854e+02 2.232e+02 3.646e+02, threshold=3.708e+02, percent-clipped=1.0 +2023-03-26 18:27:03,146 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5279, 2.3807, 1.8801, 2.6284, 2.4424, 2.0076, 2.9000, 2.5030], + device='cuda:0'), covar=tensor([0.1298, 0.2261, 0.3238, 0.2664, 0.2631, 0.1761, 0.3358, 0.1770], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0186, 0.0234, 0.0252, 0.0243, 0.0200, 0.0211, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:27:29,590 INFO [finetune.py:976] (0/7) Epoch 15, batch 3300, loss[loss=0.1799, simple_loss=0.2648, pruned_loss=0.04749, over 4804.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2533, pruned_loss=0.05878, over 951038.54 frames. ], batch size: 45, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:27:30,986 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-03-26 18:28:07,466 INFO [finetune.py:976] (0/7) Epoch 15, batch 3350, loss[loss=0.193, simple_loss=0.2649, pruned_loss=0.06049, over 4860.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2557, pruned_loss=0.05958, over 949732.93 frames. ], batch size: 34, lr: 3.49e-03, grad_scale: 64.0 +2023-03-26 18:28:14,152 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2565, 2.3117, 2.4382, 1.6022, 2.3177, 2.5088, 2.5051, 2.0680], + device='cuda:0'), covar=tensor([0.0579, 0.0583, 0.0621, 0.0932, 0.0592, 0.0629, 0.0571, 0.0937], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0136, 0.0143, 0.0125, 0.0125, 0.0143, 0.0143, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:28:14,627 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.792e+02 2.041e+02 2.510e+02 5.102e+02, threshold=4.082e+02, percent-clipped=3.0 +2023-03-26 18:28:21,854 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:28:54,246 INFO [finetune.py:976] (0/7) Epoch 15, batch 3400, loss[loss=0.1988, simple_loss=0.2671, pruned_loss=0.06528, over 4883.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2562, pruned_loss=0.05999, over 949047.27 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 64.0 +2023-03-26 18:29:24,882 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83619.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:29:37,311 INFO [finetune.py:976] (0/7) Epoch 15, batch 3450, loss[loss=0.1581, simple_loss=0.2332, pruned_loss=0.04147, over 4736.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2553, pruned_loss=0.05911, over 950291.21 frames. ], batch size: 54, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:29:39,045 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:29:42,000 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.651e+02 1.928e+02 2.236e+02 3.717e+02, threshold=3.855e+02, percent-clipped=0.0 +2023-03-26 18:29:50,501 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-26 18:29:57,379 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83667.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:30:11,001 INFO [finetune.py:976] (0/7) Epoch 15, batch 3500, loss[loss=0.1593, simple_loss=0.2262, pruned_loss=0.04617, over 4894.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.253, pruned_loss=0.05828, over 951108.86 frames. ], batch size: 32, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:30:37,629 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2106, 1.3912, 1.6518, 1.0599, 1.2128, 1.4710, 1.3823, 1.6533], + device='cuda:0'), covar=tensor([0.1375, 0.2142, 0.1143, 0.1411, 0.0945, 0.1204, 0.2978, 0.0733], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0204, 0.0193, 0.0190, 0.0176, 0.0212, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:30:43,627 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2808, 1.7922, 2.1697, 2.1614, 1.8953, 1.9110, 2.1511, 2.0199], + device='cuda:0'), covar=tensor([0.3831, 0.4175, 0.3419, 0.4090, 0.5130, 0.3730, 0.4901, 0.3205], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0239, 0.0257, 0.0269, 0.0268, 0.0240, 0.0281, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:30:44,676 INFO [finetune.py:976] (0/7) Epoch 15, batch 3550, loss[loss=0.1935, simple_loss=0.2552, pruned_loss=0.06591, over 4916.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2506, pruned_loss=0.0581, over 952393.14 frames. ], batch size: 37, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:30:49,415 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.222e+01 1.573e+02 1.880e+02 2.102e+02 4.250e+02, threshold=3.760e+02, percent-clipped=2.0 +2023-03-26 18:30:59,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:31:18,475 INFO [finetune.py:976] (0/7) Epoch 15, batch 3600, loss[loss=0.135, simple_loss=0.2033, pruned_loss=0.03336, over 4805.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2485, pruned_loss=0.05777, over 953253.74 frames. ], batch size: 25, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:31:48,185 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:31:59,979 INFO [finetune.py:976] (0/7) Epoch 15, batch 3650, loss[loss=0.2365, simple_loss=0.3023, pruned_loss=0.08536, over 4748.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2505, pruned_loss=0.05822, over 953644.24 frames. ], batch size: 54, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:32:04,778 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.603e+02 1.941e+02 2.306e+02 4.863e+02, threshold=3.882e+02, percent-clipped=1.0 +2023-03-26 18:32:09,552 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:32:09,622 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7994, 0.9773, 1.8607, 1.7451, 1.5918, 1.5431, 1.6381, 1.7090], + device='cuda:0'), covar=tensor([0.3619, 0.4120, 0.3117, 0.3538, 0.4590, 0.3632, 0.4435, 0.3074], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0239, 0.0258, 0.0269, 0.0268, 0.0240, 0.0281, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:32:33,847 INFO [finetune.py:976] (0/7) Epoch 15, batch 3700, loss[loss=0.1847, simple_loss=0.2462, pruned_loss=0.06161, over 4739.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2545, pruned_loss=0.05907, over 954287.10 frames. ], batch size: 23, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:32:41,698 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:07,588 INFO [finetune.py:976] (0/7) Epoch 15, batch 3750, loss[loss=0.1956, simple_loss=0.2755, pruned_loss=0.05783, over 4759.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2564, pruned_loss=0.05961, over 955720.32 frames. ], batch size: 28, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:33:08,900 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:11,799 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.596e+02 1.977e+02 2.275e+02 5.079e+02, threshold=3.955e+02, percent-clipped=1.0 +2023-03-26 18:33:14,837 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:20,992 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6981, 1.4909, 1.0112, 0.2611, 1.2761, 1.4777, 1.4226, 1.3740], + device='cuda:0'), covar=tensor([0.0957, 0.0782, 0.1419, 0.1929, 0.1463, 0.2238, 0.2233, 0.0872], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0194, 0.0199, 0.0183, 0.0212, 0.0206, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:33:32,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2363, 1.2392, 1.3203, 0.6357, 1.1909, 1.4858, 1.4412, 1.2557], + device='cuda:0'), covar=tensor([0.0883, 0.0555, 0.0501, 0.0565, 0.0574, 0.0618, 0.0383, 0.0714], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0128, 0.0130, 0.0126, 0.0141, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.1829e-05, 1.0960e-04, 8.8797e-05, 9.1832e-05, 9.2236e-05, 9.0969e-05, + 1.0157e-04, 1.0605e-04], device='cuda:0') +2023-03-26 18:33:43,547 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2407, 1.9714, 2.5995, 4.0784, 2.8792, 2.8114, 0.9352, 3.3925], + device='cuda:0'), covar=tensor([0.1675, 0.1429, 0.1328, 0.0459, 0.0689, 0.1487, 0.2021, 0.0408], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0131, 0.0162, 0.0099, 0.0136, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:33:53,423 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83985.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:55,619 INFO [finetune.py:976] (0/7) Epoch 15, batch 3800, loss[loss=0.1888, simple_loss=0.2599, pruned_loss=0.0589, over 4915.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2579, pruned_loss=0.06007, over 955958.35 frames. ], batch size: 42, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:33:55,678 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:34:03,488 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-84000.pt +2023-03-26 18:34:11,213 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84010.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:34:13,617 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1005, 2.0000, 1.7398, 2.0789, 1.8501, 1.8615, 1.9180, 2.6364], + device='cuda:0'), covar=tensor([0.4226, 0.4821, 0.3656, 0.4318, 0.4433, 0.2730, 0.4183, 0.1907], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0258, 0.0225, 0.0274, 0.0245, 0.0214, 0.0248, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:34:36,932 INFO [finetune.py:976] (0/7) Epoch 15, batch 3850, loss[loss=0.1785, simple_loss=0.2387, pruned_loss=0.05919, over 4865.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2546, pruned_loss=0.05806, over 957180.38 frames. ], batch size: 34, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:34:41,716 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.557e+01 1.496e+02 1.862e+02 2.338e+02 3.560e+02, threshold=3.724e+02, percent-clipped=0.0 +2023-03-26 18:34:42,452 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:34:43,078 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3392, 1.4664, 1.5798, 0.7463, 1.5172, 1.7264, 1.8358, 1.4260], + device='cuda:0'), covar=tensor([0.0853, 0.0574, 0.0459, 0.0525, 0.0424, 0.0640, 0.0301, 0.0675], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0129, 0.0130, 0.0126, 0.0142, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.2085e-05, 1.0994e-04, 8.9039e-05, 9.2218e-05, 9.2537e-05, 9.1143e-05, + 1.0219e-04, 1.0638e-04], device='cuda:0') +2023-03-26 18:35:10,032 INFO [finetune.py:976] (0/7) Epoch 15, batch 3900, loss[loss=0.2127, simple_loss=0.278, pruned_loss=0.07367, over 4918.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2523, pruned_loss=0.05775, over 957892.04 frames. ], batch size: 36, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:35:16,521 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84097.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:35:28,898 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:35:43,581 INFO [finetune.py:976] (0/7) Epoch 15, batch 3950, loss[loss=0.1505, simple_loss=0.2199, pruned_loss=0.04056, over 4779.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2489, pruned_loss=0.05664, over 958202.68 frames. ], batch size: 26, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:35:47,769 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.469e+02 1.885e+02 2.278e+02 4.120e+02, threshold=3.770e+02, percent-clipped=1.0 +2023-03-26 18:35:51,962 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 18:35:57,231 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:13,535 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 18:36:16,808 INFO [finetune.py:976] (0/7) Epoch 15, batch 4000, loss[loss=0.1976, simple_loss=0.2671, pruned_loss=0.06402, over 4911.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2479, pruned_loss=0.05605, over 958621.76 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:36:24,506 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:26,855 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4943, 1.3194, 1.5218, 0.8883, 1.5112, 1.7980, 1.6370, 1.3756], + device='cuda:0'), covar=tensor([0.0968, 0.0899, 0.0527, 0.0593, 0.0494, 0.0600, 0.0470, 0.0670], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0124, 0.0128, 0.0130, 0.0126, 0.0141, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.1550e-05, 1.0959e-04, 8.8792e-05, 9.1931e-05, 9.2165e-05, 9.0842e-05, + 1.0179e-04, 1.0597e-04], device='cuda:0') +2023-03-26 18:36:57,612 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:59,238 INFO [finetune.py:976] (0/7) Epoch 15, batch 4050, loss[loss=0.1677, simple_loss=0.2295, pruned_loss=0.05292, over 4826.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2536, pruned_loss=0.05856, over 956316.04 frames. ], batch size: 30, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:37:07,827 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.586e+02 1.909e+02 2.268e+02 5.729e+02, threshold=3.818e+02, percent-clipped=2.0 +2023-03-26 18:37:21,556 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:37:23,898 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 18:37:39,956 INFO [finetune.py:976] (0/7) Epoch 15, batch 4100, loss[loss=0.2287, simple_loss=0.2928, pruned_loss=0.08228, over 4733.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2556, pruned_loss=0.05845, over 956554.82 frames. ], batch size: 59, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:37:46,461 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:37:52,177 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84305.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:38:10,145 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 18:38:11,860 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:38:13,433 INFO [finetune.py:976] (0/7) Epoch 15, batch 4150, loss[loss=0.2067, simple_loss=0.2807, pruned_loss=0.06637, over 4812.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2572, pruned_loss=0.05943, over 955183.45 frames. ], batch size: 47, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:38:15,331 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84341.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:38:18,112 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.618e+02 1.997e+02 2.307e+02 7.274e+02, threshold=3.993e+02, percent-clipped=3.0 +2023-03-26 18:38:48,113 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 18:38:50,307 INFO [finetune.py:976] (0/7) Epoch 15, batch 4200, loss[loss=0.1625, simple_loss=0.2371, pruned_loss=0.04394, over 4889.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2567, pruned_loss=0.05948, over 950814.59 frames. ], batch size: 32, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:39:04,804 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:17,872 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84416.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:20,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7498, 3.8198, 3.6432, 1.7229, 3.9096, 2.9028, 0.7813, 2.8796], + device='cuda:0'), covar=tensor([0.2460, 0.1759, 0.1309, 0.3104, 0.0963, 0.0964, 0.4221, 0.1256], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0159, 0.0127, 0.0157, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 18:39:31,951 INFO [finetune.py:976] (0/7) Epoch 15, batch 4250, loss[loss=0.1941, simple_loss=0.2663, pruned_loss=0.06098, over 4890.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2544, pruned_loss=0.05897, over 952675.96 frames. ], batch size: 35, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:39:36,665 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.324e+01 1.559e+02 1.825e+02 2.300e+02 4.289e+02, threshold=3.650e+02, percent-clipped=1.0 +2023-03-26 18:39:47,004 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 18:39:47,477 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84453.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:58,291 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:40:14,294 INFO [finetune.py:976] (0/7) Epoch 15, batch 4300, loss[loss=0.1631, simple_loss=0.2223, pruned_loss=0.05197, over 4827.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2521, pruned_loss=0.05857, over 952240.71 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:40:36,218 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-26 18:40:39,091 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:40:39,730 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0702, 0.9392, 0.8802, 1.1399, 1.2329, 1.1461, 1.0434, 0.8886], + device='cuda:0'), covar=tensor([0.0320, 0.0304, 0.0588, 0.0289, 0.0250, 0.0372, 0.0310, 0.0372], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0142, 0.0113, 0.0100, 0.0106, 0.0097, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.2130e-05, 8.3931e-05, 1.1208e-04, 8.7162e-05, 7.7899e-05, 7.8458e-05, + 7.2903e-05, 8.1888e-05], device='cuda:0') +2023-03-26 18:40:42,179 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4361, 2.2663, 2.1079, 2.3785, 2.1442, 2.1779, 2.1897, 2.8632], + device='cuda:0'), covar=tensor([0.3731, 0.4563, 0.3207, 0.3948, 0.3975, 0.2412, 0.3707, 0.1803], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0258, 0.0225, 0.0275, 0.0247, 0.0215, 0.0248, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:40:47,844 INFO [finetune.py:976] (0/7) Epoch 15, batch 4350, loss[loss=0.1623, simple_loss=0.2358, pruned_loss=0.04439, over 4848.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2494, pruned_loss=0.05761, over 953257.75 frames. ], batch size: 49, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:40:51,073 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6171, 1.0447, 0.7147, 1.4195, 2.0253, 0.7741, 1.3496, 1.4157], + device='cuda:0'), covar=tensor([0.1524, 0.2306, 0.1934, 0.1296, 0.1892, 0.2022, 0.1626, 0.2057], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:40:52,224 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.502e+02 1.820e+02 2.196e+02 3.984e+02, threshold=3.641e+02, percent-clipped=2.0 +2023-03-26 18:40:58,843 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:19,624 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:21,103 INFO [finetune.py:976] (0/7) Epoch 15, batch 4400, loss[loss=0.187, simple_loss=0.264, pruned_loss=0.05499, over 4821.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2482, pruned_loss=0.05703, over 951728.03 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:41:24,128 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:32,774 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:46,492 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5769, 1.5046, 1.4824, 1.5475, 1.0129, 2.9339, 1.0305, 1.5108], + device='cuda:0'), covar=tensor([0.3207, 0.2473, 0.2123, 0.2343, 0.1930, 0.0300, 0.2558, 0.1254], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:41:51,814 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84633.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:54,830 INFO [finetune.py:976] (0/7) Epoch 15, batch 4450, loss[loss=0.1801, simple_loss=0.2537, pruned_loss=0.05322, over 4821.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2516, pruned_loss=0.05789, over 948724.00 frames. ], batch size: 40, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:41:57,736 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:01,983 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.595e+02 1.952e+02 2.292e+02 4.719e+02, threshold=3.904e+02, percent-clipped=1.0 +2023-03-26 18:42:02,777 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1634, 2.0922, 1.8036, 2.2440, 2.1005, 1.8757, 2.4418, 2.2221], + device='cuda:0'), covar=tensor([0.1121, 0.2053, 0.2467, 0.2211, 0.2015, 0.1371, 0.2890, 0.1457], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0188, 0.0234, 0.0254, 0.0244, 0.0201, 0.0212, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:42:07,102 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:23,812 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7744, 1.5594, 1.2626, 1.3060, 1.9934, 2.1371, 1.9174, 1.7745], + device='cuda:0'), covar=tensor([0.0343, 0.0450, 0.0804, 0.0436, 0.0304, 0.0534, 0.0317, 0.0446], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0143, 0.0114, 0.0100, 0.0107, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2789e-05, 8.4637e-05, 1.1315e-04, 8.7896e-05, 7.8446e-05, 7.9423e-05, + 7.3471e-05, 8.2759e-05], device='cuda:0') +2023-03-26 18:42:44,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5770, 1.6149, 1.9948, 3.1924, 2.1167, 2.3418, 0.9948, 2.5966], + device='cuda:0'), covar=tensor([0.1844, 0.1402, 0.1359, 0.0566, 0.0845, 0.1316, 0.1923, 0.0579], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0100, 0.0138, 0.0124, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:42:46,431 INFO [finetune.py:976] (0/7) Epoch 15, batch 4500, loss[loss=0.1864, simple_loss=0.2639, pruned_loss=0.05445, over 4791.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2539, pruned_loss=0.05804, over 949083.33 frames. ], batch size: 29, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:42:47,105 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84689.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:47,129 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.9547, 4.3234, 4.4872, 4.7617, 4.6330, 4.4547, 5.0890, 1.6639], + device='cuda:0'), covar=tensor([0.0774, 0.0837, 0.0705, 0.0835, 0.1319, 0.1423, 0.0555, 0.5679], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0276, 0.0292, 0.0334, 0.0283, 0.0300, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:42:49,441 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:50,725 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84694.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:43:01,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8634, 1.7984, 1.5774, 1.9728, 2.4635, 1.9486, 1.7239, 1.5004], + device='cuda:0'), covar=tensor([0.2261, 0.2017, 0.1971, 0.1702, 0.1679, 0.1215, 0.2283, 0.1989], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0207, 0.0209, 0.0191, 0.0241, 0.0184, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:43:04,306 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5064, 2.3628, 1.7744, 2.6889, 2.3642, 2.0433, 2.9400, 2.4733], + device='cuda:0'), covar=tensor([0.1221, 0.2178, 0.3042, 0.2421, 0.2441, 0.1655, 0.2608, 0.1716], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0188, 0.0234, 0.0253, 0.0244, 0.0201, 0.0211, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:43:06,205 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 18:43:19,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5478, 1.5649, 2.0455, 3.3088, 2.2340, 2.3242, 0.8480, 2.7250], + device='cuda:0'), covar=tensor([0.1805, 0.1395, 0.1301, 0.0612, 0.0790, 0.1522, 0.1913, 0.0514], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0133, 0.0165, 0.0100, 0.0138, 0.0125, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:43:20,117 INFO [finetune.py:976] (0/7) Epoch 15, batch 4550, loss[loss=0.1926, simple_loss=0.2629, pruned_loss=0.06115, over 4913.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2563, pruned_loss=0.05937, over 949509.43 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:43:25,286 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.592e+02 2.005e+02 2.406e+02 4.528e+02, threshold=4.009e+02, percent-clipped=3.0 +2023-03-26 18:43:25,978 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7681, 3.8256, 3.6365, 1.9549, 4.0288, 3.0236, 1.2215, 2.7591], + device='cuda:0'), covar=tensor([0.2337, 0.1590, 0.1491, 0.3197, 0.0959, 0.1019, 0.4038, 0.1535], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0160, 0.0128, 0.0158, 0.0123, 0.0145, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 18:43:30,258 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:43:53,702 INFO [finetune.py:976] (0/7) Epoch 15, batch 4600, loss[loss=0.1598, simple_loss=0.2238, pruned_loss=0.04791, over 4724.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2563, pruned_loss=0.05942, over 951156.94 frames. ], batch size: 23, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:43:56,190 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8850, 3.4052, 3.5750, 3.7796, 3.6054, 3.3306, 3.9376, 1.2129], + device='cuda:0'), covar=tensor([0.0948, 0.0977, 0.0986, 0.1003, 0.1496, 0.1724, 0.0831, 0.5854], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0245, 0.0277, 0.0292, 0.0335, 0.0284, 0.0301, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:44:06,854 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84800.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:07,395 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:28,642 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3623, 2.0272, 2.8535, 1.6769, 2.5268, 2.7034, 1.9933, 2.8449], + device='cuda:0'), covar=tensor([0.1450, 0.2199, 0.1788, 0.2339, 0.0973, 0.1519, 0.2826, 0.1013], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0205, 0.0194, 0.0191, 0.0176, 0.0214, 0.0219, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:44:36,214 INFO [finetune.py:976] (0/7) Epoch 15, batch 4650, loss[loss=0.1888, simple_loss=0.2497, pruned_loss=0.06391, over 4900.00 frames. ], tot_loss[loss=0.186, simple_loss=0.254, pruned_loss=0.05896, over 952802.53 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:44:37,524 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7688, 1.2286, 0.8102, 1.7197, 2.1755, 1.4772, 1.4450, 1.7002], + device='cuda:0'), covar=tensor([0.1456, 0.2170, 0.2037, 0.1285, 0.1815, 0.1934, 0.1548, 0.1847], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0093, 0.0118, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:44:38,844 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 18:44:40,389 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.584e+02 1.933e+02 2.372e+02 3.946e+02, threshold=3.865e+02, percent-clipped=0.0 +2023-03-26 18:44:45,933 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 18:44:47,589 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84855.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:48,939 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=6.15 vs. limit=5.0 +2023-03-26 18:44:56,259 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84861.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:17,963 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84881.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:26,374 INFO [finetune.py:976] (0/7) Epoch 15, batch 4700, loss[loss=0.1603, simple_loss=0.226, pruned_loss=0.04733, over 4923.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2508, pruned_loss=0.0576, over 953967.14 frames. ], batch size: 37, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:45:29,326 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:31,174 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3924, 1.5345, 1.8638, 1.7191, 1.6576, 3.2832, 1.3527, 1.6510], + device='cuda:0'), covar=tensor([0.0995, 0.1698, 0.1277, 0.1031, 0.1449, 0.0294, 0.1445, 0.1514], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0078, 0.0092, 0.0080, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:45:36,039 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:59,760 INFO [finetune.py:976] (0/7) Epoch 15, batch 4750, loss[loss=0.177, simple_loss=0.2574, pruned_loss=0.04835, over 4907.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2495, pruned_loss=0.05744, over 955294.24 frames. ], batch size: 43, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:46:01,515 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:04,957 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.651e+02 1.892e+02 2.436e+02 4.596e+02, threshold=3.784e+02, percent-clipped=1.0 +2023-03-26 18:46:08,784 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:16,352 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4369, 1.4124, 1.4240, 1.5604, 1.5251, 2.9092, 1.2330, 1.4871], + device='cuda:0'), covar=tensor([0.0936, 0.1774, 0.1159, 0.0987, 0.1575, 0.0286, 0.1532, 0.1683], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0092, 0.0080, 0.0085, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:46:33,699 INFO [finetune.py:976] (0/7) Epoch 15, batch 4800, loss[loss=0.1649, simple_loss=0.25, pruned_loss=0.03993, over 4836.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2514, pruned_loss=0.05771, over 955052.56 frames. ], batch size: 47, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:46:34,384 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:36,731 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:50,494 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85012.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:47:07,603 INFO [finetune.py:976] (0/7) Epoch 15, batch 4850, loss[loss=0.2054, simple_loss=0.2746, pruned_loss=0.06813, over 4843.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2541, pruned_loss=0.0585, over 952133.93 frames. ], batch size: 47, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:47:08,851 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:47:12,313 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.585e+02 1.858e+02 2.141e+02 6.123e+02, threshold=3.716e+02, percent-clipped=1.0 +2023-03-26 18:47:50,157 INFO [finetune.py:976] (0/7) Epoch 15, batch 4900, loss[loss=0.1867, simple_loss=0.254, pruned_loss=0.05971, over 4910.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2562, pruned_loss=0.05945, over 952694.10 frames. ], batch size: 37, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:47:59,528 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4494, 1.4537, 1.6895, 1.7450, 1.7487, 3.3246, 1.3918, 1.6133], + device='cuda:0'), covar=tensor([0.1000, 0.1869, 0.1090, 0.1024, 0.1528, 0.0270, 0.1529, 0.1669], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:48:06,014 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85107.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:11,337 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:26,707 INFO [finetune.py:976] (0/7) Epoch 15, batch 4950, loss[loss=0.1651, simple_loss=0.2441, pruned_loss=0.04303, over 4227.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2567, pruned_loss=0.05939, over 952675.07 frames. ], batch size: 65, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:48:31,435 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.624e+02 1.884e+02 2.194e+02 3.725e+02, threshold=3.769e+02, percent-clipped=1.0 +2023-03-26 18:48:39,195 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85156.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:47,008 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85168.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:48:52,408 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:55,823 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:49:00,448 INFO [finetune.py:976] (0/7) Epoch 15, batch 5000, loss[loss=0.1867, simple_loss=0.2482, pruned_loss=0.06258, over 4895.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2537, pruned_loss=0.05773, over 951666.64 frames. ], batch size: 35, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:49:01,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7854, 1.7191, 2.1563, 1.3586, 1.8373, 2.0798, 1.6522, 2.2639], + device='cuda:0'), covar=tensor([0.1296, 0.1993, 0.1387, 0.1824, 0.0987, 0.1295, 0.2655, 0.0877], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0200, 0.0190, 0.0188, 0.0174, 0.0210, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:49:36,185 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:49:42,104 INFO [finetune.py:976] (0/7) Epoch 15, batch 5050, loss[loss=0.1746, simple_loss=0.2397, pruned_loss=0.05476, over 4865.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2515, pruned_loss=0.05777, over 952178.17 frames. ], batch size: 34, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:49:46,816 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.593e+02 1.872e+02 2.269e+02 5.264e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 18:50:22,632 INFO [finetune.py:976] (0/7) Epoch 15, batch 5100, loss[loss=0.1769, simple_loss=0.2389, pruned_loss=0.0574, over 4787.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2485, pruned_loss=0.05681, over 954113.00 frames. ], batch size: 29, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:50:23,315 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:50:42,800 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:51:02,804 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:51:03,346 INFO [finetune.py:976] (0/7) Epoch 15, batch 5150, loss[loss=0.1862, simple_loss=0.2583, pruned_loss=0.05711, over 4827.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2489, pruned_loss=0.05769, over 955368.51 frames. ], batch size: 30, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:51:08,102 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.526e+02 1.888e+02 2.256e+02 3.382e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 18:51:37,046 INFO [finetune.py:976] (0/7) Epoch 15, batch 5200, loss[loss=0.2344, simple_loss=0.2989, pruned_loss=0.08495, over 4886.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2539, pruned_loss=0.05934, over 954269.93 frames. ], batch size: 32, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:51:50,449 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-26 18:51:54,143 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0122, 1.8769, 1.6276, 1.8901, 1.7760, 1.8370, 1.8280, 2.5319], + device='cuda:0'), covar=tensor([0.3867, 0.4656, 0.3579, 0.4469, 0.4629, 0.2517, 0.4563, 0.1738], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0259, 0.0226, 0.0275, 0.0247, 0.0214, 0.0248, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:52:04,182 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:04,980 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 18:52:09,545 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1146, 2.0465, 1.8073, 2.1493, 2.7468, 2.0343, 1.9829, 1.6032], + device='cuda:0'), covar=tensor([0.2098, 0.1894, 0.1794, 0.1598, 0.1605, 0.1141, 0.2152, 0.1758], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0206, 0.0209, 0.0190, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:52:10,614 INFO [finetune.py:976] (0/7) Epoch 15, batch 5250, loss[loss=0.1488, simple_loss=0.2273, pruned_loss=0.03511, over 4927.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2549, pruned_loss=0.05922, over 953436.00 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:52:15,821 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.657e+02 1.928e+02 2.523e+02 8.274e+02, threshold=3.856e+02, percent-clipped=2.0 +2023-03-26 18:52:23,089 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:27,680 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85463.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:52:33,021 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:40,796 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:44,235 INFO [finetune.py:976] (0/7) Epoch 15, batch 5300, loss[loss=0.1756, simple_loss=0.2518, pruned_loss=0.04965, over 4842.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2572, pruned_loss=0.06002, over 952076.06 frames. ], batch size: 47, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:52:44,962 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85489.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:54,981 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:55,154 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 18:53:14,958 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7741, 1.7165, 1.6877, 1.6856, 1.1152, 2.9808, 1.2536, 1.6438], + device='cuda:0'), covar=tensor([0.2787, 0.2073, 0.1843, 0.2058, 0.1763, 0.0252, 0.2241, 0.1128], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:53:19,696 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:53:24,906 INFO [finetune.py:976] (0/7) Epoch 15, batch 5350, loss[loss=0.1901, simple_loss=0.2586, pruned_loss=0.06083, over 4845.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2569, pruned_loss=0.05922, over 953435.30 frames. ], batch size: 47, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:53:28,678 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85544.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:53:29,177 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.542e+02 1.806e+02 2.197e+02 4.190e+02, threshold=3.613e+02, percent-clipped=2.0 +2023-03-26 18:53:58,024 INFO [finetune.py:976] (0/7) Epoch 15, batch 5400, loss[loss=0.2098, simple_loss=0.2724, pruned_loss=0.07356, over 4930.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2549, pruned_loss=0.05902, over 954037.60 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:54:00,455 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:11,212 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:22,027 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 18:54:23,532 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85625.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:31,764 INFO [finetune.py:976] (0/7) Epoch 15, batch 5450, loss[loss=0.1266, simple_loss=0.2045, pruned_loss=0.02433, over 4744.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2526, pruned_loss=0.05815, over 955241.53 frames. ], batch size: 59, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:54:41,082 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.884e+01 1.516e+02 1.902e+02 2.390e+02 5.288e+02, threshold=3.804e+02, percent-clipped=4.0 +2023-03-26 18:54:52,063 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:55:04,692 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 18:55:12,290 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.1844, 4.5924, 4.7364, 5.0024, 4.9106, 4.5550, 5.2879, 1.4851], + device='cuda:0'), covar=tensor([0.0639, 0.0805, 0.0839, 0.0837, 0.1082, 0.1574, 0.0527, 0.5757], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0242, 0.0271, 0.0289, 0.0330, 0.0279, 0.0295, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:55:13,406 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2109, 1.2935, 1.3746, 0.6591, 1.2966, 1.5038, 1.6030, 1.2823], + device='cuda:0'), covar=tensor([0.0870, 0.0581, 0.0533, 0.0519, 0.0532, 0.0673, 0.0361, 0.0657], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0153, 0.0126, 0.0130, 0.0132, 0.0128, 0.0144, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.3405e-05, 1.1138e-04, 9.0208e-05, 9.3352e-05, 9.3642e-05, 9.2668e-05, + 1.0413e-04, 1.0734e-04], device='cuda:0') +2023-03-26 18:55:16,861 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85686.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:55:17,972 INFO [finetune.py:976] (0/7) Epoch 15, batch 5500, loss[loss=0.1437, simple_loss=0.2184, pruned_loss=0.03446, over 4847.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2491, pruned_loss=0.05684, over 956657.40 frames. ], batch size: 44, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:02,559 INFO [finetune.py:976] (0/7) Epoch 15, batch 5550, loss[loss=0.2123, simple_loss=0.2902, pruned_loss=0.06722, over 4860.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2517, pruned_loss=0.05832, over 955871.31 frames. ], batch size: 49, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:03,820 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9433, 1.4812, 0.8132, 1.8153, 2.1453, 1.5682, 1.8368, 1.7990], + device='cuda:0'), covar=tensor([0.1392, 0.1857, 0.2009, 0.1189, 0.1886, 0.1932, 0.1221, 0.1915], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0093, 0.0118, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:56:03,876 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1972, 2.0679, 1.7139, 2.1132, 1.9842, 1.9274, 1.9989, 2.7001], + device='cuda:0'), covar=tensor([0.3979, 0.5188, 0.3694, 0.4454, 0.4266, 0.2647, 0.4263, 0.1851], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0260, 0.0226, 0.0275, 0.0247, 0.0215, 0.0248, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:56:06,725 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.623e+01 1.589e+02 1.875e+02 2.150e+02 4.153e+02, threshold=3.750e+02, percent-clipped=1.0 +2023-03-26 18:56:08,069 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6371, 2.4409, 2.2268, 2.5469, 2.4981, 2.3429, 2.8855, 2.5678], + device='cuda:0'), covar=tensor([0.1208, 0.2038, 0.2778, 0.2428, 0.2186, 0.1430, 0.2600, 0.1849], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0190, 0.0236, 0.0257, 0.0246, 0.0204, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:56:19,316 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:19,983 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5194, 1.3890, 1.2694, 1.4608, 1.7371, 1.4511, 1.1978, 1.2784], + device='cuda:0'), covar=tensor([0.1970, 0.1861, 0.1802, 0.1577, 0.1441, 0.1217, 0.2259, 0.1781], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0206, 0.0209, 0.0190, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:56:24,622 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:25,808 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2134, 1.6878, 2.4356, 1.5499, 2.1755, 2.4707, 1.7496, 2.5424], + device='cuda:0'), covar=tensor([0.1023, 0.2149, 0.1115, 0.1731, 0.0833, 0.1103, 0.2539, 0.0727], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0193, 0.0191, 0.0177, 0.0214, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:56:32,619 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:34,954 INFO [finetune.py:976] (0/7) Epoch 15, batch 5600, loss[loss=0.192, simple_loss=0.2631, pruned_loss=0.06044, over 4907.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.254, pruned_loss=0.05887, over 954266.60 frames. ], batch size: 43, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:48,425 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:50,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 18:56:53,122 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:53,153 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6190, 1.4728, 2.0173, 3.1298, 2.2089, 2.1086, 1.4281, 2.4986], + device='cuda:0'), covar=tensor([0.1664, 0.1490, 0.1210, 0.0584, 0.0718, 0.1672, 0.1392, 0.0544], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0139, 0.0125, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:57:04,149 INFO [finetune.py:976] (0/7) Epoch 15, batch 5650, loss[loss=0.1682, simple_loss=0.2337, pruned_loss=0.05139, over 4768.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2565, pruned_loss=0.0597, over 954310.54 frames. ], batch size: 28, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:57:04,766 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85839.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:08,244 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.126e+02 1.563e+02 1.888e+02 2.328e+02 3.522e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 18:57:08,364 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6741, 1.5368, 1.3473, 1.5272, 1.9424, 1.8141, 1.5616, 1.3466], + device='cuda:0'), covar=tensor([0.0274, 0.0287, 0.0629, 0.0326, 0.0195, 0.0454, 0.0327, 0.0373], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0144, 0.0113, 0.0101, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2951e-05, 8.4707e-05, 1.1389e-04, 8.7774e-05, 7.8576e-05, 7.9414e-05, + 7.3588e-05, 8.3081e-05], device='cuda:0') +2023-03-26 18:57:26,029 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85875.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:32,550 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:33,723 INFO [finetune.py:976] (0/7) Epoch 15, batch 5700, loss[loss=0.1359, simple_loss=0.1996, pruned_loss=0.03611, over 3898.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2527, pruned_loss=0.05893, over 937798.74 frames. ], batch size: 17, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:57:50,720 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-15.pt +2023-03-26 18:58:02,782 INFO [finetune.py:976] (0/7) Epoch 16, batch 0, loss[loss=0.2055, simple_loss=0.2745, pruned_loss=0.06827, over 4911.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2745, pruned_loss=0.06827, over 4911.00 frames. ], batch size: 46, lr: 3.46e-03, grad_scale: 64.0 +2023-03-26 18:58:02,783 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 18:58:09,950 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8241, 1.3518, 0.9296, 1.6179, 2.1377, 1.1421, 1.6183, 1.6446], + device='cuda:0'), covar=tensor([0.1318, 0.1811, 0.1732, 0.1149, 0.1727, 0.1953, 0.1250, 0.1852], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 18:58:17,921 INFO [finetune.py:1010] (0/7) Epoch 16, validation: loss=0.1572, simple_loss=0.2278, pruned_loss=0.04329, over 2265189.00 frames. +2023-03-26 18:58:17,921 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 18:58:22,808 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4829, 1.8010, 1.5247, 1.5220, 2.0819, 2.0496, 1.8249, 1.7890], + device='cuda:0'), covar=tensor([0.0559, 0.0334, 0.0569, 0.0355, 0.0279, 0.0553, 0.0326, 0.0378], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0109, 0.0143, 0.0113, 0.0100, 0.0107, 0.0097, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2658e-05, 8.4242e-05, 1.1341e-04, 8.7361e-05, 7.8285e-05, 7.8926e-05, + 7.3140e-05, 8.2796e-05], device='cuda:0') +2023-03-26 18:58:25,992 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 18:58:26,968 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85930.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:58:29,348 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6326, 1.5388, 1.4631, 1.5428, 1.0637, 3.4344, 1.3593, 1.7471], + device='cuda:0'), covar=tensor([0.3220, 0.2347, 0.2058, 0.2262, 0.1825, 0.0203, 0.2518, 0.1272], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0114, 0.0118, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 18:58:30,590 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:58:36,427 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.566e+02 1.783e+02 2.274e+02 8.459e+02, threshold=3.567e+02, percent-clipped=4.0 +2023-03-26 18:58:40,700 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4627, 2.4252, 1.9197, 2.6721, 2.5537, 2.0632, 3.0241, 2.4636], + device='cuda:0'), covar=tensor([0.1367, 0.2375, 0.3106, 0.2466, 0.2434, 0.1727, 0.3070, 0.1797], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0188, 0.0234, 0.0254, 0.0244, 0.0202, 0.0212, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 18:58:44,320 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:58:49,654 INFO [finetune.py:976] (0/7) Epoch 16, batch 50, loss[loss=0.1887, simple_loss=0.2566, pruned_loss=0.06042, over 4863.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2593, pruned_loss=0.06241, over 214281.49 frames. ], batch size: 34, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 18:58:59,303 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85981.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:59:05,877 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85991.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:59:11,762 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-86000.pt +2023-03-26 18:59:14,745 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:59:17,225 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 18:59:23,676 INFO [finetune.py:976] (0/7) Epoch 16, batch 100, loss[loss=0.1793, simple_loss=0.2589, pruned_loss=0.04983, over 4864.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2525, pruned_loss=0.05917, over 378501.84 frames. ], batch size: 34, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 18:59:24,986 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:59:44,131 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.611e+02 1.877e+02 2.147e+02 3.763e+02, threshold=3.754e+02, percent-clipped=3.0 +2023-03-26 18:59:47,930 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:00,105 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86063.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:06,620 INFO [finetune.py:976] (0/7) Epoch 16, batch 150, loss[loss=0.1623, simple_loss=0.2228, pruned_loss=0.0509, over 4763.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2472, pruned_loss=0.0575, over 507651.53 frames. ], batch size: 23, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:00:08,574 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:17,306 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-26 19:00:27,238 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86084.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:37,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4967, 1.5848, 2.1545, 1.8555, 1.7157, 4.0579, 1.4022, 1.7143], + device='cuda:0'), covar=tensor([0.0937, 0.1775, 0.1196, 0.0989, 0.1586, 0.0221, 0.1523, 0.1702], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0078, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:00:50,270 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:51,952 INFO [finetune.py:976] (0/7) Epoch 16, batch 200, loss[loss=0.1932, simple_loss=0.2732, pruned_loss=0.05657, over 4863.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2459, pruned_loss=0.05696, over 607797.34 frames. ], batch size: 34, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:01:04,020 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:05,146 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:09,822 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86139.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:14,014 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.565e+02 1.801e+02 2.285e+02 3.660e+02, threshold=3.601e+02, percent-clipped=0.0 +2023-03-26 19:01:27,190 INFO [finetune.py:976] (0/7) Epoch 16, batch 250, loss[loss=0.2088, simple_loss=0.2712, pruned_loss=0.07322, over 4892.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2483, pruned_loss=0.05732, over 684915.64 frames. ], batch size: 32, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:01:34,323 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7691, 1.7522, 1.8120, 1.1638, 1.8585, 1.8657, 1.8260, 1.4968], + device='cuda:0'), covar=tensor([0.0586, 0.0671, 0.0661, 0.0908, 0.0697, 0.0667, 0.0568, 0.1184], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0122, 0.0123, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:01:40,817 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:41,400 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86187.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:45,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2310, 2.8533, 3.0047, 3.1631, 3.0193, 2.8100, 3.2642, 1.0363], + device='cuda:0'), covar=tensor([0.1043, 0.1001, 0.1003, 0.1135, 0.1501, 0.1734, 0.1168, 0.5380], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0243, 0.0274, 0.0291, 0.0332, 0.0281, 0.0298, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:02:00,655 INFO [finetune.py:976] (0/7) Epoch 16, batch 300, loss[loss=0.1874, simple_loss=0.2634, pruned_loss=0.05574, over 4900.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2543, pruned_loss=0.05934, over 744763.21 frames. ], batch size: 43, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:02:06,173 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1779, 2.0842, 1.7853, 2.0244, 1.9193, 1.9531, 2.0116, 2.7521], + device='cuda:0'), covar=tensor([0.3880, 0.4548, 0.3582, 0.4175, 0.4112, 0.2459, 0.3922, 0.1773], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0259, 0.0225, 0.0273, 0.0246, 0.0213, 0.0247, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:02:11,158 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86231.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:12,978 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:17,976 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 19:02:20,698 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.625e+02 1.967e+02 2.251e+02 5.649e+02, threshold=3.935e+02, percent-clipped=3.0 +2023-03-26 19:02:20,809 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:34,278 INFO [finetune.py:976] (0/7) Epoch 16, batch 350, loss[loss=0.185, simple_loss=0.2542, pruned_loss=0.05797, over 4862.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2565, pruned_loss=0.0598, over 791298.74 frames. ], batch size: 34, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:02:44,487 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:47,956 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86286.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:03:01,143 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:04,817 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-03-26 19:03:05,209 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86313.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:07,485 INFO [finetune.py:976] (0/7) Epoch 16, batch 400, loss[loss=0.1655, simple_loss=0.2389, pruned_loss=0.04605, over 4820.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2564, pruned_loss=0.05919, over 827788.72 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:03:15,903 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:34,376 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.307e+01 1.538e+02 1.774e+02 2.172e+02 4.200e+02, threshold=3.548e+02, percent-clipped=1.0 +2023-03-26 19:03:45,093 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:50,328 INFO [finetune.py:976] (0/7) Epoch 16, batch 450, loss[loss=0.2154, simple_loss=0.2635, pruned_loss=0.08364, over 4875.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2543, pruned_loss=0.05813, over 857198.90 frames. ], batch size: 34, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:04:03,238 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1603, 2.0898, 1.9890, 2.3154, 2.7088, 2.1808, 2.1333, 1.7863], + device='cuda:0'), covar=tensor([0.2049, 0.1913, 0.1758, 0.1408, 0.1627, 0.1062, 0.2033, 0.1810], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0207, 0.0211, 0.0191, 0.0242, 0.0184, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:04:18,722 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:24,032 INFO [finetune.py:976] (0/7) Epoch 16, batch 500, loss[loss=0.1535, simple_loss=0.2298, pruned_loss=0.03858, over 4739.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2518, pruned_loss=0.0576, over 877724.69 frames. ], batch size: 27, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:04:30,006 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:40,612 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:43,619 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86445.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:44,074 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.779e+01 1.556e+02 1.875e+02 2.226e+02 4.465e+02, threshold=3.750e+02, percent-clipped=2.0 +2023-03-26 19:04:57,163 INFO [finetune.py:976] (0/7) Epoch 16, batch 550, loss[loss=0.2074, simple_loss=0.2773, pruned_loss=0.06876, over 4819.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.248, pruned_loss=0.05595, over 894618.94 frames. ], batch size: 45, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:05:31,541 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:05:39,616 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86506.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:05:49,684 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4055, 1.3566, 1.7703, 2.4997, 1.7123, 2.2067, 1.0956, 2.1208], + device='cuda:0'), covar=tensor([0.1669, 0.1362, 0.1027, 0.0723, 0.0874, 0.1119, 0.1345, 0.0623], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0100, 0.0138, 0.0123, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 19:05:49,744 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7872, 0.7276, 1.7364, 1.6843, 1.5827, 1.5192, 1.5574, 1.6871], + device='cuda:0'), covar=tensor([0.3703, 0.3651, 0.3314, 0.3515, 0.4340, 0.3794, 0.4139, 0.3163], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0238, 0.0258, 0.0268, 0.0268, 0.0241, 0.0280, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:05:50,204 INFO [finetune.py:976] (0/7) Epoch 16, batch 600, loss[loss=0.2189, simple_loss=0.2682, pruned_loss=0.08475, over 4833.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2491, pruned_loss=0.05671, over 909035.25 frames. ], batch size: 30, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:06:04,164 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86531.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:13,023 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-03-26 19:06:14,603 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.575e+02 1.922e+02 2.222e+02 3.111e+02, threshold=3.844e+02, percent-clipped=0.0 +2023-03-26 19:06:27,336 INFO [finetune.py:976] (0/7) Epoch 16, batch 650, loss[loss=0.1784, simple_loss=0.2483, pruned_loss=0.05423, over 4856.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2521, pruned_loss=0.05781, over 917607.73 frames. ], batch size: 31, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:06:36,234 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86579.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:37,379 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86580.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:41,052 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86586.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:06:52,264 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:59,446 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:01,141 INFO [finetune.py:976] (0/7) Epoch 16, batch 700, loss[loss=0.1206, simple_loss=0.1926, pruned_loss=0.02427, over 4789.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2542, pruned_loss=0.0581, over 927721.09 frames. ], batch size: 26, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:07:13,496 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86634.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:07:17,779 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:21,644 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.513e+02 1.866e+02 2.326e+02 3.823e+02, threshold=3.732e+02, percent-clipped=0.0 +2023-03-26 19:07:29,573 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:31,332 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86661.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:34,801 INFO [finetune.py:976] (0/7) Epoch 16, batch 750, loss[loss=0.2198, simple_loss=0.2876, pruned_loss=0.07598, over 4802.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2549, pruned_loss=0.05798, over 933737.46 frames. ], batch size: 40, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:02,103 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86706.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:03,364 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:08,617 INFO [finetune.py:976] (0/7) Epoch 16, batch 800, loss[loss=0.147, simple_loss=0.2178, pruned_loss=0.03814, over 4754.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2549, pruned_loss=0.05777, over 937929.85 frames. ], batch size: 26, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:14,183 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86725.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:17,053 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6412, 1.2306, 0.7322, 1.5790, 2.0100, 1.4520, 1.4363, 1.6014], + device='cuda:0'), covar=tensor([0.2074, 0.2924, 0.2652, 0.1789, 0.2437, 0.2620, 0.2100, 0.2826], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 19:08:22,322 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0241, 1.6640, 2.3054, 1.5628, 2.0278, 2.1632, 1.7356, 2.2924], + device='cuda:0'), covar=tensor([0.1394, 0.2166, 0.1578, 0.2110, 0.0959, 0.1540, 0.2581, 0.0924], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0206, 0.0195, 0.0192, 0.0178, 0.0216, 0.0220, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:08:28,782 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.501e+02 1.840e+02 2.208e+02 4.378e+02, threshold=3.681e+02, percent-clipped=4.0 +2023-03-26 19:08:40,243 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:49,459 INFO [finetune.py:976] (0/7) Epoch 16, batch 850, loss[loss=0.2162, simple_loss=0.28, pruned_loss=0.07618, over 4905.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2538, pruned_loss=0.05802, over 941015.90 frames. ], batch size: 36, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:54,206 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:09,590 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86796.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:13,593 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86801.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:09:21,928 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86814.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:23,014 INFO [finetune.py:976] (0/7) Epoch 16, batch 900, loss[loss=0.2034, simple_loss=0.2622, pruned_loss=0.07233, over 3893.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2519, pruned_loss=0.05771, over 941812.05 frames. ], batch size: 17, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:09:40,870 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4453, 1.3254, 1.7176, 2.4709, 1.6917, 2.1344, 0.9788, 2.1101], + device='cuda:0'), covar=tensor([0.1702, 0.1413, 0.1095, 0.0654, 0.0897, 0.1203, 0.1538, 0.0634], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0117, 0.0134, 0.0164, 0.0101, 0.0139, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 19:09:40,898 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5482, 1.4381, 1.4776, 0.9221, 1.5898, 1.5038, 1.5292, 1.3777], + device='cuda:0'), covar=tensor([0.0619, 0.0797, 0.0712, 0.0878, 0.0802, 0.0780, 0.0629, 0.1283], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0135, 0.0141, 0.0123, 0.0123, 0.0140, 0.0141, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:09:43,094 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.502e+02 1.904e+02 2.205e+02 3.944e+02, threshold=3.808e+02, percent-clipped=2.0 +2023-03-26 19:09:56,619 INFO [finetune.py:976] (0/7) Epoch 16, batch 950, loss[loss=0.1472, simple_loss=0.2185, pruned_loss=0.03798, over 4771.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2492, pruned_loss=0.05706, over 946471.44 frames. ], batch size: 28, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:10:02,707 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86875.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:04,478 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:20,381 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:40,049 INFO [finetune.py:976] (0/7) Epoch 16, batch 1000, loss[loss=0.2322, simple_loss=0.2961, pruned_loss=0.08417, over 4909.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2521, pruned_loss=0.05842, over 949813.79 frames. ], batch size: 35, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:11:01,991 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:08,299 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:12,979 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.633e+02 1.920e+02 2.320e+02 4.350e+02, threshold=3.840e+02, percent-clipped=2.0 +2023-03-26 19:11:13,131 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4761, 1.3164, 1.2687, 1.3626, 1.6840, 1.6015, 1.3632, 1.2566], + device='cuda:0'), covar=tensor([0.0312, 0.0369, 0.0599, 0.0342, 0.0244, 0.0469, 0.0388, 0.0442], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0108, 0.0143, 0.0113, 0.0099, 0.0106, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2572e-05, 8.3815e-05, 1.1299e-04, 8.7070e-05, 7.7397e-05, 7.8312e-05, + 7.3141e-05, 8.2343e-05], device='cuda:0') +2023-03-26 19:11:20,313 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:34,824 INFO [finetune.py:976] (0/7) Epoch 16, batch 1050, loss[loss=0.222, simple_loss=0.2774, pruned_loss=0.08328, over 4835.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2539, pruned_loss=0.05854, over 949347.21 frames. ], batch size: 30, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:11:58,519 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87002.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:12:05,579 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.81 vs. limit=2.0 +2023-03-26 19:12:08,331 INFO [finetune.py:976] (0/7) Epoch 16, batch 1100, loss[loss=0.1664, simple_loss=0.2397, pruned_loss=0.04654, over 4818.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2545, pruned_loss=0.05852, over 950540.54 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:12:27,409 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.613e+02 1.835e+02 2.273e+02 4.124e+02, threshold=3.670e+02, percent-clipped=1.0 +2023-03-26 19:12:31,625 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8452, 1.7775, 1.6207, 1.9460, 2.0911, 1.9810, 1.4268, 1.5982], + device='cuda:0'), covar=tensor([0.2170, 0.1964, 0.1971, 0.1674, 0.1648, 0.1165, 0.2417, 0.2002], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0242, 0.0185, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:12:39,651 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87063.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:12:41,762 INFO [finetune.py:976] (0/7) Epoch 16, batch 1150, loss[loss=0.1623, simple_loss=0.2417, pruned_loss=0.04144, over 4858.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2553, pruned_loss=0.05849, over 949912.98 frames. ], batch size: 34, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:00,935 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:04,438 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87101.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:13:15,213 INFO [finetune.py:976] (0/7) Epoch 16, batch 1200, loss[loss=0.2014, simple_loss=0.2586, pruned_loss=0.07205, over 4886.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2526, pruned_loss=0.05786, over 949200.04 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:16,018 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 19:13:31,479 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0765, 1.9932, 1.5705, 1.8319, 1.8621, 1.8568, 1.8836, 2.6410], + device='cuda:0'), covar=tensor([0.4404, 0.4592, 0.3717, 0.4356, 0.4357, 0.2782, 0.4166, 0.2014], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0259, 0.0224, 0.0273, 0.0246, 0.0213, 0.0248, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:13:33,240 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:34,371 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.093e+01 1.544e+02 1.890e+02 2.251e+02 4.242e+02, threshold=3.781e+02, percent-clipped=1.0 +2023-03-26 19:13:36,662 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87149.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:50,131 INFO [finetune.py:976] (0/7) Epoch 16, batch 1250, loss[loss=0.1268, simple_loss=0.1964, pruned_loss=0.02859, over 4701.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2511, pruned_loss=0.05697, over 950980.50 frames. ], batch size: 23, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:53,116 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87170.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:24,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1495, 1.3726, 1.4498, 0.7367, 1.4053, 1.6256, 1.6565, 1.3489], + device='cuda:0'), covar=tensor([0.0887, 0.0545, 0.0484, 0.0493, 0.0425, 0.0577, 0.0301, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0124, 0.0128, 0.0131, 0.0128, 0.0143, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.2325e-05, 1.1052e-04, 8.9083e-05, 9.1680e-05, 9.2526e-05, 9.2351e-05, + 1.0284e-04, 1.0619e-04], device='cuda:0') +2023-03-26 19:14:31,062 INFO [finetune.py:976] (0/7) Epoch 16, batch 1300, loss[loss=0.2022, simple_loss=0.2644, pruned_loss=0.06999, over 4930.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2474, pruned_loss=0.05557, over 952109.75 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:14:41,179 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87230.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:43,570 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:44,799 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:51,273 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.578e+02 1.944e+02 2.250e+02 4.130e+02, threshold=3.887e+02, percent-clipped=1.0 +2023-03-26 19:15:03,934 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7708, 1.6931, 1.4086, 1.8483, 2.4340, 1.8832, 1.4531, 1.4150], + device='cuda:0'), covar=tensor([0.2150, 0.2010, 0.1963, 0.1635, 0.1506, 0.1176, 0.2478, 0.1967], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0209, 0.0211, 0.0192, 0.0242, 0.0185, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:15:04,397 INFO [finetune.py:976] (0/7) Epoch 16, batch 1350, loss[loss=0.1868, simple_loss=0.2625, pruned_loss=0.05559, over 4933.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2479, pruned_loss=0.05597, over 952237.81 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:15:16,950 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:15:18,293 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8114, 1.3642, 1.9142, 1.8140, 1.6238, 1.5838, 1.7638, 1.7732], + device='cuda:0'), covar=tensor([0.4305, 0.4540, 0.3666, 0.3946, 0.5229, 0.4042, 0.4885, 0.3412], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0239, 0.0258, 0.0270, 0.0269, 0.0241, 0.0282, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:15:21,329 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87291.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:15:44,163 INFO [finetune.py:976] (0/7) Epoch 16, batch 1400, loss[loss=0.1987, simple_loss=0.2816, pruned_loss=0.05787, over 4933.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2517, pruned_loss=0.05677, over 953592.67 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:16:19,254 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.619e+02 1.873e+02 2.376e+02 3.982e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 19:16:31,528 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87358.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:16:38,517 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:16:41,379 INFO [finetune.py:976] (0/7) Epoch 16, batch 1450, loss[loss=0.1977, simple_loss=0.2642, pruned_loss=0.06556, over 4854.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2527, pruned_loss=0.05689, over 953498.64 frames. ], batch size: 31, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:16:50,767 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6091, 1.5246, 1.3231, 1.6548, 1.7771, 1.6166, 1.0700, 1.3902], + device='cuda:0'), covar=tensor([0.2057, 0.1996, 0.1857, 0.1535, 0.1535, 0.1241, 0.2545, 0.1831], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0208, 0.0210, 0.0191, 0.0241, 0.0184, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:17:18,468 INFO [finetune.py:976] (0/7) Epoch 16, batch 1500, loss[loss=0.1952, simple_loss=0.2691, pruned_loss=0.06064, over 4817.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2551, pruned_loss=0.05811, over 954850.36 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:17:23,365 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:39,130 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.635e+02 2.033e+02 2.421e+02 4.092e+02, threshold=4.066e+02, percent-clipped=1.0 +2023-03-26 19:17:48,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87461.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:52,196 INFO [finetune.py:976] (0/7) Epoch 16, batch 1550, loss[loss=0.2307, simple_loss=0.301, pruned_loss=0.0802, over 4816.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2552, pruned_loss=0.05811, over 953180.51 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:17:55,189 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:55,906 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 19:18:06,921 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 19:18:18,828 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:18,846 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6992, 1.5769, 1.4255, 1.7365, 2.1480, 1.7254, 1.4740, 1.4156], + device='cuda:0'), covar=tensor([0.1902, 0.1892, 0.1767, 0.1460, 0.1409, 0.1169, 0.2174, 0.1728], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0208, 0.0210, 0.0191, 0.0241, 0.0184, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:18:25,481 INFO [finetune.py:976] (0/7) Epoch 16, batch 1600, loss[loss=0.1547, simple_loss=0.2225, pruned_loss=0.04348, over 4799.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2534, pruned_loss=0.05795, over 953079.85 frames. ], batch size: 29, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:18:27,232 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87518.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:29,729 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87522.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:38,579 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87534.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:42,026 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-03-26 19:18:46,617 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.707e+01 1.405e+02 1.628e+02 1.991e+02 3.372e+02, threshold=3.256e+02, percent-clipped=0.0 +2023-03-26 19:18:57,033 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2495, 2.4416, 2.3238, 1.7820, 2.2278, 2.7382, 2.5395, 2.1468], + device='cuda:0'), covar=tensor([0.0668, 0.0571, 0.0756, 0.0883, 0.1255, 0.0631, 0.0592, 0.1040], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0123, 0.0123, 0.0140, 0.0142, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:18:59,343 INFO [finetune.py:976] (0/7) Epoch 16, batch 1650, loss[loss=0.1826, simple_loss=0.2504, pruned_loss=0.05744, over 4827.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2514, pruned_loss=0.05709, over 954599.20 frames. ], batch size: 41, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:18:59,480 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87566.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:19:10,145 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87582.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:11,440 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:13,088 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:17,442 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:36,466 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:43,485 INFO [finetune.py:976] (0/7) Epoch 16, batch 1700, loss[loss=0.2245, simple_loss=0.2892, pruned_loss=0.07996, over 4830.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2488, pruned_loss=0.05639, over 957443.02 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:03,760 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87645.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:20:04,232 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.606e+02 1.879e+02 2.372e+02 9.403e+02, threshold=3.758e+02, percent-clipped=6.0 +2023-03-26 19:20:06,850 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:20:08,002 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8539, 3.9010, 3.7374, 2.2046, 4.0519, 2.9837, 0.8408, 2.8501], + device='cuda:0'), covar=tensor([0.2603, 0.2087, 0.1615, 0.3003, 0.0937, 0.1009, 0.4533, 0.1438], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0158, 0.0127, 0.0156, 0.0122, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 19:20:11,625 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87658.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:20:16,841 INFO [finetune.py:976] (0/7) Epoch 16, batch 1750, loss[loss=0.1798, simple_loss=0.2547, pruned_loss=0.05245, over 4913.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2507, pruned_loss=0.05723, over 955915.65 frames. ], batch size: 36, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:16,973 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87666.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:20:30,661 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2786, 2.8848, 2.7652, 1.1844, 3.0305, 2.2067, 0.5416, 1.8452], + device='cuda:0'), covar=tensor([0.2459, 0.2537, 0.2002, 0.3753, 0.1350, 0.1167, 0.4402, 0.1806], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0158, 0.0127, 0.0156, 0.0122, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 19:20:36,457 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6804, 3.6604, 3.4651, 1.8027, 3.8276, 2.8910, 0.7302, 2.5998], + device='cuda:0'), covar=tensor([0.2258, 0.2138, 0.1541, 0.3418, 0.0977, 0.0966, 0.4436, 0.1538], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0157, 0.0127, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 19:20:44,149 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87706.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:20:50,643 INFO [finetune.py:976] (0/7) Epoch 16, batch 1800, loss[loss=0.2153, simple_loss=0.2711, pruned_loss=0.07974, over 4894.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2533, pruned_loss=0.05789, over 955826.50 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:51,917 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:21:13,298 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.635e+02 1.907e+02 2.399e+02 5.758e+02, threshold=3.813e+02, percent-clipped=1.0 +2023-03-26 19:21:36,191 INFO [finetune.py:976] (0/7) Epoch 16, batch 1850, loss[loss=0.1658, simple_loss=0.2357, pruned_loss=0.04793, over 4804.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2545, pruned_loss=0.05866, over 954952.17 frames. ], batch size: 45, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:22:22,284 INFO [finetune.py:976] (0/7) Epoch 16, batch 1900, loss[loss=0.1839, simple_loss=0.2605, pruned_loss=0.05365, over 4806.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2549, pruned_loss=0.05877, over 954109.02 frames. ], batch size: 40, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:22:23,001 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87817.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:22:41,851 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.250e+02 1.530e+02 1.792e+02 2.215e+02 4.706e+02, threshold=3.584e+02, percent-clipped=3.0 +2023-03-26 19:22:53,006 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87861.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:22:55,973 INFO [finetune.py:976] (0/7) Epoch 16, batch 1950, loss[loss=0.1411, simple_loss=0.2051, pruned_loss=0.0386, over 4223.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2543, pruned_loss=0.05852, over 952791.43 frames. ], batch size: 18, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:23:09,169 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:29,595 INFO [finetune.py:976] (0/7) Epoch 16, batch 2000, loss[loss=0.1657, simple_loss=0.2378, pruned_loss=0.04685, over 4817.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2513, pruned_loss=0.05743, over 952961.58 frames. ], batch size: 41, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:23:41,096 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:42,391 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8345, 2.5217, 2.0825, 1.0684, 2.3653, 2.1054, 2.0077, 2.2465], + device='cuda:0'), covar=tensor([0.0692, 0.0874, 0.1518, 0.2143, 0.1271, 0.2037, 0.1930, 0.0994], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0196, 0.0201, 0.0184, 0.0212, 0.0207, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:23:45,282 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87940.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:23:48,810 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:49,303 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.562e+02 1.812e+02 2.199e+02 5.123e+02, threshold=3.624e+02, percent-clipped=1.0 +2023-03-26 19:23:59,894 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87961.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:24:02,864 INFO [finetune.py:976] (0/7) Epoch 16, batch 2050, loss[loss=0.1678, simple_loss=0.236, pruned_loss=0.04975, over 4917.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2479, pruned_loss=0.05597, over 952423.13 frames. ], batch size: 37, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:24:25,007 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-88000.pt +2023-03-26 19:24:32,031 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1123, 1.9843, 2.1876, 1.7442, 2.0688, 2.3301, 2.2740, 1.8050], + device='cuda:0'), covar=tensor([0.0497, 0.0542, 0.0564, 0.0741, 0.0918, 0.0481, 0.0441, 0.1006], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0123, 0.0124, 0.0139, 0.0142, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:24:37,647 INFO [finetune.py:976] (0/7) Epoch 16, batch 2100, loss[loss=0.1914, simple_loss=0.256, pruned_loss=0.06339, over 4836.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2481, pruned_loss=0.05665, over 950571.34 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:24:38,971 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:24:46,834 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4117, 1.3547, 1.7631, 1.7185, 1.4797, 3.1567, 1.2063, 1.3694], + device='cuda:0'), covar=tensor([0.1161, 0.2368, 0.1435, 0.1161, 0.1933, 0.0303, 0.2013, 0.2301], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0073, 0.0077, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:24:48,067 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1509, 1.9029, 1.9767, 0.9273, 2.2063, 2.3181, 2.0679, 1.8473], + device='cuda:0'), covar=tensor([0.1011, 0.0731, 0.0450, 0.0768, 0.0484, 0.0708, 0.0496, 0.0644], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0152, 0.0125, 0.0129, 0.0132, 0.0128, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.2898e-05, 1.1056e-04, 8.9561e-05, 9.2050e-05, 9.3606e-05, 9.2169e-05, + 1.0341e-04, 1.0668e-04], device='cuda:0') +2023-03-26 19:24:59,877 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.339e+01 1.671e+02 1.963e+02 2.403e+02 4.597e+02, threshold=3.926e+02, percent-clipped=2.0 +2023-03-26 19:25:13,434 INFO [finetune.py:976] (0/7) Epoch 16, batch 2150, loss[loss=0.2082, simple_loss=0.2854, pruned_loss=0.06555, over 4819.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.251, pruned_loss=0.05733, over 952341.37 frames. ], batch size: 39, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:25:13,501 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:25:33,194 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5095, 1.3710, 1.7729, 1.8164, 1.5743, 3.2410, 1.3179, 1.4633], + device='cuda:0'), covar=tensor([0.1020, 0.1886, 0.1163, 0.0947, 0.1488, 0.0255, 0.1510, 0.1796], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0078, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:25:35,016 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88100.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:25:39,771 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-03-26 19:25:46,508 INFO [finetune.py:976] (0/7) Epoch 16, batch 2200, loss[loss=0.1546, simple_loss=0.2304, pruned_loss=0.03933, over 4814.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2534, pruned_loss=0.05809, over 951069.07 frames. ], batch size: 39, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:25:47,216 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:05,785 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.658e+02 1.940e+02 2.471e+02 6.986e+02, threshold=3.880e+02, percent-clipped=3.0 +2023-03-26 19:26:15,407 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88161.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:26:15,431 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88161.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:18,266 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:19,282 INFO [finetune.py:976] (0/7) Epoch 16, batch 2250, loss[loss=0.2075, simple_loss=0.2838, pruned_loss=0.0656, over 4811.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2553, pruned_loss=0.05943, over 950776.84 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:26:31,467 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6828, 0.7199, 1.6765, 1.6065, 1.4684, 1.4440, 1.5461, 1.6329], + device='cuda:0'), covar=tensor([0.3062, 0.3280, 0.2922, 0.3074, 0.3910, 0.3094, 0.3411, 0.2739], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0236, 0.0256, 0.0268, 0.0268, 0.0241, 0.0279, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:26:56,468 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:05,823 INFO [finetune.py:976] (0/7) Epoch 16, batch 2300, loss[loss=0.2106, simple_loss=0.2632, pruned_loss=0.07898, over 4258.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2548, pruned_loss=0.05905, over 950851.60 frames. ], batch size: 18, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:27:22,426 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2199, 3.6436, 3.8305, 4.0765, 3.9710, 3.7026, 4.3047, 1.3244], + device='cuda:0'), covar=tensor([0.0760, 0.0856, 0.0901, 0.1012, 0.1225, 0.1612, 0.0684, 0.5681], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0244, 0.0276, 0.0290, 0.0335, 0.0282, 0.0297, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:27:29,759 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:33,302 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:34,392 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.462e+02 1.840e+02 2.103e+02 3.666e+02, threshold=3.679e+02, percent-clipped=0.0 +2023-03-26 19:27:43,884 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88261.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:27:47,285 INFO [finetune.py:976] (0/7) Epoch 16, batch 2350, loss[loss=0.1418, simple_loss=0.2153, pruned_loss=0.03412, over 4764.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.252, pruned_loss=0.05792, over 953290.43 frames. ], batch size: 28, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:28:01,660 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88288.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:03,546 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88291.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:28:04,696 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:15,424 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88309.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:20,094 INFO [finetune.py:976] (0/7) Epoch 16, batch 2400, loss[loss=0.1787, simple_loss=0.2532, pruned_loss=0.05207, over 4762.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2502, pruned_loss=0.05736, over 954914.34 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:28:40,399 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.525e+02 1.766e+02 2.106e+02 3.774e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 19:28:44,074 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88352.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:28:52,862 INFO [finetune.py:976] (0/7) Epoch 16, batch 2450, loss[loss=0.1824, simple_loss=0.2357, pruned_loss=0.06454, over 4767.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2471, pruned_loss=0.05578, over 955508.05 frames. ], batch size: 28, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:29:26,833 INFO [finetune.py:976] (0/7) Epoch 16, batch 2500, loss[loss=0.1879, simple_loss=0.2648, pruned_loss=0.05549, over 4908.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2499, pruned_loss=0.05713, over 955227.36 frames. ], batch size: 43, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:29:48,234 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.743e+02 2.000e+02 2.601e+02 5.270e+02, threshold=4.000e+02, percent-clipped=5.0 +2023-03-26 19:29:54,228 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:30:00,708 INFO [finetune.py:976] (0/7) Epoch 16, batch 2550, loss[loss=0.1121, simple_loss=0.177, pruned_loss=0.02357, over 4353.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2535, pruned_loss=0.05841, over 954544.19 frames. ], batch size: 19, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:30:17,732 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1397, 2.8954, 2.6089, 1.4898, 2.7748, 2.1878, 2.3366, 2.5362], + device='cuda:0'), covar=tensor([0.1023, 0.0722, 0.1636, 0.1957, 0.1466, 0.2163, 0.1844, 0.1175], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0196, 0.0199, 0.0183, 0.0212, 0.0207, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:30:33,907 INFO [finetune.py:976] (0/7) Epoch 16, batch 2600, loss[loss=0.1692, simple_loss=0.2388, pruned_loss=0.04978, over 4033.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2537, pruned_loss=0.05757, over 954095.58 frames. ], batch size: 17, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:30:34,165 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 19:30:55,463 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.436e+01 1.666e+02 1.940e+02 2.279e+02 3.712e+02, threshold=3.880e+02, percent-clipped=0.0 +2023-03-26 19:30:59,277 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1992, 2.1082, 1.6668, 2.0009, 2.1206, 1.8287, 2.3666, 2.2122], + device='cuda:0'), covar=tensor([0.1357, 0.2072, 0.3011, 0.2805, 0.2561, 0.1644, 0.3008, 0.1683], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0187, 0.0234, 0.0252, 0.0244, 0.0201, 0.0212, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:31:07,484 INFO [finetune.py:976] (0/7) Epoch 16, batch 2650, loss[loss=0.1697, simple_loss=0.2405, pruned_loss=0.04941, over 4809.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2559, pruned_loss=0.05833, over 952764.99 frames. ], batch size: 25, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:31:12,336 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0095, 3.5013, 3.6750, 3.8912, 3.7784, 3.5099, 4.0872, 1.2591], + device='cuda:0'), covar=tensor([0.0703, 0.0820, 0.0837, 0.0829, 0.1142, 0.1528, 0.0703, 0.5423], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0245, 0.0277, 0.0292, 0.0335, 0.0283, 0.0298, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:31:41,336 INFO [finetune.py:976] (0/7) Epoch 16, batch 2700, loss[loss=0.1396, simple_loss=0.2101, pruned_loss=0.03457, over 4925.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2548, pruned_loss=0.05787, over 952853.93 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:32:05,538 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.504e+02 1.815e+02 2.296e+02 4.078e+02, threshold=3.631e+02, percent-clipped=1.0 +2023-03-26 19:32:05,615 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88647.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:32:26,940 INFO [finetune.py:976] (0/7) Epoch 16, batch 2750, loss[loss=0.1676, simple_loss=0.2297, pruned_loss=0.05274, over 4821.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2521, pruned_loss=0.05698, over 954838.63 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:32:42,723 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-03-26 19:32:46,904 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 19:33:17,054 INFO [finetune.py:976] (0/7) Epoch 16, batch 2800, loss[loss=0.1682, simple_loss=0.2364, pruned_loss=0.04997, over 4898.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2491, pruned_loss=0.05585, over 956110.78 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:33:37,857 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.516e+02 1.815e+02 2.286e+02 3.246e+02, threshold=3.631e+02, percent-clipped=0.0 +2023-03-26 19:33:43,837 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 19:33:44,962 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:33:50,910 INFO [finetune.py:976] (0/7) Epoch 16, batch 2850, loss[loss=0.2404, simple_loss=0.3018, pruned_loss=0.08948, over 4268.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.249, pruned_loss=0.05675, over 956737.46 frames. ], batch size: 65, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:33:58,291 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-03-26 19:34:17,620 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:34:24,821 INFO [finetune.py:976] (0/7) Epoch 16, batch 2900, loss[loss=0.2165, simple_loss=0.2927, pruned_loss=0.07011, over 4863.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2512, pruned_loss=0.05729, over 956152.70 frames. ], batch size: 44, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:34:45,211 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.632e+02 1.968e+02 2.500e+02 4.348e+02, threshold=3.936e+02, percent-clipped=6.0 +2023-03-26 19:34:45,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9942, 2.2497, 1.8247, 1.9309, 2.5514, 2.4646, 2.1227, 2.0917], + device='cuda:0'), covar=tensor([0.0487, 0.0327, 0.0614, 0.0340, 0.0274, 0.0744, 0.0323, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0107, 0.0142, 0.0111, 0.0098, 0.0106, 0.0097, 0.0107], + device='cuda:0'), out_proj_covar=tensor([7.2315e-05, 8.3011e-05, 1.1238e-04, 8.5885e-05, 7.6785e-05, 7.8081e-05, + 7.2487e-05, 8.1845e-05], device='cuda:0') +2023-03-26 19:34:58,812 INFO [finetune.py:976] (0/7) Epoch 16, batch 2950, loss[loss=0.1904, simple_loss=0.2666, pruned_loss=0.05711, over 4890.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2545, pruned_loss=0.0585, over 955884.58 frames. ], batch size: 35, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:35:04,888 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5781, 1.4832, 2.2219, 1.9326, 1.7179, 4.0934, 1.4401, 1.6448], + device='cuda:0'), covar=tensor([0.1002, 0.1852, 0.1206, 0.1000, 0.1632, 0.0216, 0.1535, 0.1844], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:35:32,639 INFO [finetune.py:976] (0/7) Epoch 16, batch 3000, loss[loss=0.2041, simple_loss=0.2796, pruned_loss=0.06427, over 4803.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2555, pruned_loss=0.0585, over 957254.49 frames. ], batch size: 45, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:35:32,640 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 19:35:39,020 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1987, 1.2633, 1.1547, 1.3361, 1.4802, 1.4099, 1.2502, 1.1860], + device='cuda:0'), covar=tensor([0.0458, 0.0362, 0.0577, 0.0285, 0.0238, 0.0460, 0.0358, 0.0351], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2988e-05, 8.3872e-05, 1.1301e-04, 8.6599e-05, 7.7392e-05, 7.8828e-05, + 7.3023e-05, 8.2483e-05], device='cuda:0') +2023-03-26 19:35:42,032 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0736, 1.9823, 1.7655, 1.9318, 2.1186, 1.8658, 2.3431, 2.0701], + device='cuda:0'), covar=tensor([0.1459, 0.2441, 0.3100, 0.2442, 0.2583, 0.1706, 0.3215, 0.1991], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0186, 0.0233, 0.0251, 0.0242, 0.0200, 0.0210, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:35:49,228 INFO [finetune.py:1010] (0/7) Epoch 16, validation: loss=0.1563, simple_loss=0.2263, pruned_loss=0.04316, over 2265189.00 frames. +2023-03-26 19:35:49,228 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 19:35:58,822 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1008, 2.0524, 1.6846, 2.0029, 1.9001, 1.8579, 1.9318, 2.5998], + device='cuda:0'), covar=tensor([0.3742, 0.4227, 0.3267, 0.3934, 0.4053, 0.2279, 0.4102, 0.1709], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0227, 0.0276, 0.0249, 0.0217, 0.0251, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:36:10,232 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0374, 1.8213, 1.6137, 1.6500, 1.7305, 1.7526, 1.7452, 2.4260], + device='cuda:0'), covar=tensor([0.3770, 0.4137, 0.3239, 0.3740, 0.4252, 0.2297, 0.3804, 0.1703], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0227, 0.0276, 0.0249, 0.0217, 0.0251, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:36:10,682 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.661e+02 1.990e+02 2.439e+02 3.546e+02, threshold=3.980e+02, percent-clipped=0.0 +2023-03-26 19:36:10,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88947.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:36:23,200 INFO [finetune.py:976] (0/7) Epoch 16, batch 3050, loss[loss=0.2626, simple_loss=0.3095, pruned_loss=0.1079, over 4729.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2557, pruned_loss=0.05845, over 955799.59 frames. ], batch size: 54, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:36:43,518 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88995.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:36:44,768 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0218, 1.8744, 1.5379, 1.4319, 2.3080, 2.4947, 2.0977, 1.8995], + device='cuda:0'), covar=tensor([0.0353, 0.0407, 0.0756, 0.0456, 0.0271, 0.0486, 0.0326, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.3022e-05, 8.3670e-05, 1.1323e-04, 8.6609e-05, 7.7337e-05, 7.8762e-05, + 7.2858e-05, 8.2250e-05], device='cuda:0') +2023-03-26 19:36:57,478 INFO [finetune.py:976] (0/7) Epoch 16, batch 3100, loss[loss=0.1897, simple_loss=0.2549, pruned_loss=0.06222, over 4815.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2539, pruned_loss=0.05726, over 957336.77 frames. ], batch size: 30, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:37:06,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:37:07,162 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 19:37:20,958 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.966e+01 1.506e+02 1.838e+02 2.198e+02 3.411e+02, threshold=3.676e+02, percent-clipped=0.0 +2023-03-26 19:37:33,682 INFO [finetune.py:976] (0/7) Epoch 16, batch 3150, loss[loss=0.2227, simple_loss=0.2804, pruned_loss=0.08248, over 4793.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2525, pruned_loss=0.0572, over 957018.30 frames. ], batch size: 29, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:37:56,586 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:38:25,683 INFO [finetune.py:976] (0/7) Epoch 16, batch 3200, loss[loss=0.2218, simple_loss=0.2795, pruned_loss=0.08208, over 4348.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2481, pruned_loss=0.05571, over 955459.13 frames. ], batch size: 65, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:38:34,383 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 19:38:50,101 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.609e+02 1.908e+02 2.339e+02 4.086e+02, threshold=3.816e+02, percent-clipped=1.0 +2023-03-26 19:38:50,251 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6899, 1.5235, 1.0968, 0.3406, 1.2928, 1.4584, 1.4253, 1.4881], + device='cuda:0'), covar=tensor([0.0825, 0.0741, 0.1271, 0.1821, 0.1300, 0.2236, 0.2129, 0.0819], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0194, 0.0198, 0.0181, 0.0210, 0.0204, 0.0221, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:38:50,318 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 19:38:53,275 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6383, 1.3941, 1.4846, 0.8646, 1.5539, 1.6287, 1.7748, 1.3208], + device='cuda:0'), covar=tensor([0.0904, 0.0640, 0.0511, 0.0565, 0.0430, 0.0655, 0.0305, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0126, 0.0130, 0.0127, 0.0142, 0.0146], + device='cuda:0'), out_proj_covar=tensor([9.1430e-05, 1.0925e-04, 8.7801e-05, 8.9911e-05, 9.2157e-05, 9.1760e-05, + 1.0192e-04, 1.0514e-04], device='cuda:0') +2023-03-26 19:38:59,803 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0538, 1.9180, 1.7495, 2.0595, 2.6772, 2.0008, 2.0581, 1.5668], + device='cuda:0'), covar=tensor([0.1990, 0.2019, 0.1772, 0.1531, 0.1695, 0.1143, 0.2008, 0.1720], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0193, 0.0245, 0.0186, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:39:02,083 INFO [finetune.py:976] (0/7) Epoch 16, batch 3250, loss[loss=0.2312, simple_loss=0.2874, pruned_loss=0.08749, over 4916.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2503, pruned_loss=0.05702, over 954930.67 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:39:04,010 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:39:21,682 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4123, 1.0154, 0.7579, 1.3381, 1.8698, 0.7569, 1.2511, 1.3413], + device='cuda:0'), covar=tensor([0.1583, 0.2240, 0.1830, 0.1226, 0.1963, 0.2082, 0.1466, 0.1950], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0095, 0.0111, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 19:39:28,299 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6437, 1.3669, 1.7050, 1.7394, 1.5413, 3.1709, 1.2590, 1.4964], + device='cuda:0'), covar=tensor([0.0975, 0.2038, 0.1580, 0.1078, 0.1770, 0.0323, 0.1764, 0.1994], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:39:30,723 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4743, 1.3949, 1.2645, 1.4566, 1.7799, 1.6185, 1.4406, 1.2834], + device='cuda:0'), covar=tensor([0.0305, 0.0310, 0.0635, 0.0302, 0.0174, 0.0434, 0.0362, 0.0414], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0111, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.3222e-05, 8.3484e-05, 1.1326e-04, 8.6185e-05, 7.7197e-05, 7.8920e-05, + 7.3133e-05, 8.2171e-05], device='cuda:0') +2023-03-26 19:39:35,940 INFO [finetune.py:976] (0/7) Epoch 16, batch 3300, loss[loss=0.2439, simple_loss=0.2982, pruned_loss=0.09476, over 4175.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2551, pruned_loss=0.05821, over 955115.70 frames. ], batch size: 65, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:39:37,863 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0191, 3.5020, 3.7272, 3.8137, 3.8021, 3.5878, 4.0553, 1.7132], + device='cuda:0'), covar=tensor([0.0699, 0.0814, 0.0720, 0.0957, 0.0975, 0.1317, 0.0659, 0.4753], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0247, 0.0277, 0.0294, 0.0335, 0.0283, 0.0299, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:39:45,098 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89230.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:39:56,763 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.235e+02 1.765e+02 2.004e+02 2.308e+02 3.942e+02, threshold=4.007e+02, percent-clipped=1.0 +2023-03-26 19:40:09,186 INFO [finetune.py:976] (0/7) Epoch 16, batch 3350, loss[loss=0.2083, simple_loss=0.2838, pruned_loss=0.0664, over 4911.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2568, pruned_loss=0.05931, over 951824.50 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:40:26,859 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0686, 2.7068, 2.5623, 1.2756, 2.8264, 2.1532, 0.8708, 1.8634], + device='cuda:0'), covar=tensor([0.3187, 0.1942, 0.1772, 0.2967, 0.1408, 0.1013, 0.3501, 0.1475], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0176, 0.0160, 0.0129, 0.0158, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 19:40:42,689 INFO [finetune.py:976] (0/7) Epoch 16, batch 3400, loss[loss=0.1658, simple_loss=0.2366, pruned_loss=0.0475, over 4779.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2577, pruned_loss=0.06002, over 948122.80 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:40:56,400 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4210, 1.3437, 1.3248, 1.3098, 0.8868, 2.2455, 0.7014, 1.1137], + device='cuda:0'), covar=tensor([0.3507, 0.2656, 0.2294, 0.2646, 0.1969, 0.0404, 0.2653, 0.1373], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:41:12,549 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.581e+02 1.832e+02 2.219e+02 5.301e+02, threshold=3.664e+02, percent-clipped=1.0 +2023-03-26 19:41:24,424 INFO [finetune.py:976] (0/7) Epoch 16, batch 3450, loss[loss=0.1707, simple_loss=0.2483, pruned_loss=0.04651, over 4919.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2571, pruned_loss=0.05927, over 949843.08 frames. ], batch size: 38, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:41:29,836 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:41:35,744 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89383.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:41:48,192 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1833, 2.8682, 2.7461, 1.3089, 2.9515, 2.1425, 0.7016, 1.9014], + device='cuda:0'), covar=tensor([0.3006, 0.2418, 0.1988, 0.3500, 0.1462, 0.1165, 0.4268, 0.1792], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0175, 0.0159, 0.0128, 0.0158, 0.0123, 0.0147, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 19:41:58,330 INFO [finetune.py:976] (0/7) Epoch 16, batch 3500, loss[loss=0.1983, simple_loss=0.2671, pruned_loss=0.06473, over 4928.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2541, pruned_loss=0.05844, over 950957.62 frames. ], batch size: 38, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:42:04,408 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:10,962 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:13,374 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89439.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:42:18,647 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.517e+02 1.946e+02 2.225e+02 4.216e+02, threshold=3.891e+02, percent-clipped=3.0 +2023-03-26 19:42:31,135 INFO [finetune.py:976] (0/7) Epoch 16, batch 3550, loss[loss=0.1645, simple_loss=0.236, pruned_loss=0.04651, over 4794.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2511, pruned_loss=0.05733, over 953528.61 frames. ], batch size: 29, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:42:40,909 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4405, 1.5395, 1.5998, 0.8760, 1.6410, 1.8504, 1.9236, 1.4028], + device='cuda:0'), covar=tensor([0.0909, 0.0614, 0.0451, 0.0532, 0.0466, 0.0538, 0.0286, 0.0667], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0153, 0.0125, 0.0128, 0.0132, 0.0130, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.3115e-05, 1.1122e-04, 8.9509e-05, 9.1377e-05, 9.3680e-05, 9.3461e-05, + 1.0323e-04, 1.0698e-04], device='cuda:0') +2023-03-26 19:42:43,945 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:53,369 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89500.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:42:59,262 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89509.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:43:06,067 INFO [finetune.py:976] (0/7) Epoch 16, batch 3600, loss[loss=0.1977, simple_loss=0.2651, pruned_loss=0.06511, over 4932.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2495, pruned_loss=0.05667, over 955071.89 frames. ], batch size: 42, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:43:14,208 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:43:37,730 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.924e+01 1.526e+02 1.890e+02 2.215e+02 3.895e+02, threshold=3.780e+02, percent-clipped=1.0 +2023-03-26 19:44:03,076 INFO [finetune.py:976] (0/7) Epoch 16, batch 3650, loss[loss=0.1993, simple_loss=0.2682, pruned_loss=0.06521, over 4830.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2526, pruned_loss=0.05876, over 953583.73 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:44:06,131 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89570.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:44:36,728 INFO [finetune.py:976] (0/7) Epoch 16, batch 3700, loss[loss=0.1715, simple_loss=0.2428, pruned_loss=0.0501, over 4933.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2558, pruned_loss=0.05901, over 954842.87 frames. ], batch size: 33, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:44:57,081 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.548e+01 1.593e+02 1.994e+02 2.376e+02 3.738e+02, threshold=3.989e+02, percent-clipped=0.0 +2023-03-26 19:45:10,209 INFO [finetune.py:976] (0/7) Epoch 16, batch 3750, loss[loss=0.1698, simple_loss=0.232, pruned_loss=0.05381, over 4731.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2559, pruned_loss=0.05905, over 952915.50 frames. ], batch size: 23, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:45:19,311 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:21,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89683.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:43,305 INFO [finetune.py:976] (0/7) Epoch 16, batch 3800, loss[loss=0.1968, simple_loss=0.2732, pruned_loss=0.06024, over 4832.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2561, pruned_loss=0.05897, over 952457.91 frames. ], batch size: 49, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:45:49,898 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7550, 1.6583, 1.6121, 1.6790, 1.1126, 3.5599, 1.4095, 1.9352], + device='cuda:0'), covar=tensor([0.3230, 0.2469, 0.2067, 0.2392, 0.1858, 0.0216, 0.2471, 0.1183], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0124, 0.0114, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:45:52,762 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:53,367 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89731.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:00,037 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:03,526 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.549e+02 1.882e+02 2.353e+02 4.344e+02, threshold=3.764e+02, percent-clipped=2.0 +2023-03-26 19:46:19,024 INFO [finetune.py:976] (0/7) Epoch 16, batch 3850, loss[loss=0.1891, simple_loss=0.2475, pruned_loss=0.06541, over 4131.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2546, pruned_loss=0.05841, over 952602.96 frames. ], batch size: 18, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:46:19,834 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 19:46:29,104 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:32,275 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 19:46:35,279 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2492, 2.1505, 1.8637, 2.0615, 2.0133, 1.9514, 2.0415, 2.7562], + device='cuda:0'), covar=tensor([0.3742, 0.4627, 0.3331, 0.3904, 0.3985, 0.2730, 0.3859, 0.1650], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0262, 0.0226, 0.0276, 0.0249, 0.0217, 0.0251, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:46:37,052 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 19:46:38,079 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89795.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:46:38,754 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2620, 2.2742, 1.8736, 2.0806, 2.2252, 1.9365, 2.5432, 2.2927], + device='cuda:0'), covar=tensor([0.1343, 0.1961, 0.2731, 0.2684, 0.2345, 0.1607, 0.2847, 0.1707], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0188, 0.0234, 0.0253, 0.0245, 0.0202, 0.0212, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:46:52,159 INFO [finetune.py:976] (0/7) Epoch 16, batch 3900, loss[loss=0.1926, simple_loss=0.2555, pruned_loss=0.06479, over 4925.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2529, pruned_loss=0.05873, over 953432.72 frames. ], batch size: 38, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:46:58,253 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89825.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:12,472 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.500e+02 1.857e+02 2.274e+02 5.172e+02, threshold=3.715e+02, percent-clipped=1.0 +2023-03-26 19:47:23,889 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:23,938 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:24,440 INFO [finetune.py:976] (0/7) Epoch 16, batch 3950, loss[loss=0.14, simple_loss=0.2142, pruned_loss=0.03296, over 4819.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2492, pruned_loss=0.05745, over 953781.90 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:47:29,791 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89873.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:30,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8294, 1.7478, 1.5805, 1.9544, 2.2328, 2.0233, 1.4828, 1.5023], + device='cuda:0'), covar=tensor([0.2288, 0.2041, 0.2054, 0.1682, 0.1726, 0.1104, 0.2611, 0.2083], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0209, 0.0211, 0.0192, 0.0243, 0.0185, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:47:53,552 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9739, 1.8186, 1.8969, 1.2773, 1.9095, 2.0190, 2.0579, 1.6045], + device='cuda:0'), covar=tensor([0.0636, 0.0732, 0.0755, 0.0929, 0.0713, 0.0696, 0.0569, 0.1160], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0137, 0.0144, 0.0126, 0.0126, 0.0143, 0.0144, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:47:57,667 INFO [finetune.py:976] (0/7) Epoch 16, batch 4000, loss[loss=0.1717, simple_loss=0.2228, pruned_loss=0.06028, over 4034.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2488, pruned_loss=0.05719, over 953112.20 frames. ], batch size: 17, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:48:04,743 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89926.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:48:18,302 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.676e+02 1.974e+02 2.469e+02 4.779e+02, threshold=3.947e+02, percent-clipped=6.0 +2023-03-26 19:48:18,487 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.31 vs. limit=5.0 +2023-03-26 19:48:32,941 INFO [finetune.py:976] (0/7) Epoch 16, batch 4050, loss[loss=0.1416, simple_loss=0.2002, pruned_loss=0.04152, over 4147.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2525, pruned_loss=0.05834, over 952910.83 frames. ], batch size: 17, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:49:10,352 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-90000.pt +2023-03-26 19:49:29,801 INFO [finetune.py:976] (0/7) Epoch 16, batch 4100, loss[loss=0.1897, simple_loss=0.258, pruned_loss=0.06076, over 4824.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.256, pruned_loss=0.05955, over 952049.94 frames. ], batch size: 33, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:49:29,915 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0518, 1.8256, 1.5347, 1.4059, 2.4638, 2.5935, 2.0443, 1.9882], + device='cuda:0'), covar=tensor([0.0375, 0.0434, 0.0833, 0.0501, 0.0261, 0.0493, 0.0453, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0109, 0.0145, 0.0114, 0.0101, 0.0109, 0.0099, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4050e-05, 8.4592e-05, 1.1492e-04, 8.7891e-05, 7.8555e-05, 8.0458e-05, + 7.4662e-05, 8.3595e-05], device='cuda:0') +2023-03-26 19:49:43,233 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:49:46,943 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:49:54,440 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.559e+02 1.839e+02 2.160e+02 6.359e+02, threshold=3.678e+02, percent-clipped=1.0 +2023-03-26 19:50:06,383 INFO [finetune.py:976] (0/7) Epoch 16, batch 4150, loss[loss=0.2195, simple_loss=0.2795, pruned_loss=0.07972, over 4729.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2574, pruned_loss=0.05983, over 952120.95 frames. ], batch size: 59, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:50:10,040 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5872, 1.5278, 1.3817, 1.5448, 1.9974, 1.7958, 1.6409, 1.4612], + device='cuda:0'), covar=tensor([0.0311, 0.0280, 0.0595, 0.0302, 0.0165, 0.0444, 0.0297, 0.0352], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0109, 0.0145, 0.0114, 0.0101, 0.0109, 0.0099, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4090e-05, 8.4514e-05, 1.1476e-04, 8.7787e-05, 7.8375e-05, 8.0383e-05, + 7.4583e-05, 8.3551e-05], device='cuda:0') +2023-03-26 19:50:14,191 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:16,517 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:26,423 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90095.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:50:28,342 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-26 19:50:32,506 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6218, 1.5191, 1.4668, 1.5320, 1.2160, 3.5892, 1.3881, 1.8584], + device='cuda:0'), covar=tensor([0.3293, 0.2560, 0.2216, 0.2428, 0.1797, 0.0177, 0.2621, 0.1304], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0097, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:50:39,513 INFO [finetune.py:976] (0/7) Epoch 16, batch 4200, loss[loss=0.2036, simple_loss=0.2671, pruned_loss=0.07004, over 4905.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2573, pruned_loss=0.05948, over 951357.98 frames. ], batch size: 36, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:50:45,529 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7761, 1.5824, 2.1948, 1.4245, 1.9667, 2.0486, 1.5402, 2.2431], + device='cuda:0'), covar=tensor([0.1358, 0.2207, 0.1197, 0.1799, 0.0945, 0.1428, 0.3048, 0.0852], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0192, 0.0191, 0.0176, 0.0214, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:50:47,968 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:48,627 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:57,831 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90143.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:51:00,648 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.547e+02 1.785e+02 2.134e+02 3.751e+02, threshold=3.570e+02, percent-clipped=1.0 +2023-03-26 19:51:12,341 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:12,839 INFO [finetune.py:976] (0/7) Epoch 16, batch 4250, loss[loss=0.1711, simple_loss=0.2283, pruned_loss=0.05699, over 4844.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2549, pruned_loss=0.05892, over 953451.65 frames. ], batch size: 49, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:51:29,551 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:43,662 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90213.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:45,896 INFO [finetune.py:976] (0/7) Epoch 16, batch 4300, loss[loss=0.1386, simple_loss=0.2007, pruned_loss=0.03826, over 4810.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.253, pruned_loss=0.0582, over 955339.88 frames. ], batch size: 25, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:51:46,069 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-03-26 19:51:48,951 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90221.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:51:56,160 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:52:07,167 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.551e+02 1.797e+02 2.190e+02 3.764e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-03-26 19:52:18,553 INFO [finetune.py:976] (0/7) Epoch 16, batch 4350, loss[loss=0.1799, simple_loss=0.2443, pruned_loss=0.05773, over 4846.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2502, pruned_loss=0.05712, over 957597.41 frames. ], batch size: 49, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:52:31,200 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5800, 1.4828, 1.3486, 1.6416, 1.6727, 1.6553, 0.9556, 1.3668], + device='cuda:0'), covar=tensor([0.2151, 0.1992, 0.1965, 0.1632, 0.1532, 0.1214, 0.2751, 0.1896], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0242, 0.0184, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:52:36,479 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:52:51,897 INFO [finetune.py:976] (0/7) Epoch 16, batch 4400, loss[loss=0.1967, simple_loss=0.277, pruned_loss=0.05817, over 4900.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2514, pruned_loss=0.05799, over 956610.96 frames. ], batch size: 43, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:53:04,943 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:53:13,597 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.586e+02 1.845e+02 2.241e+02 4.760e+02, threshold=3.689e+02, percent-clipped=4.0 +2023-03-26 19:53:15,485 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90351.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:53:25,409 INFO [finetune.py:976] (0/7) Epoch 16, batch 4450, loss[loss=0.1829, simple_loss=0.25, pruned_loss=0.05788, over 4755.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2541, pruned_loss=0.0582, over 956100.15 frames. ], batch size: 28, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:53:37,365 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:54:05,547 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90412.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:54:07,882 INFO [finetune.py:976] (0/7) Epoch 16, batch 4500, loss[loss=0.2208, simple_loss=0.2907, pruned_loss=0.07544, over 4913.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2557, pruned_loss=0.05866, over 955274.67 frames. ], batch size: 37, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:54:40,948 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.659e+02 2.056e+02 2.631e+02 3.688e+02, threshold=4.111e+02, percent-clipped=0.0 +2023-03-26 19:54:53,411 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:01,920 INFO [finetune.py:976] (0/7) Epoch 16, batch 4550, loss[loss=0.1905, simple_loss=0.2545, pruned_loss=0.06329, over 4688.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2568, pruned_loss=0.0589, over 954543.18 frames. ], batch size: 59, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:55:18,110 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:19,481 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 19:55:24,683 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90496.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:34,563 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9537, 1.6366, 2.0325, 1.3870, 1.8934, 2.1115, 1.5142, 2.2370], + device='cuda:0'), covar=tensor([0.1242, 0.2054, 0.1443, 0.2038, 0.0947, 0.1389, 0.2951, 0.0906], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0205, 0.0192, 0.0191, 0.0177, 0.0213, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:55:38,629 INFO [finetune.py:976] (0/7) Epoch 16, batch 4600, loss[loss=0.196, simple_loss=0.2698, pruned_loss=0.06111, over 4869.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2549, pruned_loss=0.05766, over 953048.89 frames. ], batch size: 34, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:55:41,652 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:42,205 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90521.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:55:55,837 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0935, 1.9515, 1.6822, 1.8184, 1.8298, 1.7653, 1.8733, 2.6340], + device='cuda:0'), covar=tensor([0.3795, 0.4316, 0.3331, 0.3969, 0.4070, 0.2441, 0.3724, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0261, 0.0226, 0.0276, 0.0249, 0.0216, 0.0250, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:55:59,294 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.556e+01 1.469e+02 1.715e+02 2.012e+02 4.010e+02, threshold=3.429e+02, percent-clipped=0.0 +2023-03-26 19:56:06,273 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90557.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:09,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8498, 1.3343, 0.8831, 1.7161, 2.2834, 1.4586, 1.5851, 1.7037], + device='cuda:0'), covar=tensor([0.1412, 0.2127, 0.2148, 0.1175, 0.1759, 0.1964, 0.1441, 0.1846], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0095, 0.0111, 0.0092, 0.0119, 0.0095, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 19:56:11,543 INFO [finetune.py:976] (0/7) Epoch 16, batch 4650, loss[loss=0.1847, simple_loss=0.2506, pruned_loss=0.05938, over 4920.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2518, pruned_loss=0.05697, over 953671.69 frames. ], batch size: 37, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:56:13,934 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:13,981 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1304, 1.3756, 1.3827, 1.3235, 1.6042, 2.5178, 1.2679, 1.5574], + device='cuda:0'), covar=tensor([0.1053, 0.1803, 0.0971, 0.0918, 0.1489, 0.0399, 0.1562, 0.1658], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:56:15,764 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 19:56:16,343 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-26 19:56:21,046 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0160, 1.5367, 2.0827, 1.8976, 1.6843, 1.6983, 1.8869, 1.9319], + device='cuda:0'), covar=tensor([0.3246, 0.3434, 0.2902, 0.3653, 0.4573, 0.3902, 0.4289, 0.2744], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0240, 0.0260, 0.0273, 0.0272, 0.0246, 0.0283, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:56:26,167 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:30,980 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0339, 0.9100, 0.9063, 1.1585, 1.2099, 1.1283, 1.0173, 0.9603], + device='cuda:0'), covar=tensor([0.0353, 0.0294, 0.0650, 0.0277, 0.0238, 0.0445, 0.0327, 0.0393], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0112, 0.0100, 0.0107, 0.0098, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.3460e-05, 8.3558e-05, 1.1351e-04, 8.6812e-05, 7.7535e-05, 7.9331e-05, + 7.3629e-05, 8.2753e-05], device='cuda:0') +2023-03-26 19:56:45,054 INFO [finetune.py:976] (0/7) Epoch 16, batch 4700, loss[loss=0.139, simple_loss=0.2077, pruned_loss=0.03518, over 4920.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2489, pruned_loss=0.05592, over 955108.65 frames. ], batch size: 43, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:56:59,832 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 19:57:04,045 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 19:57:05,662 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.593e+02 1.858e+02 2.163e+02 3.767e+02, threshold=3.717e+02, percent-clipped=1.0 +2023-03-26 19:57:14,608 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6879, 1.6374, 2.2294, 3.3181, 2.2617, 2.3196, 1.0156, 2.7268], + device='cuda:0'), covar=tensor([0.1545, 0.1393, 0.1205, 0.0568, 0.0777, 0.1611, 0.1787, 0.0460], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 19:57:18,492 INFO [finetune.py:976] (0/7) Epoch 16, batch 4750, loss[loss=0.1956, simple_loss=0.2682, pruned_loss=0.06152, over 4915.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2483, pruned_loss=0.05583, over 957024.22 frames. ], batch size: 46, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:57:37,726 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2942, 2.1157, 2.7641, 1.4869, 2.2798, 2.5737, 1.9270, 2.6774], + device='cuda:0'), covar=tensor([0.1505, 0.1886, 0.1610, 0.2548, 0.1014, 0.1536, 0.2778, 0.0983], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0206, 0.0193, 0.0192, 0.0177, 0.0214, 0.0220, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:57:45,611 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90707.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:57:52,458 INFO [finetune.py:976] (0/7) Epoch 16, batch 4800, loss[loss=0.262, simple_loss=0.321, pruned_loss=0.1015, over 4831.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.251, pruned_loss=0.05704, over 956172.90 frames. ], batch size: 47, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:57:53,864 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.50 vs. limit=5.0 +2023-03-26 19:58:13,286 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.629e+02 1.979e+02 2.321e+02 4.531e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-03-26 19:58:25,071 INFO [finetune.py:976] (0/7) Epoch 16, batch 4850, loss[loss=0.1879, simple_loss=0.2554, pruned_loss=0.06017, over 4905.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2543, pruned_loss=0.05792, over 956200.11 frames. ], batch size: 43, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:58:28,739 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 19:58:35,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3438, 1.4681, 1.9992, 1.7074, 1.6250, 3.4179, 1.3193, 1.5213], + device='cuda:0'), covar=tensor([0.1062, 0.1778, 0.1153, 0.1037, 0.1626, 0.0243, 0.1529, 0.1809], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0078, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 19:58:39,082 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90786.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:58:57,836 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90815.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:58:58,402 INFO [finetune.py:976] (0/7) Epoch 16, batch 4900, loss[loss=0.195, simple_loss=0.2628, pruned_loss=0.06361, over 4775.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2567, pruned_loss=0.05902, over 956486.94 frames. ], batch size: 25, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:59:04,107 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0892, 2.1180, 1.6621, 2.1965, 2.1374, 1.7868, 2.4083, 2.1496], + device='cuda:0'), covar=tensor([0.1350, 0.2081, 0.2817, 0.2403, 0.2360, 0.1619, 0.3181, 0.1734], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0187, 0.0234, 0.0253, 0.0245, 0.0202, 0.0212, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:59:11,163 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:11,784 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4676, 3.9013, 4.0746, 4.3082, 4.2348, 3.9374, 4.5812, 1.3853], + device='cuda:0'), covar=tensor([0.0753, 0.0840, 0.0811, 0.0884, 0.1171, 0.1650, 0.0587, 0.5734], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0242, 0.0274, 0.0291, 0.0331, 0.0278, 0.0296, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:59:24,122 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.543e+02 1.926e+02 2.205e+02 3.945e+02, threshold=3.852e+02, percent-clipped=0.0 +2023-03-26 19:59:25,317 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:26,597 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5665, 2.4401, 2.0964, 2.6630, 2.3527, 2.3322, 2.3578, 3.3357], + device='cuda:0'), covar=tensor([0.3832, 0.5016, 0.3552, 0.4110, 0.4386, 0.2496, 0.4200, 0.1620], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0225, 0.0275, 0.0248, 0.0216, 0.0249, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 19:59:27,117 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:42,911 INFO [finetune.py:976] (0/7) Epoch 16, batch 4950, loss[loss=0.184, simple_loss=0.2618, pruned_loss=0.05312, over 4894.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2577, pruned_loss=0.05912, over 956829.40 frames. ], batch size: 36, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:00:12,903 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:00:34,909 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90910.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:00:38,892 INFO [finetune.py:976] (0/7) Epoch 16, batch 5000, loss[loss=0.1978, simple_loss=0.2727, pruned_loss=0.06144, over 4891.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2558, pruned_loss=0.05848, over 957548.68 frames. ], batch size: 35, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:00:53,036 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:01:00,204 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.882e+01 1.531e+02 1.843e+02 2.134e+02 5.620e+02, threshold=3.687e+02, percent-clipped=2.0 +2023-03-26 20:01:11,975 INFO [finetune.py:976] (0/7) Epoch 16, batch 5050, loss[loss=0.1476, simple_loss=0.2214, pruned_loss=0.03687, over 4827.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2532, pruned_loss=0.05813, over 956851.59 frames. ], batch size: 51, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:01:40,185 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:01:45,574 INFO [finetune.py:976] (0/7) Epoch 16, batch 5100, loss[loss=0.1472, simple_loss=0.2044, pruned_loss=0.04493, over 4207.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2501, pruned_loss=0.05659, over 954843.81 frames. ], batch size: 18, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:02:07,770 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.471e+02 1.778e+02 2.096e+02 3.940e+02, threshold=3.556e+02, percent-clipped=2.0 +2023-03-26 20:02:12,116 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:02:18,685 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7766, 3.7287, 3.5124, 1.7003, 3.8393, 2.9263, 0.7913, 2.6147], + device='cuda:0'), covar=tensor([0.2521, 0.2238, 0.1568, 0.3664, 0.1047, 0.0973, 0.4734, 0.1694], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0176, 0.0159, 0.0129, 0.0158, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 20:02:19,226 INFO [finetune.py:976] (0/7) Epoch 16, batch 5150, loss[loss=0.2332, simple_loss=0.3086, pruned_loss=0.07893, over 4866.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2501, pruned_loss=0.05656, over 956174.77 frames. ], batch size: 44, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:02:52,453 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:02:52,957 INFO [finetune.py:976] (0/7) Epoch 16, batch 5200, loss[loss=0.169, simple_loss=0.2394, pruned_loss=0.04928, over 4734.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2542, pruned_loss=0.05803, over 955785.54 frames. ], batch size: 27, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:03:07,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1879, 1.7524, 2.1972, 2.0808, 1.8480, 1.9036, 2.0297, 2.0026], + device='cuda:0'), covar=tensor([0.4319, 0.4467, 0.3282, 0.4227, 0.5648, 0.4250, 0.5331, 0.3305], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0239, 0.0258, 0.0271, 0.0270, 0.0245, 0.0281, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:03:14,340 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.594e+02 1.966e+02 2.465e+02 4.658e+02, threshold=3.932e+02, percent-clipped=3.0 +2023-03-26 20:03:17,835 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91152.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:24,390 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:26,646 INFO [finetune.py:976] (0/7) Epoch 16, batch 5250, loss[loss=0.2358, simple_loss=0.2962, pruned_loss=0.08773, over 4906.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.256, pruned_loss=0.05831, over 956419.65 frames. ], batch size: 36, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:03:31,016 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0732, 1.5045, 2.1130, 2.0554, 1.8460, 1.8104, 2.0304, 1.9509], + device='cuda:0'), covar=tensor([0.3758, 0.3971, 0.3341, 0.3374, 0.4809, 0.3681, 0.4257, 0.3219], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0239, 0.0258, 0.0271, 0.0270, 0.0245, 0.0281, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:03:49,261 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91200.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:53,268 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91205.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:59,840 INFO [finetune.py:976] (0/7) Epoch 16, batch 5300, loss[loss=0.1708, simple_loss=0.2531, pruned_loss=0.04421, over 4886.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2558, pruned_loss=0.05826, over 955224.53 frames. ], batch size: 32, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:04:15,522 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 20:04:21,114 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.597e+02 1.843e+02 2.222e+02 3.769e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-26 20:04:33,500 INFO [finetune.py:976] (0/7) Epoch 16, batch 5350, loss[loss=0.184, simple_loss=0.2485, pruned_loss=0.05975, over 4181.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2559, pruned_loss=0.05778, over 955325.92 frames. ], batch size: 65, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:04:50,618 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91285.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:05:29,810 INFO [finetune.py:976] (0/7) Epoch 16, batch 5400, loss[loss=0.1716, simple_loss=0.2481, pruned_loss=0.04757, over 4828.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2541, pruned_loss=0.05714, over 955535.47 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:05:58,068 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4615, 1.4035, 1.8880, 2.9271, 1.9542, 2.1764, 1.0143, 2.4078], + device='cuda:0'), covar=tensor([0.1742, 0.1464, 0.1235, 0.0681, 0.0868, 0.1307, 0.1755, 0.0579], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0165, 0.0101, 0.0138, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:06:01,742 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:06:03,346 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.781e+01 1.583e+02 1.812e+02 2.291e+02 3.767e+02, threshold=3.624e+02, percent-clipped=1.0 +2023-03-26 20:06:15,750 INFO [finetune.py:976] (0/7) Epoch 16, batch 5450, loss[loss=0.1941, simple_loss=0.2503, pruned_loss=0.06894, over 4826.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2504, pruned_loss=0.05587, over 956375.44 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:06:49,418 INFO [finetune.py:976] (0/7) Epoch 16, batch 5500, loss[loss=0.2057, simple_loss=0.2706, pruned_loss=0.07042, over 4161.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2478, pruned_loss=0.05553, over 952400.21 frames. ], batch size: 65, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:07:10,222 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.789e+01 1.460e+02 1.744e+02 2.187e+02 6.443e+02, threshold=3.488e+02, percent-clipped=1.0 +2023-03-26 20:07:22,092 INFO [finetune.py:976] (0/7) Epoch 16, batch 5550, loss[loss=0.2178, simple_loss=0.3004, pruned_loss=0.06757, over 4852.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2513, pruned_loss=0.0578, over 952407.83 frames. ], batch size: 47, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:07:22,220 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6516, 1.3887, 0.9538, 0.2348, 1.1412, 1.4328, 1.3692, 1.3371], + device='cuda:0'), covar=tensor([0.0968, 0.0907, 0.1468, 0.1986, 0.1500, 0.2545, 0.2342, 0.0942], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0194, 0.0198, 0.0181, 0.0211, 0.0205, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:07:23,911 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7186, 1.4926, 2.0479, 1.2615, 1.9000, 1.9963, 1.4888, 2.1889], + device='cuda:0'), covar=tensor([0.1311, 0.2231, 0.1378, 0.1957, 0.0820, 0.1417, 0.2856, 0.0754], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0203, 0.0191, 0.0190, 0.0176, 0.0212, 0.0218, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:07:47,583 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:07:53,903 INFO [finetune.py:976] (0/7) Epoch 16, batch 5600, loss[loss=0.2126, simple_loss=0.2882, pruned_loss=0.06847, over 4816.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2561, pruned_loss=0.05992, over 953791.29 frames. ], batch size: 41, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:08:08,997 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91542.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:08:13,218 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.603e+02 1.969e+02 2.458e+02 5.397e+02, threshold=3.938e+02, percent-clipped=5.0 +2023-03-26 20:08:13,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4212, 1.5433, 1.5986, 0.8750, 1.6263, 1.8165, 1.8110, 1.4121], + device='cuda:0'), covar=tensor([0.0970, 0.0767, 0.0468, 0.0568, 0.0426, 0.0786, 0.0332, 0.0827], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0151, 0.0125, 0.0128, 0.0131, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.3002e-05, 1.1000e-04, 8.9363e-05, 9.1079e-05, 9.2463e-05, 9.2939e-05, + 1.0321e-04, 1.0680e-04], device='cuda:0') +2023-03-26 20:08:16,172 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:08:23,625 INFO [finetune.py:976] (0/7) Epoch 16, batch 5650, loss[loss=0.2483, simple_loss=0.3075, pruned_loss=0.09455, over 4743.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2573, pruned_loss=0.05984, over 953752.71 frames. ], batch size: 59, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:08:36,659 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 20:08:45,683 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91603.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:08:53,175 INFO [finetune.py:976] (0/7) Epoch 16, batch 5700, loss[loss=0.1285, simple_loss=0.2043, pruned_loss=0.02631, over 4386.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2527, pruned_loss=0.05844, over 938728.68 frames. ], batch size: 19, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:08:57,518 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 20:09:07,889 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:09:09,381 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-16.pt +2023-03-26 20:09:21,371 INFO [finetune.py:976] (0/7) Epoch 17, batch 0, loss[loss=0.245, simple_loss=0.3059, pruned_loss=0.09203, over 4802.00 frames. ], tot_loss[loss=0.245, simple_loss=0.3059, pruned_loss=0.09203, over 4802.00 frames. ], batch size: 39, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:09:21,372 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 20:09:23,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8033, 1.1100, 1.9015, 1.8029, 1.6703, 1.5953, 1.7320, 1.8031], + device='cuda:0'), covar=tensor([0.3890, 0.4145, 0.3662, 0.3914, 0.5238, 0.4028, 0.4867, 0.3227], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0240, 0.0259, 0.0271, 0.0270, 0.0245, 0.0282, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:09:23,627 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8086, 3.3864, 3.5029, 3.6787, 3.5905, 3.3777, 3.8968, 1.3666], + device='cuda:0'), covar=tensor([0.0836, 0.0811, 0.0857, 0.0984, 0.1250, 0.1573, 0.0755, 0.5010], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0241, 0.0272, 0.0290, 0.0330, 0.0278, 0.0296, 0.0291], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:09:24,148 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5429, 1.3067, 1.3689, 1.5064, 1.7579, 1.6679, 1.3978, 1.2937], + device='cuda:0'), covar=tensor([0.0333, 0.0359, 0.0577, 0.0318, 0.0230, 0.0386, 0.0353, 0.0442], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3808e-05, 8.3862e-05, 1.1375e-04, 8.6997e-05, 7.7713e-05, 7.9723e-05, + 7.3573e-05, 8.3274e-05], device='cuda:0') +2023-03-26 20:09:32,015 INFO [finetune.py:1010] (0/7) Epoch 17, validation: loss=0.1591, simple_loss=0.2283, pruned_loss=0.04492, over 2265189.00 frames. +2023-03-26 20:09:32,015 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 20:09:35,490 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.774e+01 1.479e+02 1.757e+02 2.057e+02 5.096e+02, threshold=3.514e+02, percent-clipped=1.0 +2023-03-26 20:10:00,083 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 20:10:07,336 INFO [finetune.py:976] (0/7) Epoch 17, batch 50, loss[loss=0.1232, simple_loss=0.2102, pruned_loss=0.01811, over 4754.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2602, pruned_loss=0.06061, over 216479.42 frames. ], batch size: 28, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:10:16,238 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6870, 2.4563, 1.9838, 0.9583, 2.1796, 2.0222, 1.8172, 2.2593], + device='cuda:0'), covar=tensor([0.0932, 0.0842, 0.1602, 0.2160, 0.1531, 0.2369, 0.2121, 0.0982], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0194, 0.0198, 0.0181, 0.0211, 0.0205, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:10:23,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 20:10:31,676 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6460, 1.3274, 2.0870, 3.2857, 2.1686, 2.4721, 1.0468, 2.7953], + device='cuda:0'), covar=tensor([0.1934, 0.2112, 0.1567, 0.0971, 0.1047, 0.1531, 0.2085, 0.0651], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0164, 0.0100, 0.0137, 0.0123, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:10:52,789 INFO [finetune.py:976] (0/7) Epoch 17, batch 100, loss[loss=0.1733, simple_loss=0.2254, pruned_loss=0.06058, over 4905.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2535, pruned_loss=0.05976, over 379711.10 frames. ], batch size: 37, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:11:01,250 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.600e+02 1.810e+02 2.096e+02 3.529e+02, threshold=3.620e+02, percent-clipped=1.0 +2023-03-26 20:11:09,182 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 20:11:25,184 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.07 vs. limit=5.0 +2023-03-26 20:11:37,625 INFO [finetune.py:976] (0/7) Epoch 17, batch 150, loss[loss=0.1691, simple_loss=0.2364, pruned_loss=0.05083, over 4862.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2465, pruned_loss=0.05617, over 507948.81 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:11,016 INFO [finetune.py:976] (0/7) Epoch 17, batch 200, loss[loss=0.2253, simple_loss=0.2856, pruned_loss=0.08254, over 4824.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2476, pruned_loss=0.05648, over 607554.38 frames. ], batch size: 38, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:11,698 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3162, 1.4345, 1.6797, 1.5655, 1.5113, 3.0060, 1.2773, 1.5088], + device='cuda:0'), covar=tensor([0.1008, 0.1868, 0.0990, 0.0957, 0.1572, 0.0322, 0.1627, 0.1868], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:12:14,524 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.595e+02 1.938e+02 2.273e+02 4.627e+02, threshold=3.876e+02, percent-clipped=4.0 +2023-03-26 20:12:43,709 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91891.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:12:44,793 INFO [finetune.py:976] (0/7) Epoch 17, batch 250, loss[loss=0.1761, simple_loss=0.2496, pruned_loss=0.05131, over 4897.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2521, pruned_loss=0.05779, over 687100.10 frames. ], batch size: 35, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:47,912 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91898.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:12:50,134 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0265, 2.0217, 1.7471, 1.8289, 2.5202, 2.5465, 2.0814, 2.0938], + device='cuda:0'), covar=tensor([0.0369, 0.0415, 0.0571, 0.0385, 0.0241, 0.0481, 0.0354, 0.0416], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.4044e-05, 8.3918e-05, 1.1392e-04, 8.7083e-05, 7.7791e-05, 7.9603e-05, + 7.3818e-05, 8.3000e-05], device='cuda:0') +2023-03-26 20:13:14,603 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7495, 1.6457, 1.7899, 1.1276, 1.7572, 1.8632, 1.8283, 1.5035], + device='cuda:0'), covar=tensor([0.0585, 0.0662, 0.0620, 0.0908, 0.0848, 0.0631, 0.0557, 0.1124], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0141, 0.0124, 0.0124, 0.0140, 0.0141, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:13:17,049 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:18,231 INFO [finetune.py:976] (0/7) Epoch 17, batch 300, loss[loss=0.1771, simple_loss=0.2454, pruned_loss=0.05442, over 4862.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.255, pruned_loss=0.05811, over 747744.26 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:13:21,758 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.873e+01 1.605e+02 2.003e+02 2.239e+02 3.510e+02, threshold=4.006e+02, percent-clipped=0.0 +2023-03-26 20:13:21,958 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 20:13:24,424 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91952.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:26,029 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6777, 3.3663, 3.3572, 1.9231, 3.4641, 2.7458, 1.2554, 2.3963], + device='cuda:0'), covar=tensor([0.2818, 0.1924, 0.1315, 0.2784, 0.1080, 0.0995, 0.3681, 0.1673], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0158, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 20:13:27,180 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:33,612 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7265, 1.7057, 1.5614, 1.6393, 1.3533, 4.4049, 1.6112, 1.9898], + device='cuda:0'), covar=tensor([0.3324, 0.2564, 0.2246, 0.2363, 0.1738, 0.0137, 0.2548, 0.1315], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0096, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:13:40,210 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6106, 1.7509, 1.4273, 1.6123, 2.1774, 2.0809, 1.7979, 1.8311], + device='cuda:0'), covar=tensor([0.0442, 0.0322, 0.0570, 0.0357, 0.0257, 0.0500, 0.0348, 0.0356], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.4055e-05, 8.3752e-05, 1.1391e-04, 8.7247e-05, 7.7784e-05, 7.9726e-05, + 7.3843e-05, 8.2929e-05], device='cuda:0') +2023-03-26 20:13:42,100 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.64 vs. limit=5.0 +2023-03-26 20:13:44,455 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91981.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:13:49,289 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:51,727 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 20:13:52,144 INFO [finetune.py:976] (0/7) Epoch 17, batch 350, loss[loss=0.1858, simple_loss=0.2643, pruned_loss=0.05367, over 4821.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2559, pruned_loss=0.05801, over 794826.64 frames. ], batch size: 33, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:13:57,042 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-92000.pt +2023-03-26 20:14:09,826 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:14:19,915 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2393, 2.8393, 2.6529, 1.2576, 2.9443, 2.2367, 0.8304, 1.9073], + device='cuda:0'), covar=tensor([0.2852, 0.2163, 0.1837, 0.3152, 0.1367, 0.1037, 0.3774, 0.1569], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0172, 0.0157, 0.0127, 0.0156, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 20:14:26,168 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92042.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 20:14:26,650 INFO [finetune.py:976] (0/7) Epoch 17, batch 400, loss[loss=0.1775, simple_loss=0.2484, pruned_loss=0.05334, over 4789.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2561, pruned_loss=0.05788, over 830299.96 frames. ], batch size: 51, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:14:30,187 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.544e+02 1.847e+02 2.163e+02 3.487e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 20:14:37,416 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 20:15:00,228 INFO [finetune.py:976] (0/7) Epoch 17, batch 450, loss[loss=0.1776, simple_loss=0.246, pruned_loss=0.05461, over 4875.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2543, pruned_loss=0.05706, over 858483.75 frames. ], batch size: 34, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:15:33,735 INFO [finetune.py:976] (0/7) Epoch 17, batch 500, loss[loss=0.1438, simple_loss=0.216, pruned_loss=0.03586, over 4768.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2514, pruned_loss=0.05618, over 880301.42 frames. ], batch size: 28, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:15:37,217 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.578e+02 1.863e+02 2.269e+02 4.074e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-26 20:16:30,572 INFO [finetune.py:976] (0/7) Epoch 17, batch 550, loss[loss=0.1753, simple_loss=0.238, pruned_loss=0.05636, over 4889.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2492, pruned_loss=0.05613, over 897555.28 frames. ], batch size: 32, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:16:33,713 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92198.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:17:13,333 INFO [finetune.py:976] (0/7) Epoch 17, batch 600, loss[loss=0.1337, simple_loss=0.2061, pruned_loss=0.03063, over 4723.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2494, pruned_loss=0.05703, over 910732.23 frames. ], batch size: 23, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:17:15,211 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92246.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:17:15,824 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:17:16,353 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.709e+02 1.980e+02 2.349e+02 5.069e+02, threshold=3.960e+02, percent-clipped=5.0 +2023-03-26 20:17:47,090 INFO [finetune.py:976] (0/7) Epoch 17, batch 650, loss[loss=0.2101, simple_loss=0.2877, pruned_loss=0.06627, over 4833.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2511, pruned_loss=0.05709, over 921426.51 frames. ], batch size: 40, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:17:49,228 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-26 20:17:58,662 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:18:09,686 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2773, 2.1986, 1.8138, 2.2106, 2.0893, 2.0830, 2.0120, 3.1016], + device='cuda:0'), covar=tensor([0.3841, 0.5208, 0.3685, 0.4804, 0.4561, 0.2456, 0.5091, 0.1601], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0226, 0.0276, 0.0250, 0.0217, 0.0248, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:18:16,776 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92337.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:18:20,698 INFO [finetune.py:976] (0/7) Epoch 17, batch 700, loss[loss=0.1898, simple_loss=0.2504, pruned_loss=0.06462, over 4780.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.251, pruned_loss=0.05625, over 929792.46 frames. ], batch size: 26, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:18:23,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.553e+02 1.844e+02 2.135e+02 3.970e+02, threshold=3.688e+02, percent-clipped=1.0 +2023-03-26 20:18:45,806 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-26 20:18:54,353 INFO [finetune.py:976] (0/7) Epoch 17, batch 750, loss[loss=0.1946, simple_loss=0.2522, pruned_loss=0.06851, over 4700.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2533, pruned_loss=0.05729, over 935013.96 frames. ], batch size: 23, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:19:01,339 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 20:19:28,150 INFO [finetune.py:976] (0/7) Epoch 17, batch 800, loss[loss=0.1749, simple_loss=0.2502, pruned_loss=0.04978, over 4826.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2527, pruned_loss=0.05635, over 939859.70 frames. ], batch size: 33, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:19:31,195 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.753e+02 1.963e+02 2.342e+02 4.288e+02, threshold=3.926e+02, percent-clipped=2.0 +2023-03-26 20:19:41,726 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 20:20:01,479 INFO [finetune.py:976] (0/7) Epoch 17, batch 850, loss[loss=0.1929, simple_loss=0.26, pruned_loss=0.06287, over 4848.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2511, pruned_loss=0.0559, over 942104.96 frames. ], batch size: 44, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:20:11,794 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4388, 1.3786, 1.6389, 2.4652, 1.6295, 2.2020, 0.8808, 2.1260], + device='cuda:0'), covar=tensor([0.1613, 0.1370, 0.1086, 0.0634, 0.0931, 0.1109, 0.1468, 0.0561], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0164, 0.0101, 0.0136, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:20:35,310 INFO [finetune.py:976] (0/7) Epoch 17, batch 900, loss[loss=0.1458, simple_loss=0.2191, pruned_loss=0.03628, over 4867.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2485, pruned_loss=0.05521, over 944472.55 frames. ], batch size: 31, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:20:38,307 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:20:38,821 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.480e+02 1.791e+02 2.296e+02 4.324e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-26 20:21:05,827 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5986, 1.5073, 2.1950, 3.2833, 2.2153, 2.3602, 1.0817, 2.6745], + device='cuda:0'), covar=tensor([0.1610, 0.1350, 0.1153, 0.0564, 0.0732, 0.1687, 0.1618, 0.0499], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0164, 0.0101, 0.0136, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:21:15,098 INFO [finetune.py:976] (0/7) Epoch 17, batch 950, loss[loss=0.1745, simple_loss=0.2419, pruned_loss=0.05353, over 4819.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2473, pruned_loss=0.05526, over 946223.28 frames. ], batch size: 39, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:21:16,911 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:21:37,249 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92611.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:21:54,767 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92625.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:05,405 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92637.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:22:13,800 INFO [finetune.py:976] (0/7) Epoch 17, batch 1000, loss[loss=0.1905, simple_loss=0.261, pruned_loss=0.05996, over 4848.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2495, pruned_loss=0.05639, over 944977.34 frames. ], batch size: 47, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:22:20,426 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.492e+01 1.711e+02 2.074e+02 2.603e+02 6.251e+02, threshold=4.148e+02, percent-clipped=4.0 +2023-03-26 20:22:27,808 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:34,802 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 20:22:45,197 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92685.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:22:45,849 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:50,928 INFO [finetune.py:976] (0/7) Epoch 17, batch 1050, loss[loss=0.1784, simple_loss=0.2507, pruned_loss=0.05307, over 4828.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2539, pruned_loss=0.05747, over 946940.11 frames. ], batch size: 39, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:23:08,331 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:23:23,714 INFO [finetune.py:976] (0/7) Epoch 17, batch 1100, loss[loss=0.1901, simple_loss=0.2761, pruned_loss=0.05208, over 4919.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2547, pruned_loss=0.05745, over 949071.54 frames. ], batch size: 42, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:23:27,193 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.694e+02 2.013e+02 2.338e+02 4.806e+02, threshold=4.026e+02, percent-clipped=2.0 +2023-03-26 20:23:44,742 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6017, 1.5033, 2.0372, 3.1835, 2.1734, 2.2264, 1.0310, 2.6756], + device='cuda:0'), covar=tensor([0.1714, 0.1405, 0.1218, 0.0549, 0.0789, 0.1443, 0.1652, 0.0451], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0136, 0.0123, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:23:48,936 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:23:57,177 INFO [finetune.py:976] (0/7) Epoch 17, batch 1150, loss[loss=0.1933, simple_loss=0.2721, pruned_loss=0.05726, over 4903.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2546, pruned_loss=0.05673, over 950890.07 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:24:16,560 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 20:24:17,666 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 20:24:22,920 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:24:31,079 INFO [finetune.py:976] (0/7) Epoch 17, batch 1200, loss[loss=0.1558, simple_loss=0.2313, pruned_loss=0.04019, over 4743.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2528, pruned_loss=0.05604, over 952806.31 frames. ], batch size: 28, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:24:34,573 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.909e+01 1.547e+02 1.742e+02 2.125e+02 5.044e+02, threshold=3.483e+02, percent-clipped=2.0 +2023-03-26 20:24:45,178 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3668, 1.2764, 1.1581, 1.4149, 1.7157, 1.5436, 1.3116, 1.1724], + device='cuda:0'), covar=tensor([0.0323, 0.0322, 0.0686, 0.0320, 0.0200, 0.0447, 0.0311, 0.0356], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0143, 0.0113, 0.0099, 0.0107, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3535e-05, 8.3317e-05, 1.1334e-04, 8.7028e-05, 7.7094e-05, 7.9030e-05, + 7.3709e-05, 8.3095e-05], device='cuda:0') +2023-03-26 20:24:50,670 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1727, 1.9962, 1.7019, 1.8083, 2.1492, 1.8304, 2.2598, 2.0743], + device='cuda:0'), covar=tensor([0.1307, 0.1922, 0.2822, 0.2374, 0.2413, 0.1700, 0.2579, 0.1744], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0189, 0.0235, 0.0254, 0.0246, 0.0203, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:25:03,709 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:04,692 INFO [finetune.py:976] (0/7) Epoch 17, batch 1250, loss[loss=0.1674, simple_loss=0.2379, pruned_loss=0.04848, over 4911.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2503, pruned_loss=0.05559, over 951907.14 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:25:08,402 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:15,489 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1786, 1.9566, 1.4501, 0.6140, 1.7338, 1.7399, 1.6030, 1.8075], + device='cuda:0'), covar=tensor([0.0915, 0.0735, 0.1412, 0.1819, 0.1257, 0.2144, 0.2214, 0.0786], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0196, 0.0200, 0.0184, 0.0213, 0.0208, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:25:29,728 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92931.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:37,457 INFO [finetune.py:976] (0/7) Epoch 17, batch 1300, loss[loss=0.1589, simple_loss=0.227, pruned_loss=0.04543, over 4782.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2475, pruned_loss=0.05486, over 950921.85 frames. ], batch size: 26, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:25:41,344 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.503e+02 1.790e+02 2.154e+02 4.064e+02, threshold=3.581e+02, percent-clipped=1.0 +2023-03-26 20:25:49,683 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:59,951 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92975.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:03,546 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:10,746 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:11,262 INFO [finetune.py:976] (0/7) Epoch 17, batch 1350, loss[loss=0.2196, simple_loss=0.2766, pruned_loss=0.08133, over 4942.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2482, pruned_loss=0.05573, over 951657.46 frames. ], batch size: 33, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:26:23,833 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93010.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:51,519 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:51,532 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:00,929 INFO [finetune.py:976] (0/7) Epoch 17, batch 1400, loss[loss=0.1789, simple_loss=0.2526, pruned_loss=0.05258, over 4933.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2511, pruned_loss=0.05665, over 950992.89 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:27:08,967 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.549e+02 1.883e+02 2.310e+02 4.523e+02, threshold=3.767e+02, percent-clipped=3.0 +2023-03-26 20:27:12,640 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 20:27:34,704 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93071.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:39,794 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:50,101 INFO [finetune.py:976] (0/7) Epoch 17, batch 1450, loss[loss=0.1311, simple_loss=0.2151, pruned_loss=0.02351, over 4751.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2509, pruned_loss=0.0556, over 951561.00 frames. ], batch size: 27, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:27:52,033 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3699, 1.2612, 1.9080, 2.9152, 1.8665, 2.0885, 0.8864, 2.4104], + device='cuda:0'), covar=tensor([0.1837, 0.1570, 0.1219, 0.0661, 0.0939, 0.1513, 0.1804, 0.0552], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0166, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:27:53,128 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93097.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:27:58,964 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4813, 1.4589, 1.4029, 1.4978, 0.9838, 3.3238, 1.2439, 1.6900], + device='cuda:0'), covar=tensor([0.3453, 0.2538, 0.2273, 0.2444, 0.1959, 0.0225, 0.2800, 0.1374], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0096, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:28:05,999 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:07,298 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 20:28:23,787 INFO [finetune.py:976] (0/7) Epoch 17, batch 1500, loss[loss=0.1609, simple_loss=0.2242, pruned_loss=0.04883, over 4856.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2536, pruned_loss=0.05698, over 951586.41 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:28:27,866 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.651e+02 1.993e+02 2.270e+02 5.642e+02, threshold=3.987e+02, percent-clipped=1.0 +2023-03-26 20:28:47,149 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:53,709 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93187.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:53,815 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 20:28:57,278 INFO [finetune.py:976] (0/7) Epoch 17, batch 1550, loss[loss=0.1812, simple_loss=0.2557, pruned_loss=0.05336, over 4845.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2526, pruned_loss=0.05641, over 952621.83 frames. ], batch size: 44, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:29:23,423 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-26 20:29:30,932 INFO [finetune.py:976] (0/7) Epoch 17, batch 1600, loss[loss=0.1427, simple_loss=0.2162, pruned_loss=0.03455, over 4848.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2524, pruned_loss=0.05722, over 953871.55 frames. ], batch size: 47, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:29:33,611 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 20:29:34,594 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.546e+02 1.807e+02 2.216e+02 3.989e+02, threshold=3.613e+02, percent-clipped=1.0 +2023-03-26 20:29:38,718 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:29:57,405 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:01,032 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93287.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:01,053 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5628, 1.4764, 1.9176, 1.7765, 1.6555, 3.2480, 1.4196, 1.6223], + device='cuda:0'), covar=tensor([0.0987, 0.1712, 0.1236, 0.0973, 0.1475, 0.0222, 0.1387, 0.1602], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0078, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:30:04,615 INFO [finetune.py:976] (0/7) Epoch 17, batch 1650, loss[loss=0.1539, simple_loss=0.2114, pruned_loss=0.04819, over 4919.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2487, pruned_loss=0.05618, over 954621.96 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:30:28,975 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:30,203 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93331.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:38,310 INFO [finetune.py:976] (0/7) Epoch 17, batch 1700, loss[loss=0.195, simple_loss=0.2652, pruned_loss=0.06241, over 4895.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2467, pruned_loss=0.05522, over 955417.86 frames. ], batch size: 43, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:30:41,943 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.487e+02 1.694e+02 2.142e+02 3.933e+02, threshold=3.388e+02, percent-clipped=2.0 +2023-03-26 20:30:43,329 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3507, 1.4872, 1.2296, 1.4865, 1.6979, 1.7681, 1.4531, 1.4597], + device='cuda:0'), covar=tensor([0.0356, 0.0314, 0.0531, 0.0291, 0.0202, 0.0336, 0.0365, 0.0298], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0108, 0.0142, 0.0112, 0.0099, 0.0107, 0.0098, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.3144e-05, 8.3442e-05, 1.1255e-04, 8.6618e-05, 7.6965e-05, 7.8871e-05, + 7.3317e-05, 8.2823e-05], device='cuda:0') +2023-03-26 20:30:48,108 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1729, 1.8003, 2.1557, 2.1139, 1.8276, 1.8957, 2.0851, 2.0088], + device='cuda:0'), covar=tensor([0.3988, 0.4320, 0.3285, 0.4225, 0.5186, 0.3942, 0.5048, 0.3301], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0241, 0.0259, 0.0273, 0.0271, 0.0246, 0.0283, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:30:53,335 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93366.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:31:00,369 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:31:05,673 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6311, 1.5828, 1.2932, 1.6539, 1.9937, 1.8748, 1.5107, 1.4925], + device='cuda:0'), covar=tensor([0.0314, 0.0336, 0.0637, 0.0310, 0.0202, 0.0516, 0.0329, 0.0412], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0113, 0.0099, 0.0107, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3322e-05, 8.3665e-05, 1.1296e-04, 8.6909e-05, 7.7269e-05, 7.9135e-05, + 7.3553e-05, 8.3133e-05], device='cuda:0') +2023-03-26 20:31:06,277 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2090, 2.1458, 2.1354, 1.5988, 2.1953, 2.3277, 2.3962, 1.7798], + device='cuda:0'), covar=tensor([0.0556, 0.0619, 0.0708, 0.0895, 0.0637, 0.0617, 0.0538, 0.1087], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0136, 0.0143, 0.0124, 0.0125, 0.0141, 0.0143, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:31:11,620 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93392.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:31:12,183 INFO [finetune.py:976] (0/7) Epoch 17, batch 1750, loss[loss=0.1627, simple_loss=0.2341, pruned_loss=0.04562, over 4867.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2482, pruned_loss=0.0556, over 954020.54 frames. ], batch size: 31, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:31:33,278 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93424.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:31:48,778 INFO [finetune.py:976] (0/7) Epoch 17, batch 1800, loss[loss=0.2205, simple_loss=0.2732, pruned_loss=0.08394, over 4155.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2511, pruned_loss=0.05664, over 954621.73 frames. ], batch size: 65, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:31:56,895 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.552e+02 1.846e+02 2.179e+02 3.576e+02, threshold=3.692e+02, percent-clipped=3.0 +2023-03-26 20:32:20,906 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:32:40,555 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93487.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:32:49,592 INFO [finetune.py:976] (0/7) Epoch 17, batch 1850, loss[loss=0.1516, simple_loss=0.2266, pruned_loss=0.03828, over 4789.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2526, pruned_loss=0.05738, over 955206.60 frames. ], batch size: 25, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:33:20,425 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93535.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:26,142 INFO [finetune.py:976] (0/7) Epoch 17, batch 1900, loss[loss=0.1526, simple_loss=0.2346, pruned_loss=0.03526, over 4814.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2538, pruned_loss=0.05754, over 954516.16 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:33:30,352 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.618e+02 1.925e+02 2.327e+02 3.543e+02, threshold=3.851e+02, percent-clipped=0.0 +2023-03-26 20:33:32,495 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-03-26 20:33:34,128 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:55,449 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93587.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:59,455 INFO [finetune.py:976] (0/7) Epoch 17, batch 1950, loss[loss=0.1401, simple_loss=0.2288, pruned_loss=0.02575, over 4762.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2525, pruned_loss=0.05654, over 955978.42 frames. ], batch size: 28, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:34:06,616 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93603.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:24,526 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93631.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:27,388 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:32,610 INFO [finetune.py:976] (0/7) Epoch 17, batch 2000, loss[loss=0.1371, simple_loss=0.2043, pruned_loss=0.03491, over 4817.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2514, pruned_loss=0.05683, over 955949.93 frames. ], batch size: 25, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:34:37,203 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.489e+01 1.528e+02 1.753e+02 2.103e+02 5.258e+02, threshold=3.506e+02, percent-clipped=1.0 +2023-03-26 20:34:43,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3781, 1.3613, 1.7365, 2.4219, 1.6852, 2.1916, 0.9001, 2.0846], + device='cuda:0'), covar=tensor([0.1697, 0.1413, 0.1057, 0.0692, 0.0867, 0.1204, 0.1553, 0.0609], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0165, 0.0101, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:34:48,128 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:56,474 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:01,047 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3083, 2.9231, 3.0779, 3.2185, 3.0761, 2.9201, 3.3173, 0.9757], + device='cuda:0'), covar=tensor([0.0989, 0.0990, 0.1024, 0.1146, 0.1594, 0.1763, 0.1172, 0.5427], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0292, 0.0333, 0.0281, 0.0300, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:35:05,795 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:06,327 INFO [finetune.py:976] (0/7) Epoch 17, batch 2050, loss[loss=0.1395, simple_loss=0.2157, pruned_loss=0.03163, over 4915.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2483, pruned_loss=0.05587, over 955652.18 frames. ], batch size: 35, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:35:08,320 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 20:35:20,564 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93714.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:29,714 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 20:35:37,800 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:39,563 INFO [finetune.py:976] (0/7) Epoch 17, batch 2100, loss[loss=0.1902, simple_loss=0.2533, pruned_loss=0.0636, over 4822.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2487, pruned_loss=0.05638, over 956012.47 frames. ], batch size: 30, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:35:43,619 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.402e+01 1.568e+02 1.860e+02 2.232e+02 5.340e+02, threshold=3.720e+02, percent-clipped=4.0 +2023-03-26 20:35:58,017 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:58,598 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:12,207 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2874, 2.8725, 3.0202, 3.2084, 3.0314, 2.8599, 3.3258, 0.9939], + device='cuda:0'), covar=tensor([0.1191, 0.1126, 0.1122, 0.1429, 0.1879, 0.1968, 0.1232, 0.5743], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0292, 0.0334, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:36:13,275 INFO [finetune.py:976] (0/7) Epoch 17, batch 2150, loss[loss=0.286, simple_loss=0.3347, pruned_loss=0.1187, over 4925.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2523, pruned_loss=0.05741, over 953376.77 frames. ], batch size: 38, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:36:24,176 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 20:36:26,355 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5659, 1.3798, 2.2708, 3.3344, 2.2808, 2.3070, 1.0888, 2.6379], + device='cuda:0'), covar=tensor([0.1788, 0.1613, 0.1188, 0.0511, 0.0773, 0.1716, 0.1798, 0.0563], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:36:31,172 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:38,624 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:47,327 INFO [finetune.py:976] (0/7) Epoch 17, batch 2200, loss[loss=0.2147, simple_loss=0.2819, pruned_loss=0.07378, over 4912.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2528, pruned_loss=0.05694, over 954018.71 frames. ], batch size: 36, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:36:51,478 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.569e+02 1.869e+02 2.308e+02 4.137e+02, threshold=3.738e+02, percent-clipped=1.0 +2023-03-26 20:37:01,713 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3846, 1.3123, 1.4761, 1.5215, 1.4502, 2.9380, 1.2037, 1.4054], + device='cuda:0'), covar=tensor([0.0962, 0.1936, 0.1244, 0.1029, 0.1783, 0.0275, 0.1644, 0.1839], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0079, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:37:36,162 INFO [finetune.py:976] (0/7) Epoch 17, batch 2250, loss[loss=0.177, simple_loss=0.2484, pruned_loss=0.05281, over 4816.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2541, pruned_loss=0.05764, over 954904.01 frames. ], batch size: 45, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:38:30,052 INFO [finetune.py:976] (0/7) Epoch 17, batch 2300, loss[loss=0.1539, simple_loss=0.2255, pruned_loss=0.04112, over 4765.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2549, pruned_loss=0.0575, over 955757.77 frames. ], batch size: 26, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:38:34,185 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.601e+02 1.890e+02 2.328e+02 3.292e+02, threshold=3.781e+02, percent-clipped=0.0 +2023-03-26 20:38:56,104 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:39:03,825 INFO [finetune.py:976] (0/7) Epoch 17, batch 2350, loss[loss=0.1864, simple_loss=0.2469, pruned_loss=0.06297, over 4729.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.253, pruned_loss=0.05684, over 954028.65 frames. ], batch size: 54, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:39:08,699 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-94000.pt +2023-03-26 20:39:35,518 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2978, 1.9400, 2.2903, 2.1437, 1.9376, 1.9325, 2.1576, 2.1009], + device='cuda:0'), covar=tensor([0.4168, 0.4533, 0.3658, 0.4299, 0.5560, 0.4339, 0.5358, 0.3394], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0273, 0.0272, 0.0247, 0.0284, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:39:37,901 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94042.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:39:38,383 INFO [finetune.py:976] (0/7) Epoch 17, batch 2400, loss[loss=0.163, simple_loss=0.2254, pruned_loss=0.05025, over 4914.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2504, pruned_loss=0.05615, over 952141.72 frames. ], batch size: 43, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:39:42,505 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.475e+02 1.781e+02 2.110e+02 4.538e+02, threshold=3.563e+02, percent-clipped=2.0 +2023-03-26 20:40:06,875 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6471, 1.6237, 2.0100, 1.8622, 1.7162, 3.6056, 1.5440, 1.6447], + device='cuda:0'), covar=tensor([0.0882, 0.1697, 0.1065, 0.0948, 0.1472, 0.0196, 0.1418, 0.1676], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:40:11,636 INFO [finetune.py:976] (0/7) Epoch 17, batch 2450, loss[loss=0.2076, simple_loss=0.281, pruned_loss=0.06711, over 4818.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2474, pruned_loss=0.05522, over 954697.01 frames. ], batch size: 45, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:40:29,749 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 20:40:34,652 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:40:37,218 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 20:40:42,588 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 20:40:45,406 INFO [finetune.py:976] (0/7) Epoch 17, batch 2500, loss[loss=0.2305, simple_loss=0.2967, pruned_loss=0.0821, over 4903.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2499, pruned_loss=0.05651, over 954049.12 frames. ], batch size: 35, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:40:49,550 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.530e+01 1.683e+02 1.909e+02 2.220e+02 4.342e+02, threshold=3.819e+02, percent-clipped=2.0 +2023-03-26 20:41:09,044 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3327, 3.7565, 3.9863, 4.1976, 4.0908, 3.8669, 4.4200, 1.3156], + device='cuda:0'), covar=tensor([0.0731, 0.0806, 0.0817, 0.1065, 0.1107, 0.1462, 0.0619, 0.5667], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0276, 0.0293, 0.0335, 0.0281, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:41:18,599 INFO [finetune.py:976] (0/7) Epoch 17, batch 2550, loss[loss=0.2265, simple_loss=0.2968, pruned_loss=0.07808, over 4823.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2533, pruned_loss=0.05689, over 953856.32 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:41:18,707 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9880, 1.5030, 2.3334, 1.5139, 2.1424, 2.1145, 1.5690, 2.2928], + device='cuda:0'), covar=tensor([0.1229, 0.2081, 0.1268, 0.1861, 0.0777, 0.1392, 0.2770, 0.0750], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0203, 0.0188, 0.0188, 0.0175, 0.0211, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:41:36,393 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9776, 2.0891, 1.6814, 1.8489, 2.5170, 2.5961, 2.0732, 2.0056], + device='cuda:0'), covar=tensor([0.0346, 0.0286, 0.0560, 0.0340, 0.0192, 0.0431, 0.0366, 0.0400], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0106, 0.0142, 0.0111, 0.0098, 0.0106, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2644e-05, 8.2039e-05, 1.1184e-04, 8.5786e-05, 7.6515e-05, 7.8633e-05, + 7.2730e-05, 8.2582e-05], device='cuda:0') +2023-03-26 20:41:38,777 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94223.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:41:48,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7381, 1.6094, 1.4334, 1.7296, 2.1292, 1.7954, 1.4623, 1.4217], + device='cuda:0'), covar=tensor([0.2054, 0.1951, 0.1791, 0.1659, 0.1636, 0.1195, 0.2302, 0.1835], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0212, 0.0192, 0.0241, 0.0185, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:41:52,380 INFO [finetune.py:976] (0/7) Epoch 17, batch 2600, loss[loss=0.1901, simple_loss=0.26, pruned_loss=0.06007, over 4719.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2545, pruned_loss=0.05727, over 953016.99 frames. ], batch size: 59, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:41:56,016 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.647e+02 1.955e+02 2.265e+02 3.573e+02, threshold=3.911e+02, percent-clipped=0.0 +2023-03-26 20:41:56,141 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5016, 1.6124, 1.3368, 1.5768, 1.9094, 1.8323, 1.5622, 1.4737], + device='cuda:0'), covar=tensor([0.0341, 0.0280, 0.0697, 0.0293, 0.0209, 0.0464, 0.0353, 0.0362], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0106, 0.0142, 0.0111, 0.0099, 0.0107, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.2950e-05, 8.2321e-05, 1.1228e-04, 8.6046e-05, 7.6890e-05, 7.8957e-05, + 7.3081e-05, 8.2857e-05], device='cuda:0') +2023-03-26 20:42:17,835 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7557, 1.3364, 0.8393, 1.5675, 2.1426, 1.2462, 1.4344, 1.5290], + device='cuda:0'), covar=tensor([0.1431, 0.2028, 0.1869, 0.1209, 0.1844, 0.1984, 0.1462, 0.1937], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:42:19,684 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94284.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:42:22,078 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 20:42:25,432 INFO [finetune.py:976] (0/7) Epoch 17, batch 2650, loss[loss=0.2099, simple_loss=0.2795, pruned_loss=0.07019, over 4919.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2558, pruned_loss=0.05771, over 955263.45 frames. ], batch size: 38, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:43:12,618 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:43:20,895 INFO [finetune.py:976] (0/7) Epoch 17, batch 2700, loss[loss=0.1849, simple_loss=0.2685, pruned_loss=0.05063, over 4885.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2535, pruned_loss=0.05608, over 954868.06 frames. ], batch size: 43, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:43:24,107 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:43:28,207 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.514e+02 1.766e+02 2.145e+02 4.618e+02, threshold=3.532e+02, percent-clipped=2.0 +2023-03-26 20:43:30,766 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4409, 1.5847, 1.8846, 1.8269, 1.6444, 3.4185, 1.4462, 1.7193], + device='cuda:0'), covar=tensor([0.0945, 0.1598, 0.1065, 0.0897, 0.1502, 0.0202, 0.1354, 0.1555], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0079, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:43:40,049 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9290, 1.7818, 1.5382, 1.4189, 1.9233, 1.6302, 1.8455, 1.8785], + device='cuda:0'), covar=tensor([0.1399, 0.2002, 0.3044, 0.2681, 0.2662, 0.1828, 0.2897, 0.1858], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0189, 0.0235, 0.0254, 0.0245, 0.0203, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:44:10,179 INFO [finetune.py:976] (0/7) Epoch 17, batch 2750, loss[loss=0.1988, simple_loss=0.2473, pruned_loss=0.07519, over 4910.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.251, pruned_loss=0.0557, over 952865.08 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:44:20,195 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94409.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:44:22,511 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5573, 2.2705, 1.6707, 0.7755, 1.9269, 2.0426, 1.9083, 1.9499], + device='cuda:0'), covar=tensor([0.0811, 0.0788, 0.1696, 0.2173, 0.1567, 0.2444, 0.2154, 0.1018], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0197, 0.0200, 0.0184, 0.0213, 0.0208, 0.0225, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:44:31,914 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94426.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:44:43,081 INFO [finetune.py:976] (0/7) Epoch 17, batch 2800, loss[loss=0.1996, simple_loss=0.263, pruned_loss=0.06811, over 4907.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2481, pruned_loss=0.05521, over 954422.64 frames. ], batch size: 37, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:44:47,181 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.577e+01 1.587e+02 1.887e+02 2.313e+02 4.372e+02, threshold=3.775e+02, percent-clipped=5.0 +2023-03-26 20:45:02,860 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:45:16,199 INFO [finetune.py:976] (0/7) Epoch 17, batch 2850, loss[loss=0.2152, simple_loss=0.2724, pruned_loss=0.07902, over 4759.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2479, pruned_loss=0.05549, over 955291.77 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:45:19,536 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 20:45:49,603 INFO [finetune.py:976] (0/7) Epoch 17, batch 2900, loss[loss=0.192, simple_loss=0.286, pruned_loss=0.04903, over 4801.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2493, pruned_loss=0.05586, over 952174.73 frames. ], batch size: 41, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:45:53,201 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.550e+02 1.801e+02 2.117e+02 3.911e+02, threshold=3.601e+02, percent-clipped=1.0 +2023-03-26 20:46:00,457 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4385, 1.4840, 2.0657, 3.1755, 2.1257, 2.1353, 1.1478, 2.6222], + device='cuda:0'), covar=tensor([0.1829, 0.1489, 0.1233, 0.0548, 0.0838, 0.1620, 0.1686, 0.0500], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0165, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:46:12,363 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94579.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:46:19,598 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-26 20:46:22,371 INFO [finetune.py:976] (0/7) Epoch 17, batch 2950, loss[loss=0.2123, simple_loss=0.2735, pruned_loss=0.07555, over 4896.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2533, pruned_loss=0.0574, over 950869.43 frames. ], batch size: 43, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:46:52,133 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94637.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:46:56,165 INFO [finetune.py:976] (0/7) Epoch 17, batch 3000, loss[loss=0.2069, simple_loss=0.2765, pruned_loss=0.06868, over 4889.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2525, pruned_loss=0.05663, over 950066.29 frames. ], batch size: 35, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:46:56,168 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 20:46:58,307 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2446, 1.4251, 1.4361, 0.7885, 1.3803, 1.5955, 1.6705, 1.3324], + device='cuda:0'), covar=tensor([0.0783, 0.0477, 0.0526, 0.0457, 0.0458, 0.0569, 0.0271, 0.0520], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0153, 0.0125, 0.0129, 0.0132, 0.0131, 0.0144, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.2855e-05, 1.1092e-04, 8.9644e-05, 9.1542e-05, 9.3197e-05, 9.3963e-05, + 1.0418e-04, 1.0779e-04], device='cuda:0') +2023-03-26 20:47:06,772 INFO [finetune.py:1010] (0/7) Epoch 17, validation: loss=0.1562, simple_loss=0.2257, pruned_loss=0.04335, over 2265189.00 frames. +2023-03-26 20:47:06,773 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 20:47:10,433 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.605e+02 1.916e+02 2.337e+02 3.800e+02, threshold=3.832e+02, percent-clipped=2.0 +2023-03-26 20:47:10,637 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-03-26 20:47:18,295 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 20:47:23,530 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7091, 1.6604, 1.6327, 1.0362, 1.8246, 1.9634, 1.9484, 1.5135], + device='cuda:0'), covar=tensor([0.0870, 0.0515, 0.0537, 0.0563, 0.0427, 0.0540, 0.0320, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0153, 0.0125, 0.0129, 0.0132, 0.0131, 0.0145, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.2826e-05, 1.1087e-04, 8.9727e-05, 9.1426e-05, 9.3251e-05, 9.3873e-05, + 1.0421e-04, 1.0779e-04], device='cuda:0') +2023-03-26 20:47:33,734 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:47:34,361 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2232, 3.7845, 3.9780, 3.9336, 3.7619, 3.6405, 4.3242, 1.5063], + device='cuda:0'), covar=tensor([0.1034, 0.1456, 0.1357, 0.1632, 0.1995, 0.2090, 0.1046, 0.7299], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0276, 0.0294, 0.0336, 0.0281, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:47:38,997 INFO [finetune.py:976] (0/7) Epoch 17, batch 3050, loss[loss=0.1966, simple_loss=0.2678, pruned_loss=0.06273, over 4801.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.253, pruned_loss=0.05625, over 951699.06 frames. ], batch size: 51, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:47:46,829 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94704.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 20:48:19,143 INFO [finetune.py:976] (0/7) Epoch 17, batch 3100, loss[loss=0.1875, simple_loss=0.2486, pruned_loss=0.06315, over 4727.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2516, pruned_loss=0.05635, over 950875.32 frames. ], batch size: 54, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:48:20,429 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1631, 3.5659, 3.7484, 3.9747, 3.9136, 3.6959, 4.2337, 1.3858], + device='cuda:0'), covar=tensor([0.0805, 0.1022, 0.0920, 0.1009, 0.1328, 0.1564, 0.0731, 0.5726], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0276, 0.0293, 0.0335, 0.0281, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:48:27,684 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.533e+02 1.793e+02 2.269e+02 8.706e+02, threshold=3.585e+02, percent-clipped=3.0 +2023-03-26 20:48:29,748 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 20:48:47,535 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5723, 1.4568, 1.2895, 1.5559, 1.9352, 1.6517, 1.4630, 1.3137], + device='cuda:0'), covar=tensor([0.0342, 0.0345, 0.0610, 0.0308, 0.0181, 0.0584, 0.0355, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0106, 0.0141, 0.0111, 0.0098, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.2781e-05, 8.1947e-05, 1.1153e-04, 8.5737e-05, 7.6541e-05, 7.8824e-05, + 7.2723e-05, 8.2813e-05], device='cuda:0') +2023-03-26 20:49:17,517 INFO [finetune.py:976] (0/7) Epoch 17, batch 3150, loss[loss=0.177, simple_loss=0.245, pruned_loss=0.05451, over 4904.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2494, pruned_loss=0.05591, over 952092.38 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:49:28,562 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.00 vs. limit=5.0 +2023-03-26 20:49:40,249 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 20:49:51,404 INFO [finetune.py:976] (0/7) Epoch 17, batch 3200, loss[loss=0.1863, simple_loss=0.2581, pruned_loss=0.05728, over 4826.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2473, pruned_loss=0.05542, over 953323.13 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:49:55,534 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.463e+02 1.766e+02 2.027e+02 4.168e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 20:49:57,701 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 20:50:16,321 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94879.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:50:16,350 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9729, 1.8975, 1.7552, 1.9797, 2.3824, 2.0284, 2.0207, 1.6867], + device='cuda:0'), covar=tensor([0.1772, 0.1774, 0.1555, 0.1493, 0.1637, 0.1085, 0.2052, 0.1603], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0211, 0.0191, 0.0241, 0.0185, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:50:25,250 INFO [finetune.py:976] (0/7) Epoch 17, batch 3250, loss[loss=0.1835, simple_loss=0.254, pruned_loss=0.05651, over 4148.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2471, pruned_loss=0.05525, over 952450.92 frames. ], batch size: 65, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:50:29,668 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 20:50:35,612 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-26 20:50:48,425 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94927.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:50:58,674 INFO [finetune.py:976] (0/7) Epoch 17, batch 3300, loss[loss=0.1671, simple_loss=0.2243, pruned_loss=0.05495, over 3919.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2512, pruned_loss=0.05663, over 952598.09 frames. ], batch size: 17, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:51:02,382 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.785e+02 2.188e+02 2.532e+02 5.228e+02, threshold=4.375e+02, percent-clipped=4.0 +2023-03-26 20:51:28,034 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 20:51:32,705 INFO [finetune.py:976] (0/7) Epoch 17, batch 3350, loss[loss=0.1993, simple_loss=0.2493, pruned_loss=0.07467, over 4795.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2535, pruned_loss=0.0573, over 953750.97 frames. ], batch size: 25, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:51:36,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8111, 1.7805, 1.6224, 1.7898, 1.3083, 4.2832, 1.5967, 1.8078], + device='cuda:0'), covar=tensor([0.3146, 0.2436, 0.2111, 0.2344, 0.1714, 0.0130, 0.2443, 0.1330], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0119, 0.0122, 0.0113, 0.0096, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:51:39,673 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:51:44,407 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7453, 1.3427, 0.8644, 1.7056, 2.0951, 1.4139, 1.4623, 1.7808], + device='cuda:0'), covar=tensor([0.1448, 0.1977, 0.1940, 0.1193, 0.1819, 0.2061, 0.1354, 0.1714], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0095, 0.0110, 0.0093, 0.0119, 0.0095, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:51:46,139 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5032, 1.3826, 1.4894, 0.9159, 1.4810, 1.4839, 1.4698, 1.3088], + device='cuda:0'), covar=tensor([0.0608, 0.0843, 0.0755, 0.0959, 0.0843, 0.0810, 0.0700, 0.1307], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0122, 0.0123, 0.0140, 0.0141, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:52:06,494 INFO [finetune.py:976] (0/7) Epoch 17, batch 3400, loss[loss=0.1853, simple_loss=0.2641, pruned_loss=0.05324, over 4822.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2537, pruned_loss=0.05711, over 954260.83 frames. ], batch size: 25, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:52:10,134 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.553e+02 1.863e+02 2.086e+02 3.757e+02, threshold=3.727e+02, percent-clipped=0.0 +2023-03-26 20:52:12,033 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:52:40,298 INFO [finetune.py:976] (0/7) Epoch 17, batch 3450, loss[loss=0.1994, simple_loss=0.2673, pruned_loss=0.06574, over 4878.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2533, pruned_loss=0.05677, over 954171.66 frames. ], batch size: 35, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:52:47,678 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95105.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:52:48,250 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1384, 1.6713, 1.0355, 2.1279, 2.3311, 2.0049, 1.7906, 2.1201], + device='cuda:0'), covar=tensor([0.1231, 0.1649, 0.1879, 0.1022, 0.1642, 0.1894, 0.1196, 0.1530], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0093, 0.0119, 0.0095, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:52:57,089 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8204, 1.6998, 1.6599, 1.6939, 1.6048, 3.9202, 1.6605, 1.8663], + device='cuda:0'), covar=tensor([0.3139, 0.2262, 0.2003, 0.2209, 0.1434, 0.0163, 0.2525, 0.1272], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0119, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:53:02,391 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6835, 1.3340, 0.8328, 1.6541, 2.0209, 1.4106, 1.4431, 1.7537], + device='cuda:0'), covar=tensor([0.1371, 0.1914, 0.1904, 0.1155, 0.1881, 0.1946, 0.1407, 0.1642], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0095, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:53:13,090 INFO [finetune.py:976] (0/7) Epoch 17, batch 3500, loss[loss=0.1971, simple_loss=0.2555, pruned_loss=0.06934, over 4185.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2514, pruned_loss=0.05664, over 950981.52 frames. ], batch size: 65, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:53:17,182 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.619e+02 1.940e+02 2.279e+02 3.817e+02, threshold=3.880e+02, percent-clipped=1.0 +2023-03-26 20:53:26,806 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3927, 1.3337, 1.4311, 0.7857, 1.4463, 1.4318, 1.3817, 1.3044], + device='cuda:0'), covar=tensor([0.0536, 0.0754, 0.0630, 0.0845, 0.0851, 0.0614, 0.0583, 0.1122], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0134, 0.0140, 0.0122, 0.0122, 0.0140, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:53:35,214 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95166.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:54:05,084 INFO [finetune.py:976] (0/7) Epoch 17, batch 3550, loss[loss=0.1641, simple_loss=0.23, pruned_loss=0.04913, over 4739.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05577, over 950393.95 frames. ], batch size: 54, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:54:51,062 INFO [finetune.py:976] (0/7) Epoch 17, batch 3600, loss[loss=0.1661, simple_loss=0.2247, pruned_loss=0.05377, over 4740.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2472, pruned_loss=0.05576, over 950239.11 frames. ], batch size: 23, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:54:54,641 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.580e+02 1.871e+02 2.182e+02 4.206e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-26 20:55:01,987 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:06,898 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:07,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9834, 1.9760, 1.6191, 1.7321, 1.9725, 1.6750, 2.2110, 1.9998], + device='cuda:0'), covar=tensor([0.1315, 0.1858, 0.2802, 0.2479, 0.2501, 0.1614, 0.2859, 0.1654], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0188, 0.0235, 0.0253, 0.0246, 0.0203, 0.0213, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:55:18,014 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7613, 1.6516, 1.5391, 1.8635, 2.2854, 1.8741, 1.4787, 1.3939], + device='cuda:0'), covar=tensor([0.2179, 0.2083, 0.1910, 0.1690, 0.1564, 0.1308, 0.2480, 0.1954], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0209, 0.0212, 0.0191, 0.0241, 0.0186, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:55:24,757 INFO [finetune.py:976] (0/7) Epoch 17, batch 3650, loss[loss=0.1234, simple_loss=0.1942, pruned_loss=0.02627, over 4723.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2493, pruned_loss=0.05644, over 952670.18 frames. ], batch size: 23, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:55:42,967 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:48,244 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:55,818 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95339.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:58,602 INFO [finetune.py:976] (0/7) Epoch 17, batch 3700, loss[loss=0.2218, simple_loss=0.2925, pruned_loss=0.07552, over 4751.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2524, pruned_loss=0.05697, over 951697.98 frames. ], batch size: 54, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:56:02,234 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.605e+02 1.907e+02 2.409e+02 3.957e+02, threshold=3.813e+02, percent-clipped=4.0 +2023-03-26 20:56:31,735 INFO [finetune.py:976] (0/7) Epoch 17, batch 3750, loss[loss=0.179, simple_loss=0.2651, pruned_loss=0.04648, over 4841.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.254, pruned_loss=0.05734, over 951643.71 frames. ], batch size: 49, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:56:36,717 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:56:56,516 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1730, 1.7396, 2.1027, 2.0965, 1.7807, 1.8472, 2.0053, 1.9399], + device='cuda:0'), covar=tensor([0.4196, 0.4333, 0.3551, 0.4300, 0.5464, 0.4364, 0.5451, 0.3420], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0238, 0.0256, 0.0271, 0.0269, 0.0245, 0.0280, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:57:04,527 INFO [finetune.py:976] (0/7) Epoch 17, batch 3800, loss[loss=0.193, simple_loss=0.2561, pruned_loss=0.065, over 4880.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2547, pruned_loss=0.05758, over 951394.34 frames. ], batch size: 35, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:57:08,652 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7850, 1.6438, 1.5026, 1.6891, 2.0657, 1.9738, 1.6740, 1.5188], + device='cuda:0'), covar=tensor([0.0265, 0.0329, 0.0524, 0.0300, 0.0190, 0.0442, 0.0313, 0.0368], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0108, 0.0143, 0.0113, 0.0099, 0.0109, 0.0098, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.3625e-05, 8.3063e-05, 1.1277e-04, 8.6846e-05, 7.6910e-05, 8.0391e-05, + 7.3540e-05, 8.3921e-05], device='cuda:0') +2023-03-26 20:57:09,547 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.488e+02 1.737e+02 2.235e+02 4.648e+02, threshold=3.475e+02, percent-clipped=3.0 +2023-03-26 20:57:16,908 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95461.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:57:37,562 INFO [finetune.py:976] (0/7) Epoch 17, batch 3850, loss[loss=0.1658, simple_loss=0.2421, pruned_loss=0.04469, over 4834.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.254, pruned_loss=0.05754, over 951768.18 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:57:43,976 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.26 vs. limit=5.0 +2023-03-26 20:57:46,396 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0624, 1.9788, 1.6899, 1.8053, 2.0422, 1.7227, 2.2289, 2.0275], + device='cuda:0'), covar=tensor([0.1264, 0.1807, 0.2765, 0.2282, 0.2449, 0.1659, 0.2825, 0.1662], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0189, 0.0236, 0.0255, 0.0247, 0.0204, 0.0214, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 20:58:10,766 INFO [finetune.py:976] (0/7) Epoch 17, batch 3900, loss[loss=0.1576, simple_loss=0.2355, pruned_loss=0.03983, over 4862.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2532, pruned_loss=0.05838, over 952451.30 frames. ], batch size: 31, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:58:15,378 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.550e+02 1.834e+02 2.229e+02 4.290e+02, threshold=3.669e+02, percent-clipped=3.0 +2023-03-26 20:58:16,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6412, 1.5398, 1.5103, 1.5565, 0.9231, 3.2542, 1.2166, 1.6031], + device='cuda:0'), covar=tensor([0.3225, 0.2439, 0.2166, 0.2357, 0.1973, 0.0219, 0.2826, 0.1355], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 20:58:46,336 INFO [finetune.py:976] (0/7) Epoch 17, batch 3950, loss[loss=0.1685, simple_loss=0.2362, pruned_loss=0.05034, over 4911.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2495, pruned_loss=0.05683, over 952140.37 frames. ], batch size: 37, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:59:04,356 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:05,018 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95617.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:13,743 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:34,007 INFO [finetune.py:976] (0/7) Epoch 17, batch 4000, loss[loss=0.2148, simple_loss=0.2744, pruned_loss=0.07762, over 4711.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2488, pruned_loss=0.05702, over 952909.99 frames. ], batch size: 59, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:59:42,361 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.540e+02 1.979e+02 2.285e+02 3.877e+02, threshold=3.958e+02, percent-clipped=2.0 +2023-03-26 20:59:44,296 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4266, 1.6398, 0.7674, 2.1734, 2.5833, 1.8817, 2.0015, 2.1615], + device='cuda:0'), covar=tensor([0.1311, 0.2099, 0.2161, 0.1153, 0.1702, 0.1802, 0.1411, 0.1828], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0095, 0.0110, 0.0092, 0.0119, 0.0095, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 20:59:46,668 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5205, 0.7840, 1.6054, 1.4577, 1.3870, 1.2953, 1.4795, 1.4706], + device='cuda:0'), covar=tensor([0.2695, 0.2857, 0.2234, 0.2621, 0.3136, 0.2820, 0.2869, 0.2155], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0238, 0.0257, 0.0272, 0.0270, 0.0246, 0.0281, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:00:12,621 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95678.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:00:26,082 INFO [finetune.py:976] (0/7) Epoch 17, batch 4050, loss[loss=0.2073, simple_loss=0.288, pruned_loss=0.06325, over 4921.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2516, pruned_loss=0.05769, over 951985.44 frames. ], batch size: 36, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:00:27,353 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95695.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:00:37,385 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9026, 1.8617, 1.5362, 1.8713, 1.6806, 1.6965, 1.7627, 2.4649], + device='cuda:0'), covar=tensor([0.3976, 0.4141, 0.3357, 0.3838, 0.4134, 0.2466, 0.3829, 0.1720], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0227, 0.0277, 0.0251, 0.0219, 0.0251, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:00:59,870 INFO [finetune.py:976] (0/7) Epoch 17, batch 4100, loss[loss=0.1495, simple_loss=0.2137, pruned_loss=0.04268, over 4740.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2549, pruned_loss=0.05877, over 953450.58 frames. ], batch size: 23, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:01:04,071 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.600e+02 1.864e+02 2.304e+02 4.240e+02, threshold=3.729e+02, percent-clipped=2.0 +2023-03-26 21:01:12,307 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95761.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:01:27,355 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-26 21:01:31,971 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5516, 1.3841, 1.2530, 1.5587, 1.5844, 1.6161, 0.9318, 1.3199], + device='cuda:0'), covar=tensor([0.2253, 0.2255, 0.2087, 0.1735, 0.1691, 0.1352, 0.2726, 0.2012], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0191, 0.0241, 0.0186, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:01:33,030 INFO [finetune.py:976] (0/7) Epoch 17, batch 4150, loss[loss=0.1568, simple_loss=0.2307, pruned_loss=0.04151, over 4731.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2564, pruned_loss=0.05968, over 951803.01 frames. ], batch size: 27, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:01:44,397 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95809.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:01:44,469 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5096, 2.4351, 2.0317, 2.4798, 2.4171, 2.0521, 2.9288, 2.5541], + device='cuda:0'), covar=tensor([0.1225, 0.2234, 0.2870, 0.2847, 0.2474, 0.1623, 0.3249, 0.1550], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0191, 0.0238, 0.0257, 0.0248, 0.0205, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:01:46,227 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95812.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:02:06,758 INFO [finetune.py:976] (0/7) Epoch 17, batch 4200, loss[loss=0.2063, simple_loss=0.2692, pruned_loss=0.07169, over 4832.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2571, pruned_loss=0.0592, over 954038.93 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:02:09,303 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95847.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:02:11,507 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.538e+02 1.932e+02 2.354e+02 8.206e+02, threshold=3.863e+02, percent-clipped=2.0 +2023-03-26 21:02:14,047 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9511, 4.5060, 4.2407, 2.4927, 4.6477, 3.6742, 1.1317, 3.1943], + device='cuda:0'), covar=tensor([0.2127, 0.1591, 0.1384, 0.2987, 0.0667, 0.0722, 0.4212, 0.1289], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0175, 0.0159, 0.0129, 0.0158, 0.0123, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 21:02:27,865 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95873.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:02:39,923 INFO [finetune.py:976] (0/7) Epoch 17, batch 4250, loss[loss=0.1584, simple_loss=0.2304, pruned_loss=0.04321, over 4795.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2551, pruned_loss=0.05831, over 954693.21 frames. ], batch size: 29, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:02:50,079 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95908.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:02:55,835 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:02:57,066 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0460, 1.9860, 1.9844, 1.4257, 1.9957, 2.0811, 2.0531, 1.7116], + device='cuda:0'), covar=tensor([0.0539, 0.0574, 0.0704, 0.0907, 0.0742, 0.0685, 0.0600, 0.1059], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0136, 0.0142, 0.0125, 0.0125, 0.0142, 0.0144, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:03:02,169 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:13,535 INFO [finetune.py:976] (0/7) Epoch 17, batch 4300, loss[loss=0.1352, simple_loss=0.2116, pruned_loss=0.02943, over 4825.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.252, pruned_loss=0.05734, over 955708.11 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:03:18,258 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.465e+02 1.654e+02 2.123e+02 3.225e+02, threshold=3.308e+02, percent-clipped=0.0 +2023-03-26 21:03:27,286 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:33,594 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:34,211 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95973.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:39,473 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:47,255 INFO [finetune.py:976] (0/7) Epoch 17, batch 4350, loss[loss=0.1434, simple_loss=0.2114, pruned_loss=0.03776, over 4776.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2485, pruned_loss=0.05608, over 954640.61 frames. ], batch size: 28, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:03:48,556 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:48,658 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 21:03:51,661 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-96000.pt +2023-03-26 21:04:20,839 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:04:21,943 INFO [finetune.py:976] (0/7) Epoch 17, batch 4400, loss[loss=0.2018, simple_loss=0.2566, pruned_loss=0.07351, over 4820.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2488, pruned_loss=0.05599, over 954467.84 frames. ], batch size: 30, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:04:22,001 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96043.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:04:28,704 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.398e+01 1.490e+02 1.749e+02 2.200e+02 3.209e+02, threshold=3.497e+02, percent-clipped=0.0 +2023-03-26 21:04:47,915 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7834, 1.7238, 1.6899, 1.7591, 1.4248, 3.3274, 1.7272, 1.9890], + device='cuda:0'), covar=tensor([0.2940, 0.2236, 0.1934, 0.2023, 0.1602, 0.0270, 0.2293, 0.1124], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0114, 0.0119, 0.0122, 0.0112, 0.0095, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:04:59,193 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96080.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:05:13,436 INFO [finetune.py:976] (0/7) Epoch 17, batch 4450, loss[loss=0.2057, simple_loss=0.279, pruned_loss=0.06616, over 4887.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2523, pruned_loss=0.05725, over 955282.10 frames. ], batch size: 32, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:05:44,682 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-26 21:05:58,469 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96141.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:05:59,553 INFO [finetune.py:976] (0/7) Epoch 17, batch 4500, loss[loss=0.1943, simple_loss=0.278, pruned_loss=0.05527, over 4838.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.254, pruned_loss=0.05752, over 955382.30 frames. ], batch size: 49, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:06:03,840 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.724e+02 1.946e+02 2.358e+02 4.504e+02, threshold=3.891e+02, percent-clipped=3.0 +2023-03-26 21:06:08,321 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 21:06:13,996 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:06:15,762 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96168.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:06:33,241 INFO [finetune.py:976] (0/7) Epoch 17, batch 4550, loss[loss=0.2282, simple_loss=0.2875, pruned_loss=0.08451, over 4903.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2553, pruned_loss=0.05754, over 956000.78 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:06:39,478 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96203.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:06:54,572 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96226.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:07:05,721 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 21:07:07,173 INFO [finetune.py:976] (0/7) Epoch 17, batch 4600, loss[loss=0.1655, simple_loss=0.2382, pruned_loss=0.04636, over 4810.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2539, pruned_loss=0.05654, over 955376.95 frames. ], batch size: 41, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:07:11,422 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.530e+02 1.886e+02 2.340e+02 4.335e+02, threshold=3.772e+02, percent-clipped=2.0 +2023-03-26 21:07:14,559 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2909, 2.9319, 3.0550, 3.2147, 3.0698, 2.9001, 3.3267, 0.9520], + device='cuda:0'), covar=tensor([0.1165, 0.1040, 0.1131, 0.1157, 0.1615, 0.1722, 0.1173, 0.5473], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0245, 0.0276, 0.0292, 0.0335, 0.0282, 0.0302, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:07:26,466 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:07:40,044 INFO [finetune.py:976] (0/7) Epoch 17, batch 4650, loss[loss=0.1523, simple_loss=0.2288, pruned_loss=0.03787, over 4840.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2519, pruned_loss=0.05619, over 954842.78 frames. ], batch size: 49, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:07:58,451 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:00,267 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:06,367 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8561, 1.7865, 1.5656, 2.0630, 2.3886, 2.0301, 1.6857, 1.5178], + device='cuda:0'), covar=tensor([0.2261, 0.2130, 0.2035, 0.1694, 0.1789, 0.1222, 0.2552, 0.2029], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0192, 0.0242, 0.0186, 0.0215, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:08:08,004 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:13,200 INFO [finetune.py:976] (0/7) Epoch 17, batch 4700, loss[loss=0.1779, simple_loss=0.2409, pruned_loss=0.05746, over 4937.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2486, pruned_loss=0.055, over 957077.85 frames. ], batch size: 38, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:08:18,318 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.531e+02 1.882e+02 2.216e+02 4.319e+02, threshold=3.764e+02, percent-clipped=2.0 +2023-03-26 21:08:26,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7431, 2.8117, 2.4961, 2.0163, 2.7888, 2.9795, 2.9854, 2.5286], + device='cuda:0'), covar=tensor([0.0605, 0.0557, 0.0842, 0.0886, 0.0545, 0.0702, 0.0584, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0135, 0.0141, 0.0123, 0.0124, 0.0141, 0.0142, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:08:40,429 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:43,973 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96390.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:45,651 INFO [finetune.py:976] (0/7) Epoch 17, batch 4750, loss[loss=0.1916, simple_loss=0.2657, pruned_loss=0.05877, over 4772.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2469, pruned_loss=0.05461, over 955514.11 frames. ], batch size: 28, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:09:14,755 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:09:16,089 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-26 21:09:19,420 INFO [finetune.py:976] (0/7) Epoch 17, batch 4800, loss[loss=0.2249, simple_loss=0.3033, pruned_loss=0.07322, over 4813.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2494, pruned_loss=0.05571, over 952903.96 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:09:25,021 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.609e+02 1.875e+02 2.422e+02 6.864e+02, threshold=3.750e+02, percent-clipped=2.0 +2023-03-26 21:09:25,754 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96451.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:09:29,371 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7483, 1.6912, 1.3670, 1.8238, 2.2967, 1.8310, 1.5371, 1.3703], + device='cuda:0'), covar=tensor([0.2095, 0.1899, 0.1936, 0.1583, 0.1552, 0.1161, 0.2379, 0.1835], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0212, 0.0191, 0.0241, 0.0186, 0.0214, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:09:36,646 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96468.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:09:46,908 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 21:09:54,757 INFO [finetune.py:976] (0/7) Epoch 17, batch 4850, loss[loss=0.1797, simple_loss=0.2631, pruned_loss=0.04814, over 4813.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2519, pruned_loss=0.05655, over 952135.78 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:10:02,900 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96503.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:10:06,411 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4484, 1.1264, 0.7671, 1.3456, 1.8165, 0.7791, 1.2940, 1.4479], + device='cuda:0'), covar=tensor([0.1409, 0.1930, 0.1718, 0.1151, 0.1914, 0.2055, 0.1347, 0.1707], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0109, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 21:10:15,683 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96516.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:10:15,738 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.5779, 3.1807, 2.8331, 1.5459, 3.0290, 2.6085, 2.5360, 2.7669], + device='cuda:0'), covar=tensor([0.0686, 0.0700, 0.1194, 0.1830, 0.1354, 0.1778, 0.1671, 0.0946], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0197, 0.0200, 0.0185, 0.0214, 0.0209, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:10:18,741 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:10:38,709 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1146, 2.0521, 1.6736, 1.9910, 1.8424, 1.9055, 1.9524, 2.6884], + device='cuda:0'), covar=tensor([0.3499, 0.4258, 0.3107, 0.3919, 0.4132, 0.2271, 0.3795, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0261, 0.0227, 0.0276, 0.0251, 0.0219, 0.0251, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:10:45,275 INFO [finetune.py:976] (0/7) Epoch 17, batch 4900, loss[loss=0.1689, simple_loss=0.2321, pruned_loss=0.05288, over 4150.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2533, pruned_loss=0.05713, over 952149.26 frames. ], batch size: 17, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:10:54,622 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.592e+02 1.896e+02 2.164e+02 3.347e+02, threshold=3.792e+02, percent-clipped=0.0 +2023-03-26 21:10:55,803 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:11:10,542 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5471, 1.4309, 2.1559, 3.1248, 2.0457, 2.2825, 1.0681, 2.6065], + device='cuda:0'), covar=tensor([0.1769, 0.1546, 0.1262, 0.0594, 0.0844, 0.1471, 0.1833, 0.0505], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0166, 0.0102, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 21:11:26,124 INFO [finetune.py:976] (0/7) Epoch 17, batch 4950, loss[loss=0.175, simple_loss=0.2568, pruned_loss=0.04663, over 4904.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2555, pruned_loss=0.05773, over 952386.61 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:11:55,610 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96636.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:11:59,759 INFO [finetune.py:976] (0/7) Epoch 17, batch 5000, loss[loss=0.1906, simple_loss=0.2521, pruned_loss=0.06453, over 4906.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2528, pruned_loss=0.05671, over 953253.13 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:12:04,407 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.531e+02 1.819e+02 2.156e+02 3.437e+02, threshold=3.638e+02, percent-clipped=0.0 +2023-03-26 21:12:24,600 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:12:26,957 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:12:33,328 INFO [finetune.py:976] (0/7) Epoch 17, batch 5050, loss[loss=0.1967, simple_loss=0.2657, pruned_loss=0.06388, over 4820.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.25, pruned_loss=0.05581, over 951661.99 frames. ], batch size: 39, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:12:48,820 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 21:13:01,826 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96736.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:06,443 INFO [finetune.py:976] (0/7) Epoch 17, batch 5100, loss[loss=0.2051, simple_loss=0.2589, pruned_loss=0.07571, over 4920.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2481, pruned_loss=0.05534, over 952628.25 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:13:08,316 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96746.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:10,588 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.513e+02 1.848e+02 2.246e+02 3.685e+02, threshold=3.695e+02, percent-clipped=1.0 +2023-03-26 21:13:28,941 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:33,645 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:39,088 INFO [finetune.py:976] (0/7) Epoch 17, batch 5150, loss[loss=0.1704, simple_loss=0.222, pruned_loss=0.05938, over 4383.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.247, pruned_loss=0.05517, over 951984.95 frames. ], batch size: 19, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:13:42,712 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2244, 1.4300, 1.5910, 1.0765, 1.3287, 1.4833, 1.3594, 1.5902], + device='cuda:0'), covar=tensor([0.1299, 0.2037, 0.1203, 0.1588, 0.0914, 0.1375, 0.2831, 0.0841], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0202, 0.0188, 0.0188, 0.0175, 0.0211, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:13:58,204 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:05,908 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4462, 2.2895, 2.8652, 1.7494, 2.6559, 2.7955, 2.0022, 2.7335], + device='cuda:0'), covar=tensor([0.1243, 0.1570, 0.1325, 0.1998, 0.0810, 0.1519, 0.2369, 0.0903], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0203, 0.0189, 0.0189, 0.0176, 0.0212, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:14:08,361 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96837.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:11,885 INFO [finetune.py:976] (0/7) Epoch 17, batch 5200, loss[loss=0.2151, simple_loss=0.2811, pruned_loss=0.07453, over 4769.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2499, pruned_loss=0.05617, over 953971.05 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:14:16,597 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.635e+02 1.977e+02 2.300e+02 5.939e+02, threshold=3.955e+02, percent-clipped=5.0 +2023-03-26 21:14:29,993 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96869.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:44,941 INFO [finetune.py:976] (0/7) Epoch 17, batch 5250, loss[loss=0.2143, simple_loss=0.2822, pruned_loss=0.07321, over 4893.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2512, pruned_loss=0.05651, over 953608.62 frames. ], batch size: 43, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:15:04,270 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 21:15:21,072 INFO [finetune.py:976] (0/7) Epoch 17, batch 5300, loss[loss=0.1903, simple_loss=0.2672, pruned_loss=0.0567, over 4814.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2534, pruned_loss=0.05711, over 953734.67 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:15:30,005 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.644e+02 1.959e+02 2.379e+02 3.599e+02, threshold=3.918e+02, percent-clipped=0.0 +2023-03-26 21:15:31,511 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 21:16:00,344 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:16,631 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 21:16:17,650 INFO [finetune.py:976] (0/7) Epoch 17, batch 5350, loss[loss=0.2138, simple_loss=0.2709, pruned_loss=0.07829, over 4920.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2542, pruned_loss=0.05722, over 955833.11 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:16:19,671 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9892, 1.2553, 1.9089, 1.9167, 1.7124, 1.6926, 1.8486, 1.7444], + device='cuda:0'), covar=tensor([0.3538, 0.3710, 0.2989, 0.3279, 0.4285, 0.3564, 0.4030, 0.3058], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0241, 0.0259, 0.0275, 0.0273, 0.0248, 0.0283, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:16:25,309 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97000.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:44,613 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:54,581 INFO [finetune.py:976] (0/7) Epoch 17, batch 5400, loss[loss=0.1374, simple_loss=0.2017, pruned_loss=0.03655, over 4716.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2529, pruned_loss=0.05743, over 954195.28 frames. ], batch size: 23, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:16:56,506 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:58,803 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.591e+02 1.860e+02 2.348e+02 4.043e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 21:17:06,253 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97061.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:17:11,086 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3849, 2.6353, 2.2834, 1.8935, 2.6517, 2.7297, 2.7892, 2.4147], + device='cuda:0'), covar=tensor([0.0655, 0.0586, 0.0795, 0.0876, 0.0585, 0.0747, 0.0588, 0.0918], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0122, 0.0124, 0.0139, 0.0141, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:17:27,400 INFO [finetune.py:976] (0/7) Epoch 17, batch 5450, loss[loss=0.2273, simple_loss=0.269, pruned_loss=0.09282, over 4198.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2492, pruned_loss=0.05637, over 954455.14 frames. ], batch size: 65, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:17:28,067 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97094.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:17:52,424 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:18:00,547 INFO [finetune.py:976] (0/7) Epoch 17, batch 5500, loss[loss=0.1942, simple_loss=0.271, pruned_loss=0.05866, over 4901.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2458, pruned_loss=0.05527, over 954318.13 frames. ], batch size: 35, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:18:04,744 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.221e+01 1.533e+02 1.769e+02 2.042e+02 3.202e+02, threshold=3.539e+02, percent-clipped=0.0 +2023-03-26 21:18:06,128 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4152, 1.5779, 2.2342, 1.6092, 1.7504, 3.8258, 1.5056, 1.7211], + device='cuda:0'), covar=tensor([0.0985, 0.1579, 0.1169, 0.0963, 0.1430, 0.0213, 0.1363, 0.1649], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:18:33,718 INFO [finetune.py:976] (0/7) Epoch 17, batch 5550, loss[loss=0.2077, simple_loss=0.2773, pruned_loss=0.06904, over 4894.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2468, pruned_loss=0.05508, over 951368.61 frames. ], batch size: 37, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:05,113 INFO [finetune.py:976] (0/7) Epoch 17, batch 5600, loss[loss=0.191, simple_loss=0.2648, pruned_loss=0.05857, over 4776.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2504, pruned_loss=0.05631, over 952361.08 frames. ], batch size: 26, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:09,079 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.641e+02 1.914e+02 2.406e+02 4.422e+02, threshold=3.827e+02, percent-clipped=1.0 +2023-03-26 21:19:14,879 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:19:16,618 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3108, 2.2508, 2.8010, 1.6731, 2.5757, 2.6294, 1.9679, 2.7176], + device='cuda:0'), covar=tensor([0.1447, 0.1714, 0.1535, 0.2226, 0.1029, 0.1665, 0.2780, 0.0829], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0202, 0.0188, 0.0189, 0.0175, 0.0212, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:19:20,214 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 21:19:34,291 INFO [finetune.py:976] (0/7) Epoch 17, batch 5650, loss[loss=0.1832, simple_loss=0.251, pruned_loss=0.05774, over 4891.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2534, pruned_loss=0.05714, over 952833.64 frames. ], batch size: 32, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:51,445 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:04,078 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97342.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:04,601 INFO [finetune.py:976] (0/7) Epoch 17, batch 5700, loss[loss=0.1534, simple_loss=0.2116, pruned_loss=0.04756, over 4217.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2487, pruned_loss=0.05606, over 934678.64 frames. ], batch size: 18, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:20:07,045 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97347.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:08,735 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.528e+01 1.533e+02 1.739e+02 2.212e+02 3.283e+02, threshold=3.478e+02, percent-clipped=0.0 +2023-03-26 21:20:12,306 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97356.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:15,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5056, 2.3786, 2.1670, 2.3612, 2.2978, 2.2743, 2.2573, 2.9817], + device='cuda:0'), covar=tensor([0.3613, 0.4496, 0.3182, 0.3876, 0.4079, 0.2483, 0.4120, 0.1575], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0228, 0.0276, 0.0252, 0.0219, 0.0252, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:20:21,443 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-17.pt +2023-03-26 21:20:35,933 INFO [finetune.py:976] (0/7) Epoch 18, batch 0, loss[loss=0.203, simple_loss=0.2791, pruned_loss=0.06342, over 4878.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2791, pruned_loss=0.06342, over 4878.00 frames. ], batch size: 35, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:20:35,934 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 21:20:46,783 INFO [finetune.py:1010] (0/7) Epoch 18, validation: loss=0.1584, simple_loss=0.2281, pruned_loss=0.0444, over 2265189.00 frames. +2023-03-26 21:20:46,783 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 21:20:49,159 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:21:26,107 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97403.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:21:30,206 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:21:45,570 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 21:21:47,679 INFO [finetune.py:976] (0/7) Epoch 18, batch 50, loss[loss=0.2, simple_loss=0.2817, pruned_loss=0.05913, over 4812.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2557, pruned_loss=0.05665, over 214081.10 frames. ], batch size: 39, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:21:58,971 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:00,274 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 21:22:00,765 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:09,807 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.163e+01 1.563e+02 1.902e+02 2.308e+02 3.615e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-03-26 21:22:25,131 INFO [finetune.py:976] (0/7) Epoch 18, batch 100, loss[loss=0.182, simple_loss=0.2435, pruned_loss=0.0602, over 4748.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2495, pruned_loss=0.05598, over 378431.69 frames. ], batch size: 59, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:22:31,667 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97480.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:58,721 INFO [finetune.py:976] (0/7) Epoch 18, batch 150, loss[loss=0.1819, simple_loss=0.249, pruned_loss=0.05743, over 4762.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2458, pruned_loss=0.05469, over 507769.31 frames. ], batch size: 26, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:23:17,209 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.559e+02 1.792e+02 2.227e+02 6.409e+02, threshold=3.584e+02, percent-clipped=2.0 +2023-03-26 21:23:19,755 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1578, 1.9257, 2.0679, 0.8918, 2.3668, 2.4913, 2.2524, 1.8269], + device='cuda:0'), covar=tensor([0.0884, 0.0794, 0.0555, 0.0736, 0.0486, 0.0678, 0.0442, 0.0818], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0127, 0.0130, 0.0129, 0.0143, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.1278e-05, 1.0890e-04, 8.8327e-05, 9.0232e-05, 9.1370e-05, 9.2552e-05, + 1.0281e-04, 1.0582e-04], device='cuda:0') +2023-03-26 21:23:20,338 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:23:26,110 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1676, 1.3606, 1.4842, 0.7267, 1.4210, 1.6572, 1.6629, 1.3489], + device='cuda:0'), covar=tensor([0.0978, 0.0602, 0.0521, 0.0546, 0.0475, 0.0569, 0.0368, 0.0737], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0127, 0.0130, 0.0129, 0.0143, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.1197e-05, 1.0878e-04, 8.8282e-05, 9.0162e-05, 9.1357e-05, 9.2496e-05, + 1.0278e-04, 1.0576e-04], device='cuda:0') +2023-03-26 21:23:32,346 INFO [finetune.py:976] (0/7) Epoch 18, batch 200, loss[loss=0.2194, simple_loss=0.2769, pruned_loss=0.08094, over 4108.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2455, pruned_loss=0.05581, over 607762.67 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:24:01,867 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:01,912 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:05,307 INFO [finetune.py:976] (0/7) Epoch 18, batch 250, loss[loss=0.1732, simple_loss=0.2512, pruned_loss=0.04756, over 4899.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2489, pruned_loss=0.0566, over 684740.62 frames. ], batch size: 35, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:24:23,171 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 21:24:24,125 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.615e+02 1.960e+02 2.417e+02 4.168e+02, threshold=3.921e+02, percent-clipped=3.0 +2023-03-26 21:24:27,905 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:37,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8188, 1.2636, 1.7080, 1.7935, 1.5737, 1.5917, 1.6912, 1.6550], + device='cuda:0'), covar=tensor([0.4902, 0.4827, 0.4754, 0.4496, 0.6119, 0.4821, 0.5863, 0.4281], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0239, 0.0259, 0.0273, 0.0272, 0.0247, 0.0281, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:24:37,893 INFO [finetune.py:976] (0/7) Epoch 18, batch 300, loss[loss=0.187, simple_loss=0.2661, pruned_loss=0.05393, over 4826.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2533, pruned_loss=0.05842, over 742234.07 frames. ], batch size: 33, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:24:40,276 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97674.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:24:56,233 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97698.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:24:59,277 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97703.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:59,859 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97704.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:25:10,597 INFO [finetune.py:976] (0/7) Epoch 18, batch 350, loss[loss=0.1993, simple_loss=0.274, pruned_loss=0.0623, over 4914.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2547, pruned_loss=0.05895, over 788081.45 frames. ], batch size: 36, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:25:17,532 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:25:20,586 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97735.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:25:30,598 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.611e+02 1.882e+02 2.387e+02 3.928e+02, threshold=3.763e+02, percent-clipped=1.0 +2023-03-26 21:25:39,966 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-26 21:25:43,299 INFO [finetune.py:976] (0/7) Epoch 18, batch 400, loss[loss=0.1701, simple_loss=0.2423, pruned_loss=0.04891, over 4908.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2551, pruned_loss=0.05852, over 826920.19 frames. ], batch size: 36, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:26:22,831 INFO [finetune.py:976] (0/7) Epoch 18, batch 450, loss[loss=0.1807, simple_loss=0.2503, pruned_loss=0.05558, over 4822.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2534, pruned_loss=0.05781, over 853710.86 frames. ], batch size: 41, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:26:33,942 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8800, 1.8613, 1.7015, 1.7955, 1.5863, 4.3722, 1.6266, 2.0758], + device='cuda:0'), covar=tensor([0.3236, 0.2511, 0.2067, 0.2193, 0.1540, 0.0131, 0.2770, 0.1281], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0095, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:27:01,042 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.519e+02 1.772e+02 2.128e+02 3.513e+02, threshold=3.544e+02, percent-clipped=0.0 +2023-03-26 21:27:16,934 INFO [finetune.py:976] (0/7) Epoch 18, batch 500, loss[loss=0.1592, simple_loss=0.2296, pruned_loss=0.04434, over 4091.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2509, pruned_loss=0.05654, over 876425.82 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:27:29,763 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:44,603 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:47,670 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:50,646 INFO [finetune.py:976] (0/7) Epoch 18, batch 550, loss[loss=0.179, simple_loss=0.2468, pruned_loss=0.05562, over 4830.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2477, pruned_loss=0.05557, over 895541.73 frames. ], batch size: 51, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:28:19,277 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:19,731 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.294e+01 1.530e+02 1.816e+02 2.060e+02 3.951e+02, threshold=3.633e+02, percent-clipped=3.0 +2023-03-26 21:28:24,688 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1936, 1.2318, 1.3990, 0.6132, 1.3645, 1.5995, 1.5352, 1.3382], + device='cuda:0'), covar=tensor([0.0761, 0.0604, 0.0450, 0.0465, 0.0385, 0.0435, 0.0330, 0.0513], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0126, 0.0129, 0.0129, 0.0142, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.1192e-05, 1.0881e-04, 8.8000e-05, 8.9867e-05, 9.1022e-05, 9.2420e-05, + 1.0220e-04, 1.0578e-04], device='cuda:0') +2023-03-26 21:28:28,256 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:32,536 INFO [finetune.py:976] (0/7) Epoch 18, batch 600, loss[loss=0.1912, simple_loss=0.2708, pruned_loss=0.05586, over 4930.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2467, pruned_loss=0.05478, over 911074.12 frames. ], batch size: 38, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:28:51,972 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97998.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:53,725 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-98000.pt +2023-03-26 21:28:56,785 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98003.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:07,659 INFO [finetune.py:976] (0/7) Epoch 18, batch 650, loss[loss=0.2447, simple_loss=0.3055, pruned_loss=0.09201, over 4841.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2491, pruned_loss=0.05511, over 921975.41 frames. ], batch size: 47, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:29:13,276 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98030.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:29:13,308 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:25,503 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:28,301 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.543e+02 1.878e+02 2.128e+02 3.672e+02, threshold=3.757e+02, percent-clipped=1.0 +2023-03-26 21:29:28,989 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:41,536 INFO [finetune.py:976] (0/7) Epoch 18, batch 700, loss[loss=0.1495, simple_loss=0.2186, pruned_loss=0.04022, over 4678.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2514, pruned_loss=0.05635, over 929055.33 frames. ], batch size: 23, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:29:45,864 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:05,715 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:11,004 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9438, 4.0859, 3.7976, 2.2987, 4.1554, 3.1420, 0.9919, 2.9989], + device='cuda:0'), covar=tensor([0.2182, 0.1912, 0.1638, 0.3027, 0.1071, 0.0930, 0.4652, 0.1387], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0177, 0.0159, 0.0128, 0.0160, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 21:30:14,143 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7995, 1.7982, 2.0391, 1.1446, 2.0060, 2.2438, 2.1475, 1.7240], + device='cuda:0'), covar=tensor([0.0751, 0.0635, 0.0455, 0.0529, 0.0407, 0.0576, 0.0339, 0.0603], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0127, 0.0130, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.1361e-05, 1.0967e-04, 8.8496e-05, 9.0323e-05, 9.1903e-05, 9.2905e-05, + 1.0280e-04, 1.0652e-04], device='cuda:0') +2023-03-26 21:30:15,239 INFO [finetune.py:976] (0/7) Epoch 18, batch 750, loss[loss=0.1896, simple_loss=0.2555, pruned_loss=0.06188, over 4800.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2518, pruned_loss=0.05603, over 931072.77 frames. ], batch size: 25, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:30:34,724 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.501e+02 1.785e+02 2.309e+02 4.193e+02, threshold=3.569e+02, percent-clipped=2.0 +2023-03-26 21:30:46,060 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:48,361 INFO [finetune.py:976] (0/7) Epoch 18, batch 800, loss[loss=0.1599, simple_loss=0.2366, pruned_loss=0.04159, over 4787.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2529, pruned_loss=0.0563, over 936615.39 frames. ], batch size: 25, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:31:15,630 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:31:22,189 INFO [finetune.py:976] (0/7) Epoch 18, batch 850, loss[loss=0.1499, simple_loss=0.224, pruned_loss=0.03797, over 4771.00 frames. ], tot_loss[loss=0.18, simple_loss=0.25, pruned_loss=0.05498, over 939079.38 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:31:45,827 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:31:55,530 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.511e+02 1.794e+02 2.111e+02 3.360e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-26 21:32:06,589 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:32:15,532 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2982, 2.3424, 2.0364, 2.4354, 2.8521, 2.2991, 2.2436, 1.7874], + device='cuda:0'), covar=tensor([0.2079, 0.1762, 0.1815, 0.1516, 0.1578, 0.1103, 0.1875, 0.1816], + device='cuda:0'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0191, 0.0242, 0.0187, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:32:25,141 INFO [finetune.py:976] (0/7) Epoch 18, batch 900, loss[loss=0.1876, simple_loss=0.2513, pruned_loss=0.06193, over 4865.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2477, pruned_loss=0.05423, over 944622.27 frames. ], batch size: 34, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:33:02,861 INFO [finetune.py:976] (0/7) Epoch 18, batch 950, loss[loss=0.1728, simple_loss=0.2404, pruned_loss=0.05257, over 4829.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2468, pruned_loss=0.05465, over 946155.90 frames. ], batch size: 33, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:33:08,450 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98330.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:33:23,086 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.511e+02 1.760e+02 2.160e+02 3.441e+02, threshold=3.521e+02, percent-clipped=0.0 +2023-03-26 21:33:37,315 INFO [finetune.py:976] (0/7) Epoch 18, batch 1000, loss[loss=0.144, simple_loss=0.2243, pruned_loss=0.03179, over 4834.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2481, pruned_loss=0.05505, over 947851.38 frames. ], batch size: 25, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:33:42,165 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98378.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:34:02,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98410.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:34:06,236 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5349, 1.6003, 2.0369, 1.7841, 1.7015, 3.5911, 1.6262, 1.7052], + device='cuda:0'), covar=tensor([0.0984, 0.1858, 0.1094, 0.1011, 0.1539, 0.0215, 0.1390, 0.1711], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:34:14,800 INFO [finetune.py:976] (0/7) Epoch 18, batch 1050, loss[loss=0.1882, simple_loss=0.2675, pruned_loss=0.05445, over 4927.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2517, pruned_loss=0.05588, over 949901.40 frames. ], batch size: 33, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:34:26,450 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9967, 1.8261, 1.5244, 1.5871, 1.7255, 1.7329, 1.7860, 2.4723], + device='cuda:0'), covar=tensor([0.4031, 0.4228, 0.3547, 0.3922, 0.4203, 0.2742, 0.4038, 0.1848], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0262, 0.0227, 0.0275, 0.0252, 0.0220, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:34:36,631 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.558e+02 1.856e+02 2.287e+02 5.753e+02, threshold=3.713e+02, percent-clipped=4.0 +2023-03-26 21:34:44,673 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98462.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:34:51,516 INFO [finetune.py:976] (0/7) Epoch 18, batch 1100, loss[loss=0.2375, simple_loss=0.2938, pruned_loss=0.09062, over 4816.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2541, pruned_loss=0.0569, over 952112.00 frames. ], batch size: 38, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:34:51,648 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:08,486 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6034, 1.6959, 1.3737, 1.5889, 2.0611, 2.0113, 1.6246, 1.5707], + device='cuda:0'), covar=tensor([0.0384, 0.0328, 0.0580, 0.0366, 0.0249, 0.0494, 0.0339, 0.0384], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0108, 0.0144, 0.0113, 0.0101, 0.0109, 0.0099, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.5148e-05, 8.3451e-05, 1.1378e-04, 8.6666e-05, 7.8786e-05, 8.0849e-05, + 7.4099e-05, 8.4446e-05], device='cuda:0') +2023-03-26 21:35:14,499 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9971, 1.4589, 1.9635, 1.9903, 1.7682, 1.6950, 1.9297, 1.8041], + device='cuda:0'), covar=tensor([0.3841, 0.3737, 0.3447, 0.3487, 0.5012, 0.3718, 0.4251, 0.3194], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0274, 0.0273, 0.0248, 0.0283, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:35:24,231 INFO [finetune.py:976] (0/7) Epoch 18, batch 1150, loss[loss=0.1463, simple_loss=0.2251, pruned_loss=0.03376, over 4849.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.254, pruned_loss=0.05647, over 952834.84 frames. ], batch size: 49, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:35:39,051 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:39,639 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98544.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:43,755 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.197e+02 1.665e+02 1.912e+02 2.323e+02 4.830e+02, threshold=3.825e+02, percent-clipped=3.0 +2023-03-26 21:35:57,262 INFO [finetune.py:976] (0/7) Epoch 18, batch 1200, loss[loss=0.1957, simple_loss=0.2656, pruned_loss=0.06289, over 4909.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2527, pruned_loss=0.05626, over 952413.53 frames. ], batch size: 36, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:35:59,148 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2255, 2.1230, 1.7760, 1.9907, 1.9661, 1.9660, 2.0006, 2.6997], + device='cuda:0'), covar=tensor([0.3683, 0.4395, 0.3302, 0.3739, 0.4027, 0.2435, 0.3834, 0.1718], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0228, 0.0276, 0.0253, 0.0220, 0.0252, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:36:12,052 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:36:19,572 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:36:31,240 INFO [finetune.py:976] (0/7) Epoch 18, batch 1250, loss[loss=0.1565, simple_loss=0.2113, pruned_loss=0.0508, over 4163.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2499, pruned_loss=0.05555, over 952698.43 frames. ], batch size: 17, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:37:01,600 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.465e+02 1.816e+02 2.333e+02 4.053e+02, threshold=3.632e+02, percent-clipped=1.0 +2023-03-26 21:37:29,823 INFO [finetune.py:976] (0/7) Epoch 18, batch 1300, loss[loss=0.1512, simple_loss=0.2321, pruned_loss=0.03517, over 4903.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2468, pruned_loss=0.05455, over 954951.98 frames. ], batch size: 32, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:37:54,835 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2985, 2.7828, 2.5623, 1.2932, 2.8181, 2.3983, 2.4310, 2.5879], + device='cuda:0'), covar=tensor([0.0922, 0.1011, 0.1834, 0.2235, 0.1781, 0.2183, 0.1806, 0.1087], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0196, 0.0201, 0.0183, 0.0214, 0.0208, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:38:07,463 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8215, 1.5795, 2.2014, 1.4412, 2.0073, 2.1033, 1.3907, 2.3104], + device='cuda:0'), covar=tensor([0.1487, 0.2266, 0.1305, 0.2063, 0.1116, 0.1535, 0.3072, 0.0971], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0202, 0.0191, 0.0189, 0.0175, 0.0211, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:38:11,398 INFO [finetune.py:976] (0/7) Epoch 18, batch 1350, loss[loss=0.1977, simple_loss=0.2537, pruned_loss=0.07079, over 4813.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2456, pruned_loss=0.05434, over 952102.61 frames. ], batch size: 45, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:38:16,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7698, 1.7911, 1.7162, 1.7573, 1.3887, 3.2710, 1.6112, 1.9660], + device='cuda:0'), covar=tensor([0.2821, 0.2108, 0.1788, 0.1947, 0.1542, 0.0291, 0.2712, 0.1139], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0121, 0.0122, 0.0113, 0.0096, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:38:28,088 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 21:38:31,462 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.676e+02 1.899e+02 2.271e+02 3.691e+02, threshold=3.798e+02, percent-clipped=1.0 +2023-03-26 21:38:38,219 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98762.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:38:40,579 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:38:44,614 INFO [finetune.py:976] (0/7) Epoch 18, batch 1400, loss[loss=0.1994, simple_loss=0.2789, pruned_loss=0.0599, over 4762.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2502, pruned_loss=0.05588, over 951862.22 frames. ], batch size: 54, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:39:10,803 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:39:17,933 INFO [finetune.py:976] (0/7) Epoch 18, batch 1450, loss[loss=0.1744, simple_loss=0.2494, pruned_loss=0.04972, over 4896.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.252, pruned_loss=0.05592, over 954032.54 frames. ], batch size: 35, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:39:40,442 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.618e+02 1.947e+02 2.343e+02 3.750e+02, threshold=3.895e+02, percent-clipped=0.0 +2023-03-26 21:39:52,508 INFO [finetune.py:976] (0/7) Epoch 18, batch 1500, loss[loss=0.1702, simple_loss=0.2499, pruned_loss=0.0453, over 4915.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2533, pruned_loss=0.05634, over 951966.76 frames. ], batch size: 38, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:40:06,459 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98890.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:07,896 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 21:40:10,016 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98895.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:12,865 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:26,172 INFO [finetune.py:976] (0/7) Epoch 18, batch 1550, loss[loss=0.1693, simple_loss=0.2336, pruned_loss=0.05251, over 4691.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2539, pruned_loss=0.05666, over 951535.55 frames. ], batch size: 23, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:40:47,923 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.516e+02 1.782e+02 2.221e+02 4.511e+02, threshold=3.564e+02, percent-clipped=1.0 +2023-03-26 21:40:48,065 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:51,132 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98956.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:41:00,138 INFO [finetune.py:976] (0/7) Epoch 18, batch 1600, loss[loss=0.1757, simple_loss=0.2498, pruned_loss=0.05079, over 4892.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2502, pruned_loss=0.05512, over 953156.12 frames. ], batch size: 32, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:41:00,250 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98971.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:41:33,949 INFO [finetune.py:976] (0/7) Epoch 18, batch 1650, loss[loss=0.1386, simple_loss=0.2111, pruned_loss=0.03303, over 4812.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2483, pruned_loss=0.05526, over 955452.90 frames. ], batch size: 51, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:41:36,672 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 21:41:41,715 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99032.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:02,532 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.819e+01 1.633e+02 2.015e+02 2.343e+02 3.841e+02, threshold=4.030e+02, percent-clipped=3.0 +2023-03-26 21:42:02,667 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:17,579 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:25,675 INFO [finetune.py:976] (0/7) Epoch 18, batch 1700, loss[loss=0.1497, simple_loss=0.2304, pruned_loss=0.03453, over 4815.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2468, pruned_loss=0.05501, over 956059.75 frames. ], batch size: 51, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:01,003 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99112.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:43:02,159 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99114.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:43:10,653 INFO [finetune.py:976] (0/7) Epoch 18, batch 1750, loss[loss=0.205, simple_loss=0.2749, pruned_loss=0.06757, over 4857.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2473, pruned_loss=0.05511, over 956040.51 frames. ], batch size: 31, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:32,523 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8620, 3.3977, 3.5192, 3.7025, 3.6013, 3.3686, 3.9365, 1.2702], + device='cuda:0'), covar=tensor([0.1019, 0.0969, 0.0933, 0.1186, 0.1513, 0.1664, 0.0834, 0.5788], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0241, 0.0275, 0.0289, 0.0330, 0.0278, 0.0297, 0.0291], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:43:38,844 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.106e+01 1.643e+02 1.917e+02 2.422e+02 4.876e+02, threshold=3.835e+02, percent-clipped=2.0 +2023-03-26 21:43:51,916 INFO [finetune.py:976] (0/7) Epoch 18, batch 1800, loss[loss=0.191, simple_loss=0.2595, pruned_loss=0.06121, over 4888.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2484, pruned_loss=0.055, over 955976.47 frames. ], batch size: 32, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:58,265 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 21:44:10,877 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:13,959 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1844, 2.0763, 1.7424, 2.2906, 2.1125, 1.9590, 2.4690, 2.1498], + device='cuda:0'), covar=tensor([0.1261, 0.2290, 0.2987, 0.2355, 0.2409, 0.1647, 0.2636, 0.1771], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0188, 0.0234, 0.0253, 0.0245, 0.0202, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:44:25,712 INFO [finetune.py:976] (0/7) Epoch 18, batch 1850, loss[loss=0.2369, simple_loss=0.2915, pruned_loss=0.09118, over 4216.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.251, pruned_loss=0.05638, over 954473.25 frames. ], batch size: 65, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:44:30,709 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:41,937 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:43,016 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:45,844 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.618e+02 1.939e+02 2.302e+02 3.831e+02, threshold=3.878e+02, percent-clipped=0.0 +2023-03-26 21:44:45,920 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:47,172 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5988, 2.4863, 1.9861, 0.8930, 2.1947, 1.9221, 1.8502, 2.2026], + device='cuda:0'), covar=tensor([0.0927, 0.0768, 0.1731, 0.2222, 0.1533, 0.2374, 0.2177, 0.1011], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0194, 0.0201, 0.0183, 0.0214, 0.0209, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:44:57,632 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:59,345 INFO [finetune.py:976] (0/7) Epoch 18, batch 1900, loss[loss=0.1876, simple_loss=0.2499, pruned_loss=0.06263, over 4284.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2535, pruned_loss=0.05653, over 956370.95 frames. ], batch size: 65, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:45:11,537 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:33,104 INFO [finetune.py:976] (0/7) Epoch 18, batch 1950, loss[loss=0.2249, simple_loss=0.2773, pruned_loss=0.08627, over 4183.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2509, pruned_loss=0.05538, over 957131.80 frames. ], batch size: 65, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:45:35,072 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:36,810 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99327.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:38,102 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:52,712 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.452e+02 1.660e+02 1.987e+02 4.820e+02, threshold=3.320e+02, percent-clipped=1.0 +2023-03-26 21:46:06,274 INFO [finetune.py:976] (0/7) Epoch 18, batch 2000, loss[loss=0.1273, simple_loss=0.2071, pruned_loss=0.02373, over 4745.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2488, pruned_loss=0.05477, over 958386.18 frames. ], batch size: 23, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:46:15,393 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:23,712 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.66 vs. limit=5.0 +2023-03-26 21:46:29,083 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8317, 1.7003, 1.5468, 1.3658, 1.9179, 1.6194, 1.8034, 1.8194], + device='cuda:0'), covar=tensor([0.1357, 0.2018, 0.2859, 0.2420, 0.2502, 0.1705, 0.2788, 0.1944], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0189, 0.0235, 0.0254, 0.0246, 0.0203, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:46:30,176 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:39,572 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:40,078 INFO [finetune.py:976] (0/7) Epoch 18, batch 2050, loss[loss=0.1991, simple_loss=0.2617, pruned_loss=0.06823, over 4835.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2461, pruned_loss=0.05416, over 957642.21 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:46:59,827 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.434e+02 1.742e+02 2.145e+02 5.049e+02, threshold=3.484e+02, percent-clipped=5.0 +2023-03-26 21:47:17,709 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99468.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:47:19,436 INFO [finetune.py:976] (0/7) Epoch 18, batch 2100, loss[loss=0.2133, simple_loss=0.2843, pruned_loss=0.07119, over 4865.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2455, pruned_loss=0.05395, over 956197.52 frames. ], batch size: 31, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:47:31,490 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:47:39,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7263, 1.7738, 1.5173, 1.9187, 2.0500, 1.9072, 1.5422, 1.4453], + device='cuda:0'), covar=tensor([0.2275, 0.1986, 0.2021, 0.1583, 0.1878, 0.1286, 0.2406, 0.1904], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0211, 0.0190, 0.0239, 0.0185, 0.0214, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:47:54,060 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6288, 1.5240, 1.5203, 1.5472, 1.2065, 3.6831, 1.3780, 1.7439], + device='cuda:0'), covar=tensor([0.3541, 0.2650, 0.2275, 0.2461, 0.1812, 0.0186, 0.2826, 0.1400], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 21:47:58,125 INFO [finetune.py:976] (0/7) Epoch 18, batch 2150, loss[loss=0.2036, simple_loss=0.2701, pruned_loss=0.06855, over 4846.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2495, pruned_loss=0.05489, over 956331.95 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:48:08,430 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99529.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:48:16,778 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0552, 1.9275, 1.5574, 0.6746, 1.7261, 1.6513, 1.5349, 1.7768], + device='cuda:0'), covar=tensor([0.0996, 0.0777, 0.1354, 0.1974, 0.1291, 0.2241, 0.2144, 0.0949], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0196, 0.0201, 0.0184, 0.0215, 0.0210, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:48:28,288 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99546.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:35,644 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.546e+02 1.925e+02 2.366e+02 3.688e+02, threshold=3.850e+02, percent-clipped=2.0 +2023-03-26 21:48:35,733 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:51,474 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99568.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:53,717 INFO [finetune.py:976] (0/7) Epoch 18, batch 2200, loss[loss=0.1414, simple_loss=0.2167, pruned_loss=0.03306, over 4769.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2521, pruned_loss=0.05566, over 957236.39 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:49:03,267 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:05,756 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:08,730 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99594.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:11,759 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99599.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:11,829 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7035, 2.4320, 1.9130, 0.9903, 2.1979, 2.0982, 1.9141, 2.1512], + device='cuda:0'), covar=tensor([0.0811, 0.0781, 0.1616, 0.2020, 0.1352, 0.2254, 0.1984, 0.0953], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0195, 0.0200, 0.0184, 0.0214, 0.0209, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:49:27,189 INFO [finetune.py:976] (0/7) Epoch 18, batch 2250, loss[loss=0.1777, simple_loss=0.2435, pruned_loss=0.05589, over 4747.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2528, pruned_loss=0.05588, over 957088.96 frames. ], batch size: 27, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:49:29,578 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:31,396 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99627.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:33,078 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:39,630 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3371, 2.3493, 1.7124, 2.4543, 2.3308, 1.9901, 2.8661, 2.4400], + device='cuda:0'), covar=tensor([0.1206, 0.2149, 0.2971, 0.2596, 0.2426, 0.1507, 0.2916, 0.1668], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0187, 0.0233, 0.0252, 0.0244, 0.0201, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:49:46,332 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:46,820 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.914e+01 1.516e+02 1.824e+02 2.092e+02 3.162e+02, threshold=3.647e+02, percent-clipped=0.0 +2023-03-26 21:49:55,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1674, 2.0220, 1.7089, 1.9789, 1.9111, 1.8823, 1.9537, 2.6668], + device='cuda:0'), covar=tensor([0.3544, 0.4348, 0.3228, 0.3835, 0.4217, 0.2534, 0.3914, 0.1574], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0261, 0.0227, 0.0275, 0.0251, 0.0219, 0.0251, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:50:00,832 INFO [finetune.py:976] (0/7) Epoch 18, batch 2300, loss[loss=0.1417, simple_loss=0.2261, pruned_loss=0.02863, over 4769.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2522, pruned_loss=0.05551, over 957691.88 frames. ], batch size: 27, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:50:03,801 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99675.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:06,804 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:06,831 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:24,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99707.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:34,081 INFO [finetune.py:976] (0/7) Epoch 18, batch 2350, loss[loss=0.1758, simple_loss=0.2394, pruned_loss=0.05613, over 4762.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2503, pruned_loss=0.05491, over 958020.53 frames. ], batch size: 54, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:50:47,880 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:52,120 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 21:50:54,364 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.551e+02 1.845e+02 2.144e+02 4.060e+02, threshold=3.690e+02, percent-clipped=1.0 +2023-03-26 21:50:56,922 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99755.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:51:08,062 INFO [finetune.py:976] (0/7) Epoch 18, batch 2400, loss[loss=0.1491, simple_loss=0.223, pruned_loss=0.03759, over 4906.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2475, pruned_loss=0.05456, over 960686.77 frames. ], batch size: 35, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:51:11,642 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:51:23,522 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6012, 1.5079, 2.3062, 3.2257, 2.1962, 2.3735, 1.2842, 2.6312], + device='cuda:0'), covar=tensor([0.1608, 0.1374, 0.1154, 0.0550, 0.0790, 0.1397, 0.1577, 0.0522], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0165, 0.0101, 0.0135, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 21:51:39,935 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 21:51:41,396 INFO [finetune.py:976] (0/7) Epoch 18, batch 2450, loss[loss=0.1657, simple_loss=0.2465, pruned_loss=0.0425, over 4905.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2451, pruned_loss=0.0539, over 959185.26 frames. ], batch size: 36, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:51:43,689 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99824.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:51:57,452 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 21:52:01,734 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.091e+01 1.409e+02 1.705e+02 2.080e+02 4.896e+02, threshold=3.409e+02, percent-clipped=2.0 +2023-03-26 21:52:14,325 INFO [finetune.py:976] (0/7) Epoch 18, batch 2500, loss[loss=0.1972, simple_loss=0.275, pruned_loss=0.05969, over 4903.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2481, pruned_loss=0.05494, over 960314.17 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:52:26,318 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:50,111 INFO [finetune.py:976] (0/7) Epoch 18, batch 2550, loss[loss=0.2052, simple_loss=0.269, pruned_loss=0.0707, over 4870.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2516, pruned_loss=0.05621, over 957770.69 frames. ], batch size: 31, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:52:52,498 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:52,515 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:58,895 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:06,642 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:10,119 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3858, 0.9276, 0.6899, 1.2273, 1.7770, 0.6892, 1.1067, 1.2264], + device='cuda:0'), covar=tensor([0.1335, 0.1984, 0.1524, 0.1146, 0.1631, 0.1729, 0.1388, 0.1832], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0096, 0.0110, 0.0092, 0.0120, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 21:53:10,619 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.647e+02 1.933e+02 2.438e+02 4.501e+02, threshold=3.867e+02, percent-clipped=7.0 +2023-03-26 21:53:14,349 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-26 21:53:20,823 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4451, 2.4165, 2.3205, 1.7072, 2.3783, 2.6149, 2.6505, 2.1503], + device='cuda:0'), covar=tensor([0.0625, 0.0663, 0.0762, 0.0872, 0.0758, 0.0666, 0.0648, 0.1055], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0134, 0.0141, 0.0121, 0.0123, 0.0139, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:53:29,758 INFO [finetune.py:976] (0/7) Epoch 18, batch 2600, loss[loss=0.2401, simple_loss=0.3094, pruned_loss=0.08538, over 4907.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2539, pruned_loss=0.05674, over 957726.07 frames. ], batch size: 38, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:53:30,434 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:40,421 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:02,729 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-100000.pt +2023-03-26 21:54:12,862 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100007.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:54:24,615 INFO [finetune.py:976] (0/7) Epoch 18, batch 2650, loss[loss=0.1858, simple_loss=0.2646, pruned_loss=0.05354, over 4892.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2548, pruned_loss=0.05698, over 958191.26 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:54:29,387 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:35,198 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:42,276 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8397, 1.5759, 1.8234, 1.2735, 1.8487, 1.8363, 1.8346, 1.3026], + device='cuda:0'), covar=tensor([0.0747, 0.1056, 0.0812, 0.1098, 0.0880, 0.0790, 0.0785, 0.1921], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0135, 0.0141, 0.0122, 0.0124, 0.0139, 0.0141, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:54:46,209 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.530e+02 1.912e+02 2.295e+02 4.144e+02, threshold=3.823e+02, percent-clipped=1.0 +2023-03-26 21:54:56,607 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100068.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:54:58,304 INFO [finetune.py:976] (0/7) Epoch 18, batch 2700, loss[loss=0.1884, simple_loss=0.2562, pruned_loss=0.06025, over 4844.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2525, pruned_loss=0.05579, over 955841.59 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:55:01,426 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:55:04,445 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-03-26 21:55:15,699 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-03-26 21:55:31,862 INFO [finetune.py:976] (0/7) Epoch 18, batch 2750, loss[loss=0.1399, simple_loss=0.209, pruned_loss=0.03535, over 4842.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2495, pruned_loss=0.05494, over 954058.94 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:55:33,699 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100124.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:55:33,751 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100124.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:55:52,996 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.510e+02 1.766e+02 2.096e+02 4.575e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 21:56:05,347 INFO [finetune.py:976] (0/7) Epoch 18, batch 2800, loss[loss=0.1847, simple_loss=0.2436, pruned_loss=0.06287, over 4936.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.246, pruned_loss=0.05353, over 956099.04 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:56:06,008 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100172.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:56:11,427 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9380, 1.3965, 1.8664, 1.8249, 1.6664, 1.6523, 1.7336, 1.7450], + device='cuda:0'), covar=tensor([0.5042, 0.4803, 0.4164, 0.4331, 0.5798, 0.4926, 0.5607, 0.4183], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0239, 0.0259, 0.0274, 0.0273, 0.0248, 0.0284, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:56:20,134 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 21:56:38,930 INFO [finetune.py:976] (0/7) Epoch 18, batch 2850, loss[loss=0.1398, simple_loss=0.2151, pruned_loss=0.03224, over 4769.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.245, pruned_loss=0.05395, over 954909.59 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:56:40,880 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100224.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:44,570 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0020, 1.9559, 1.5581, 1.8151, 1.7711, 1.7244, 1.8488, 2.5113], + device='cuda:0'), covar=tensor([0.3556, 0.3916, 0.3191, 0.3571, 0.3873, 0.2445, 0.3657, 0.1579], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0226, 0.0272, 0.0249, 0.0218, 0.0249, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:56:49,825 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100238.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:54,568 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:59,192 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.635e+02 1.899e+02 2.309e+02 4.393e+02, threshold=3.799e+02, percent-clipped=4.0 +2023-03-26 21:57:11,713 INFO [finetune.py:976] (0/7) Epoch 18, batch 2900, loss[loss=0.1987, simple_loss=0.2676, pruned_loss=0.06487, over 4833.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2468, pruned_loss=0.05416, over 956240.54 frames. ], batch size: 30, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:57:12,862 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:26,151 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:30,854 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:35,737 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-03-26 21:57:39,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5093, 1.3172, 1.1950, 1.4652, 1.7010, 1.5759, 0.9861, 1.2666], + device='cuda:0'), covar=tensor([0.2283, 0.2242, 0.2060, 0.1774, 0.1528, 0.1261, 0.2639, 0.1962], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0208, 0.0211, 0.0190, 0.0240, 0.0185, 0.0215, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 21:57:45,539 INFO [finetune.py:976] (0/7) Epoch 18, batch 2950, loss[loss=0.2184, simple_loss=0.2898, pruned_loss=0.07351, over 4804.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2495, pruned_loss=0.05514, over 955039.34 frames. ], batch size: 41, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:57:55,262 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:58:06,321 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.490e+02 1.822e+02 2.174e+02 4.072e+02, threshold=3.643e+02, percent-clipped=2.0 +2023-03-26 21:58:13,572 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100363.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:58:18,413 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 21:58:18,818 INFO [finetune.py:976] (0/7) Epoch 18, batch 3000, loss[loss=0.1653, simple_loss=0.2372, pruned_loss=0.04671, over 4845.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2516, pruned_loss=0.0559, over 954969.62 frames. ], batch size: 30, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:58:18,819 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 21:58:31,198 INFO [finetune.py:1010] (0/7) Epoch 18, validation: loss=0.1568, simple_loss=0.2261, pruned_loss=0.04375, over 2265189.00 frames. +2023-03-26 21:58:31,199 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 21:58:44,386 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:59:06,994 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100402.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:59:11,280 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 21:59:29,743 INFO [finetune.py:976] (0/7) Epoch 18, batch 3050, loss[loss=0.1917, simple_loss=0.2631, pruned_loss=0.0602, over 4840.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2512, pruned_loss=0.0552, over 953435.86 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:59:53,773 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.572e+02 1.917e+02 2.276e+02 3.597e+02, threshold=3.833e+02, percent-clipped=0.0 +2023-03-26 22:00:01,168 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100463.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:00:07,400 INFO [finetune.py:976] (0/7) Epoch 18, batch 3100, loss[loss=0.1521, simple_loss=0.2255, pruned_loss=0.03934, over 4760.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2502, pruned_loss=0.05521, over 954367.85 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 22:00:25,919 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-03-26 22:00:40,538 INFO [finetune.py:976] (0/7) Epoch 18, batch 3150, loss[loss=0.2481, simple_loss=0.3015, pruned_loss=0.09731, over 4798.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.248, pruned_loss=0.05474, over 954535.24 frames. ], batch size: 45, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 22:00:53,532 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2884, 2.2064, 1.7129, 2.2185, 2.1762, 1.9105, 2.5362, 2.3042], + device='cuda:0'), covar=tensor([0.1211, 0.2314, 0.3075, 0.2825, 0.2648, 0.1707, 0.3622, 0.1722], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0187, 0.0234, 0.0252, 0.0244, 0.0202, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:00:59,246 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7973, 1.5478, 2.1905, 1.8579, 1.8270, 4.3128, 1.5558, 1.8121], + device='cuda:0'), covar=tensor([0.0950, 0.1827, 0.1191, 0.1041, 0.1570, 0.0225, 0.1544, 0.1805], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0090, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:01:00,942 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.523e+02 1.835e+02 2.195e+02 4.344e+02, threshold=3.670e+02, percent-clipped=3.0 +2023-03-26 22:01:12,892 INFO [finetune.py:976] (0/7) Epoch 18, batch 3200, loss[loss=0.1584, simple_loss=0.2265, pruned_loss=0.04513, over 4757.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2452, pruned_loss=0.05383, over 955648.54 frames. ], batch size: 28, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:01:28,940 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100594.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:01:30,210 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.4687, 2.9041, 2.6749, 1.3849, 2.9135, 2.4870, 2.4750, 2.5869], + device='cuda:0'), covar=tensor([0.0848, 0.0935, 0.2167, 0.2327, 0.1656, 0.2143, 0.1956, 0.1208], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0193, 0.0199, 0.0182, 0.0212, 0.0207, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:01:46,315 INFO [finetune.py:976] (0/7) Epoch 18, batch 3250, loss[loss=0.2127, simple_loss=0.2793, pruned_loss=0.07304, over 4810.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2467, pruned_loss=0.0548, over 954841.71 frames. ], batch size: 45, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:08,114 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.502e+02 1.862e+02 2.235e+02 4.464e+02, threshold=3.723e+02, percent-clipped=3.0 +2023-03-26 22:02:14,915 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100663.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:02:20,154 INFO [finetune.py:976] (0/7) Epoch 18, batch 3300, loss[loss=0.167, simple_loss=0.2433, pruned_loss=0.04535, over 4901.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2511, pruned_loss=0.05565, over 955635.56 frames. ], batch size: 32, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:47,027 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100711.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:02:53,405 INFO [finetune.py:976] (0/7) Epoch 18, batch 3350, loss[loss=0.1413, simple_loss=0.2065, pruned_loss=0.038, over 4062.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2521, pruned_loss=0.05581, over 954390.38 frames. ], batch size: 17, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:59,508 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100730.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:03:08,296 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0773, 0.9812, 0.9871, 1.1569, 1.2429, 1.1697, 1.0170, 0.9632], + device='cuda:0'), covar=tensor([0.0332, 0.0293, 0.0609, 0.0312, 0.0281, 0.0440, 0.0303, 0.0384], + device='cuda:0'), in_proj_covar=tensor([0.0096, 0.0107, 0.0143, 0.0111, 0.0101, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.4173e-05, 8.2478e-05, 1.1276e-04, 8.4889e-05, 7.8444e-05, 7.9916e-05, + 7.3440e-05, 8.3366e-05], device='cuda:0') +2023-03-26 22:03:14,083 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.901e+01 1.577e+02 1.865e+02 2.249e+02 4.268e+02, threshold=3.731e+02, percent-clipped=3.0 +2023-03-26 22:03:14,393 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 22:03:18,260 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100758.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:03:26,646 INFO [finetune.py:976] (0/7) Epoch 18, batch 3400, loss[loss=0.2062, simple_loss=0.2734, pruned_loss=0.06954, over 4896.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2542, pruned_loss=0.0566, over 957267.67 frames. ], batch size: 36, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:03:27,978 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4588, 1.3017, 1.9262, 2.8714, 1.8613, 2.2069, 1.0749, 2.3838], + device='cuda:0'), covar=tensor([0.1581, 0.1402, 0.1119, 0.0593, 0.0830, 0.1458, 0.1472, 0.0545], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0166, 0.0102, 0.0136, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:03:40,310 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100791.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:04:13,288 INFO [finetune.py:976] (0/7) Epoch 18, batch 3450, loss[loss=0.2017, simple_loss=0.271, pruned_loss=0.0662, over 4901.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2538, pruned_loss=0.05634, over 958277.72 frames. ], batch size: 37, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:04:14,015 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5136, 1.4568, 1.3265, 1.4550, 1.7946, 1.7047, 1.4880, 1.2624], + device='cuda:0'), covar=tensor([0.0299, 0.0308, 0.0635, 0.0305, 0.0232, 0.0519, 0.0381, 0.0399], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0107, 0.0143, 0.0110, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.4024e-05, 8.2353e-05, 1.1261e-04, 8.4594e-05, 7.8234e-05, 7.9916e-05, + 7.3467e-05, 8.3214e-05], device='cuda:0') +2023-03-26 22:04:49,164 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.551e+02 1.762e+02 2.136e+02 3.810e+02, threshold=3.524e+02, percent-clipped=1.0 +2023-03-26 22:05:03,936 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.15 vs. limit=5.0 +2023-03-26 22:05:04,969 INFO [finetune.py:976] (0/7) Epoch 18, batch 3500, loss[loss=0.1516, simple_loss=0.2303, pruned_loss=0.03643, over 4913.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2514, pruned_loss=0.05548, over 956344.80 frames. ], batch size: 43, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:05:21,049 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100894.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:05:38,763 INFO [finetune.py:976] (0/7) Epoch 18, batch 3550, loss[loss=0.1668, simple_loss=0.2398, pruned_loss=0.04689, over 4728.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.249, pruned_loss=0.05482, over 957542.49 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:05:49,118 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1939, 2.0288, 1.6917, 1.9408, 2.1296, 1.8751, 2.3182, 2.1493], + device='cuda:0'), covar=tensor([0.1376, 0.2083, 0.2994, 0.2635, 0.2547, 0.1707, 0.3403, 0.1734], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0186, 0.0232, 0.0251, 0.0243, 0.0201, 0.0212, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:05:52,400 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100942.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:05:59,326 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.557e+01 1.553e+02 1.835e+02 2.348e+02 4.609e+02, threshold=3.670e+02, percent-clipped=4.0 +2023-03-26 22:06:12,132 INFO [finetune.py:976] (0/7) Epoch 18, batch 3600, loss[loss=0.1828, simple_loss=0.2471, pruned_loss=0.05926, over 4753.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2467, pruned_loss=0.05422, over 956651.38 frames. ], batch size: 54, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:06:41,390 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7915, 3.3710, 3.5067, 3.6742, 3.5422, 3.3327, 3.8646, 1.1992], + device='cuda:0'), covar=tensor([0.0947, 0.0913, 0.0833, 0.1028, 0.1415, 0.1696, 0.0862, 0.5518], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0278, 0.0291, 0.0334, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:06:46,076 INFO [finetune.py:976] (0/7) Epoch 18, batch 3650, loss[loss=0.2024, simple_loss=0.2908, pruned_loss=0.057, over 4834.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2486, pruned_loss=0.05506, over 955290.26 frames. ], batch size: 47, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:07:06,790 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.559e+02 1.860e+02 2.177e+02 4.070e+02, threshold=3.719e+02, percent-clipped=1.0 +2023-03-26 22:07:10,511 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:18,906 INFO [finetune.py:976] (0/7) Epoch 18, batch 3700, loss[loss=0.1722, simple_loss=0.2549, pruned_loss=0.0447, over 4905.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2514, pruned_loss=0.05576, over 955432.20 frames. ], batch size: 37, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:07:28,499 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101086.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:07:37,442 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2201, 3.6401, 3.8359, 4.0908, 4.0059, 3.7308, 4.3059, 1.3944], + device='cuda:0'), covar=tensor([0.0845, 0.0848, 0.0822, 0.0884, 0.1289, 0.1604, 0.0747, 0.5691], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0243, 0.0277, 0.0289, 0.0333, 0.0281, 0.0299, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:07:42,146 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:43,250 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101107.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:52,640 INFO [finetune.py:976] (0/7) Epoch 18, batch 3750, loss[loss=0.1496, simple_loss=0.224, pruned_loss=0.03765, over 4762.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2534, pruned_loss=0.05602, over 955549.89 frames. ], batch size: 27, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:07:54,580 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7334, 1.6381, 2.1654, 1.8842, 1.8881, 4.2744, 1.6808, 1.8545], + device='cuda:0'), covar=tensor([0.0995, 0.1926, 0.1316, 0.0985, 0.1632, 0.0201, 0.1489, 0.1833], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:07:54,595 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6814, 2.7934, 2.6517, 1.8849, 2.6065, 2.9233, 2.8894, 2.4416], + device='cuda:0'), covar=tensor([0.0559, 0.0563, 0.0652, 0.0869, 0.0599, 0.0597, 0.0579, 0.0904], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0135, 0.0140, 0.0120, 0.0123, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:07:59,884 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5519, 1.4021, 1.8168, 1.7444, 1.5178, 3.3506, 1.3240, 1.5061], + device='cuda:0'), covar=tensor([0.0985, 0.1996, 0.1142, 0.0963, 0.1694, 0.0249, 0.1562, 0.1834], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:08:12,837 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.641e+02 1.814e+02 2.131e+02 4.110e+02, threshold=3.627e+02, percent-clipped=2.0 +2023-03-26 22:08:24,456 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:08:26,163 INFO [finetune.py:976] (0/7) Epoch 18, batch 3800, loss[loss=0.188, simple_loss=0.2624, pruned_loss=0.05683, over 4781.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2541, pruned_loss=0.05629, over 955423.40 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:08:41,814 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 22:08:59,899 INFO [finetune.py:976] (0/7) Epoch 18, batch 3850, loss[loss=0.1814, simple_loss=0.2503, pruned_loss=0.05626, over 4832.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2519, pruned_loss=0.05494, over 956387.89 frames. ], batch size: 30, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:09:02,512 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 22:09:28,089 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 22:09:28,671 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5281, 2.3638, 1.8270, 2.5243, 2.4956, 2.1447, 2.9752, 2.5408], + device='cuda:0'), covar=tensor([0.1287, 0.2292, 0.2973, 0.2703, 0.2402, 0.1547, 0.3388, 0.1759], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0189, 0.0236, 0.0254, 0.0246, 0.0204, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:09:30,622 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-03-26 22:09:30,993 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.525e+02 1.862e+02 2.180e+02 4.556e+02, threshold=3.724e+02, percent-clipped=2.0 +2023-03-26 22:09:57,716 INFO [finetune.py:976] (0/7) Epoch 18, batch 3900, loss[loss=0.167, simple_loss=0.2355, pruned_loss=0.04924, over 4825.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2489, pruned_loss=0.05473, over 954079.76 frames. ], batch size: 33, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:10:31,567 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-03-26 22:10:41,619 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 22:10:41,970 INFO [finetune.py:976] (0/7) Epoch 18, batch 3950, loss[loss=0.1783, simple_loss=0.2459, pruned_loss=0.05535, over 4822.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2461, pruned_loss=0.05417, over 955650.69 frames. ], batch size: 41, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:11:02,251 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.470e+02 1.754e+02 2.083e+02 3.090e+02, threshold=3.508e+02, percent-clipped=0.0 +2023-03-26 22:11:15,361 INFO [finetune.py:976] (0/7) Epoch 18, batch 4000, loss[loss=0.1849, simple_loss=0.2584, pruned_loss=0.05568, over 4777.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2449, pruned_loss=0.05341, over 955718.87 frames. ], batch size: 28, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:11:26,018 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101386.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:11:28,515 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-03-26 22:11:43,639 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5717, 1.4742, 1.4188, 1.5504, 1.2168, 3.6802, 1.4775, 1.8140], + device='cuda:0'), covar=tensor([0.3348, 0.2591, 0.2299, 0.2475, 0.1768, 0.0154, 0.2525, 0.1259], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:11:49,438 INFO [finetune.py:976] (0/7) Epoch 18, batch 4050, loss[loss=0.2251, simple_loss=0.281, pruned_loss=0.08458, over 4929.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2496, pruned_loss=0.05497, over 956176.64 frames. ], batch size: 33, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:11:52,412 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4818, 1.7239, 1.4381, 1.4107, 1.9840, 1.8966, 1.7768, 1.6710], + device='cuda:0'), covar=tensor([0.0573, 0.0338, 0.0618, 0.0406, 0.0371, 0.0696, 0.0385, 0.0428], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0105, 0.0141, 0.0109, 0.0099, 0.0107, 0.0097, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3272e-05, 8.1387e-05, 1.1121e-04, 8.3783e-05, 7.7272e-05, 7.8965e-05, + 7.2532e-05, 8.2914e-05], device='cuda:0') +2023-03-26 22:11:58,801 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101434.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:12:05,531 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6614, 1.0099, 1.0132, 1.4644, 1.9055, 1.5357, 1.2826, 1.4382], + device='cuda:0'), covar=tensor([0.1501, 0.2520, 0.1800, 0.1328, 0.2103, 0.2104, 0.1585, 0.2140], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0109, 0.0092, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:12:10,021 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.670e+02 2.040e+02 2.363e+02 9.256e+02, threshold=4.080e+02, percent-clipped=2.0 +2023-03-26 22:12:17,268 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:12:22,996 INFO [finetune.py:976] (0/7) Epoch 18, batch 4100, loss[loss=0.1535, simple_loss=0.2154, pruned_loss=0.04581, over 4381.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2527, pruned_loss=0.0564, over 955900.10 frames. ], batch size: 19, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:12:34,992 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 22:12:56,254 INFO [finetune.py:976] (0/7) Epoch 18, batch 4150, loss[loss=0.1959, simple_loss=0.2782, pruned_loss=0.05681, over 4845.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2535, pruned_loss=0.05648, over 954208.91 frames. ], batch size: 49, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:13:16,878 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.508e+02 1.851e+02 2.208e+02 3.984e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 22:13:24,820 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 22:13:29,469 INFO [finetune.py:976] (0/7) Epoch 18, batch 4200, loss[loss=0.197, simple_loss=0.2685, pruned_loss=0.06278, over 4909.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2532, pruned_loss=0.05568, over 955798.16 frames. ], batch size: 46, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:00,302 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-26 22:14:03,051 INFO [finetune.py:976] (0/7) Epoch 18, batch 4250, loss[loss=0.1996, simple_loss=0.2596, pruned_loss=0.06985, over 4901.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2503, pruned_loss=0.05485, over 956485.49 frames. ], batch size: 35, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:24,253 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.580e+02 1.774e+02 2.219e+02 3.425e+02, threshold=3.547e+02, percent-clipped=0.0 +2023-03-26 22:14:38,481 INFO [finetune.py:976] (0/7) Epoch 18, batch 4300, loss[loss=0.1641, simple_loss=0.2295, pruned_loss=0.04939, over 4815.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2459, pruned_loss=0.05341, over 955362.50 frames. ], batch size: 45, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:54,335 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:15:28,659 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 22:15:33,151 INFO [finetune.py:976] (0/7) Epoch 18, batch 4350, loss[loss=0.188, simple_loss=0.2433, pruned_loss=0.06634, over 4823.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2431, pruned_loss=0.0529, over 957046.30 frames. ], batch size: 30, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:16:00,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5766, 1.4851, 1.4484, 0.7250, 1.5968, 1.7470, 1.7673, 1.3397], + device='cuda:0'), covar=tensor([0.0871, 0.0631, 0.0484, 0.0542, 0.0396, 0.0526, 0.0324, 0.0640], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0124, 0.0127, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.1589e-05, 1.1029e-04, 8.8882e-05, 9.0281e-05, 9.2503e-05, 9.2933e-05, + 1.0198e-04, 1.0700e-04], device='cuda:0') +2023-03-26 22:16:05,133 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:10,238 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.272e+01 1.483e+02 1.686e+02 2.087e+02 3.591e+02, threshold=3.373e+02, percent-clipped=1.0 +2023-03-26 22:16:16,963 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:21,724 INFO [finetune.py:976] (0/7) Epoch 18, batch 4400, loss[loss=0.1466, simple_loss=0.2115, pruned_loss=0.0408, over 4719.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2446, pruned_loss=0.05405, over 954770.17 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:16:49,623 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:55,630 INFO [finetune.py:976] (0/7) Epoch 18, batch 4450, loss[loss=0.157, simple_loss=0.2291, pruned_loss=0.04249, over 4869.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2478, pruned_loss=0.05505, over 952781.34 frames. ], batch size: 31, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:17:01,634 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1833, 1.9627, 1.4703, 0.5809, 1.6174, 1.7558, 1.6229, 1.7642], + device='cuda:0'), covar=tensor([0.0951, 0.0751, 0.1525, 0.2041, 0.1466, 0.2592, 0.2526, 0.0882], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0200, 0.0184, 0.0213, 0.0209, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:17:05,389 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 22:17:16,755 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.606e+02 1.893e+02 2.313e+02 4.401e+02, threshold=3.785e+02, percent-clipped=7.0 +2023-03-26 22:17:29,381 INFO [finetune.py:976] (0/7) Epoch 18, batch 4500, loss[loss=0.1879, simple_loss=0.259, pruned_loss=0.0584, over 4836.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2493, pruned_loss=0.05503, over 953138.97 frames. ], batch size: 49, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:02,636 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101920.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:18:03,132 INFO [finetune.py:976] (0/7) Epoch 18, batch 4550, loss[loss=0.2005, simple_loss=0.2752, pruned_loss=0.0629, over 4913.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2514, pruned_loss=0.05622, over 952467.02 frames. ], batch size: 36, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:09,833 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8599, 1.6682, 1.4127, 1.2094, 1.6068, 1.6304, 1.6136, 2.1807], + device='cuda:0'), covar=tensor([0.3949, 0.3790, 0.3261, 0.3660, 0.3745, 0.2650, 0.3520, 0.1979], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0227, 0.0272, 0.0250, 0.0218, 0.0250, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:18:14,467 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9264, 1.4930, 1.9907, 1.8793, 1.6934, 1.6845, 1.8458, 1.8221], + device='cuda:0'), covar=tensor([0.3573, 0.3478, 0.2762, 0.3390, 0.4182, 0.3658, 0.3803, 0.2732], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0240, 0.0259, 0.0275, 0.0275, 0.0248, 0.0283, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:18:24,137 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.522e+02 1.794e+02 2.336e+02 4.256e+02, threshold=3.587e+02, percent-clipped=2.0 +2023-03-26 22:18:36,744 INFO [finetune.py:976] (0/7) Epoch 18, batch 4600, loss[loss=0.226, simple_loss=0.2925, pruned_loss=0.07974, over 4230.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2506, pruned_loss=0.05545, over 952126.20 frames. ], batch size: 65, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:38,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9121, 1.4086, 1.8779, 1.9013, 1.6651, 1.6642, 1.8210, 1.7426], + device='cuda:0'), covar=tensor([0.4214, 0.4142, 0.3555, 0.3947, 0.5231, 0.4070, 0.4789, 0.3340], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0241, 0.0260, 0.0276, 0.0276, 0.0249, 0.0284, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:18:42,888 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101981.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:18:47,488 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:18:53,533 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-03-26 22:18:55,907 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-102000.pt +2023-03-26 22:19:01,298 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2993, 2.3235, 1.8986, 2.4200, 2.2237, 2.2404, 2.1891, 3.1226], + device='cuda:0'), covar=tensor([0.3983, 0.4811, 0.3619, 0.4227, 0.4397, 0.2513, 0.4558, 0.1737], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0226, 0.0271, 0.0249, 0.0218, 0.0249, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:19:11,218 INFO [finetune.py:976] (0/7) Epoch 18, batch 4650, loss[loss=0.1668, simple_loss=0.2358, pruned_loss=0.04886, over 4799.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2493, pruned_loss=0.05526, over 954400.09 frames. ], batch size: 29, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:19:23,873 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:19:29,229 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:19:31,548 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.564e+02 1.867e+02 2.217e+02 4.281e+02, threshold=3.734e+02, percent-clipped=4.0 +2023-03-26 22:19:37,698 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 22:19:39,767 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7651, 1.8144, 2.0574, 2.0048, 1.8875, 3.5629, 1.6342, 1.8650], + device='cuda:0'), covar=tensor([0.1083, 0.2023, 0.1145, 0.0980, 0.1534, 0.0271, 0.1676, 0.2034], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:19:45,055 INFO [finetune.py:976] (0/7) Epoch 18, batch 4700, loss[loss=0.1759, simple_loss=0.2433, pruned_loss=0.0543, over 4922.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2458, pruned_loss=0.05424, over 953138.41 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:19:57,270 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.4266, 4.6717, 4.9691, 5.2614, 5.1995, 4.8635, 5.5391, 1.9540], + device='cuda:0'), covar=tensor([0.0688, 0.0836, 0.0717, 0.0702, 0.1103, 0.1473, 0.0446, 0.5133], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0278, 0.0290, 0.0331, 0.0281, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:20:31,468 INFO [finetune.py:976] (0/7) Epoch 18, batch 4750, loss[loss=0.1719, simple_loss=0.2546, pruned_loss=0.04459, over 4807.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2442, pruned_loss=0.05351, over 953699.82 frames. ], batch size: 45, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:20:40,654 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2773, 2.9197, 3.0323, 3.1976, 3.0379, 2.8717, 3.3434, 1.0221], + device='cuda:0'), covar=tensor([0.1191, 0.1091, 0.1092, 0.1341, 0.1785, 0.1921, 0.1097, 0.5418], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0278, 0.0290, 0.0332, 0.0281, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:20:56,162 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.538e+02 1.904e+02 2.326e+02 4.380e+02, threshold=3.807e+02, percent-clipped=2.0 +2023-03-26 22:20:56,435 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 22:21:02,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2538, 2.1461, 1.7026, 2.0991, 2.1682, 1.9006, 2.5022, 2.2807], + device='cuda:0'), covar=tensor([0.1386, 0.1932, 0.2957, 0.2523, 0.2379, 0.1587, 0.2600, 0.1674], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0189, 0.0235, 0.0254, 0.0245, 0.0203, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:21:23,316 INFO [finetune.py:976] (0/7) Epoch 18, batch 4800, loss[loss=0.2027, simple_loss=0.2834, pruned_loss=0.061, over 4900.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2481, pruned_loss=0.05501, over 954763.52 frames. ], batch size: 43, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:21:33,224 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 22:21:51,162 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0850, 1.9717, 1.6196, 1.9337, 1.8561, 1.8883, 1.9887, 2.6519], + device='cuda:0'), covar=tensor([0.3763, 0.4278, 0.3284, 0.3927, 0.4156, 0.2479, 0.3798, 0.1675], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0227, 0.0272, 0.0250, 0.0218, 0.0250, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:22:00,409 INFO [finetune.py:976] (0/7) Epoch 18, batch 4850, loss[loss=0.2107, simple_loss=0.2925, pruned_loss=0.0645, over 4767.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2527, pruned_loss=0.05624, over 954501.92 frames. ], batch size: 54, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:22:09,470 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7126, 1.1548, 0.9224, 1.6099, 2.0918, 1.5531, 1.5496, 1.5492], + device='cuda:0'), covar=tensor([0.1554, 0.2139, 0.1802, 0.1214, 0.1990, 0.1998, 0.1432, 0.1959], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:22:13,792 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3918, 1.2961, 1.2411, 1.3154, 1.5551, 1.5140, 1.3160, 1.2088], + device='cuda:0'), covar=tensor([0.0378, 0.0310, 0.0528, 0.0321, 0.0274, 0.0417, 0.0338, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0106, 0.0141, 0.0109, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:0'), out_proj_covar=tensor([7.3108e-05, 8.1627e-05, 1.1144e-04, 8.3793e-05, 7.6871e-05, 7.8986e-05, + 7.2675e-05, 8.2704e-05], device='cuda:0') +2023-03-26 22:22:20,215 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.897e+01 1.478e+02 1.803e+02 2.234e+02 3.533e+02, threshold=3.606e+02, percent-clipped=0.0 +2023-03-26 22:22:33,608 INFO [finetune.py:976] (0/7) Epoch 18, batch 4900, loss[loss=0.1752, simple_loss=0.2521, pruned_loss=0.04915, over 4925.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2543, pruned_loss=0.05654, over 955422.86 frames. ], batch size: 41, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:22:37,663 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102276.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 22:23:06,927 INFO [finetune.py:976] (0/7) Epoch 18, batch 4950, loss[loss=0.1515, simple_loss=0.2277, pruned_loss=0.03766, over 4677.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2559, pruned_loss=0.05713, over 953270.59 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:23:19,538 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102340.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:21,340 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:26,705 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.970e+01 1.449e+02 1.847e+02 2.188e+02 4.191e+02, threshold=3.694e+02, percent-clipped=1.0 +2023-03-26 22:23:40,083 INFO [finetune.py:976] (0/7) Epoch 18, batch 5000, loss[loss=0.1547, simple_loss=0.2321, pruned_loss=0.03862, over 4799.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2529, pruned_loss=0.05615, over 954149.07 frames. ], batch size: 51, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:23:47,726 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-26 22:23:51,808 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:58,593 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9887, 1.3853, 2.0278, 1.9639, 1.8046, 1.7501, 1.9153, 1.8697], + device='cuda:0'), covar=tensor([0.3483, 0.3632, 0.2860, 0.3318, 0.4432, 0.3495, 0.4026, 0.2846], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0240, 0.0258, 0.0274, 0.0273, 0.0247, 0.0282, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:24:08,528 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0725, 1.9745, 2.6273, 3.7331, 2.6301, 2.8745, 1.5805, 3.0370], + device='cuda:0'), covar=tensor([0.1436, 0.1195, 0.1066, 0.0490, 0.0725, 0.1020, 0.1488, 0.0459], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0163, 0.0099, 0.0134, 0.0123, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:24:13,295 INFO [finetune.py:976] (0/7) Epoch 18, batch 5050, loss[loss=0.1437, simple_loss=0.2216, pruned_loss=0.03292, over 4810.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2508, pruned_loss=0.05613, over 954592.47 frames. ], batch size: 51, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:24:33,970 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.223e+01 1.582e+02 1.966e+02 2.354e+02 3.513e+02, threshold=3.932e+02, percent-clipped=0.0 +2023-03-26 22:24:45,178 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2827, 2.9135, 3.0646, 3.2056, 3.0655, 2.9261, 3.3408, 1.0040], + device='cuda:0'), covar=tensor([0.1149, 0.1152, 0.1123, 0.1201, 0.1598, 0.1890, 0.1118, 0.5578], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0280, 0.0292, 0.0334, 0.0283, 0.0303, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:24:46,889 INFO [finetune.py:976] (0/7) Epoch 18, batch 5100, loss[loss=0.1358, simple_loss=0.2126, pruned_loss=0.02947, over 4761.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2462, pruned_loss=0.05425, over 955425.34 frames. ], batch size: 28, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:25:20,567 INFO [finetune.py:976] (0/7) Epoch 18, batch 5150, loss[loss=0.1801, simple_loss=0.2366, pruned_loss=0.0618, over 4145.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2464, pruned_loss=0.05462, over 953820.17 frames. ], batch size: 18, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:25:53,892 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.631e+02 1.989e+02 2.441e+02 4.766e+02, threshold=3.977e+02, percent-clipped=3.0 +2023-03-26 22:26:14,502 INFO [finetune.py:976] (0/7) Epoch 18, batch 5200, loss[loss=0.1612, simple_loss=0.2438, pruned_loss=0.03927, over 4898.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2519, pruned_loss=0.05668, over 954589.26 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:26:22,147 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102576.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:26:56,628 INFO [finetune.py:976] (0/7) Epoch 18, batch 5250, loss[loss=0.2207, simple_loss=0.2884, pruned_loss=0.07649, over 4899.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2523, pruned_loss=0.05648, over 952368.08 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:26:58,548 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102624.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:27:00,892 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2201, 1.3836, 1.5633, 1.0032, 1.3275, 1.4206, 1.3618, 1.6188], + device='cuda:0'), covar=tensor([0.1403, 0.2213, 0.1337, 0.1674, 0.1055, 0.1453, 0.3113, 0.0995], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0204, 0.0190, 0.0190, 0.0176, 0.0215, 0.0218, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:27:06,342 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 22:27:11,558 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:27:17,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2670, 1.8786, 2.4508, 4.0365, 2.8059, 2.7078, 0.9551, 3.4306], + device='cuda:0'), covar=tensor([0.1543, 0.1449, 0.1448, 0.0442, 0.0725, 0.1414, 0.1949, 0.0371], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0133, 0.0163, 0.0100, 0.0134, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:27:17,760 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.670e+02 1.981e+02 2.217e+02 4.217e+02, threshold=3.962e+02, percent-clipped=1.0 +2023-03-26 22:27:29,391 INFO [finetune.py:976] (0/7) Epoch 18, batch 5300, loss[loss=0.1538, simple_loss=0.2442, pruned_loss=0.03164, over 4778.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2536, pruned_loss=0.05677, over 954558.94 frames. ], batch size: 29, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:27:44,025 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102691.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:28:03,083 INFO [finetune.py:976] (0/7) Epoch 18, batch 5350, loss[loss=0.1625, simple_loss=0.2396, pruned_loss=0.04272, over 4735.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.253, pruned_loss=0.05594, over 954641.97 frames. ], batch size: 54, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:28:25,303 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.536e+02 1.872e+02 2.228e+02 4.473e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 22:28:34,089 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.76 vs. limit=5.0 +2023-03-26 22:28:36,910 INFO [finetune.py:976] (0/7) Epoch 18, batch 5400, loss[loss=0.1547, simple_loss=0.2213, pruned_loss=0.044, over 4793.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2507, pruned_loss=0.05551, over 954064.67 frames. ], batch size: 26, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:28:44,163 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0087, 1.9850, 2.0156, 2.0744, 1.8102, 4.8218, 1.9967, 2.3243], + device='cuda:0'), covar=tensor([0.3130, 0.2259, 0.1919, 0.2259, 0.1443, 0.0104, 0.2220, 0.1128], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0114, 0.0120, 0.0123, 0.0113, 0.0095, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:28:58,485 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6788, 3.9711, 4.1901, 4.4548, 4.3605, 4.0647, 4.7764, 1.5696], + device='cuda:0'), covar=tensor([0.0863, 0.0938, 0.0951, 0.1123, 0.1605, 0.1787, 0.0691, 0.5796], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0246, 0.0281, 0.0294, 0.0336, 0.0285, 0.0306, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:29:09,136 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0841, 1.5498, 2.1036, 1.9958, 1.8784, 1.8187, 1.9551, 1.9313], + device='cuda:0'), covar=tensor([0.3742, 0.3890, 0.3351, 0.3735, 0.4798, 0.3765, 0.4626, 0.3150], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0240, 0.0259, 0.0276, 0.0274, 0.0248, 0.0283, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:29:10,820 INFO [finetune.py:976] (0/7) Epoch 18, batch 5450, loss[loss=0.1712, simple_loss=0.2322, pruned_loss=0.05512, over 4824.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2472, pruned_loss=0.05437, over 954533.23 frames. ], batch size: 39, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:29:27,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6016, 3.5767, 3.3328, 1.7169, 3.6282, 2.7501, 0.8179, 2.4815], + device='cuda:0'), covar=tensor([0.2590, 0.1912, 0.1825, 0.3347, 0.1264, 0.0974, 0.4487, 0.1559], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0161, 0.0130, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 22:29:30,968 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:29:31,448 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.551e+01 1.511e+02 1.793e+02 2.102e+02 5.113e+02, threshold=3.586e+02, percent-clipped=1.0 +2023-03-26 22:29:44,409 INFO [finetune.py:976] (0/7) Epoch 18, batch 5500, loss[loss=0.1849, simple_loss=0.261, pruned_loss=0.05445, over 4812.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2447, pruned_loss=0.05366, over 953382.32 frames. ], batch size: 51, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:29:53,485 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0954, 1.0020, 1.0125, 0.4629, 0.9787, 1.2267, 1.2461, 0.9960], + device='cuda:0'), covar=tensor([0.0923, 0.0560, 0.0580, 0.0544, 0.0581, 0.0531, 0.0418, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0125, 0.0127, 0.0132, 0.0130, 0.0143, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.1876e-05, 1.1017e-04, 8.9458e-05, 9.0578e-05, 9.3094e-05, 9.3261e-05, + 1.0266e-04, 1.0737e-04], device='cuda:0') +2023-03-26 22:29:54,061 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102887.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:30:12,674 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102912.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:30:18,101 INFO [finetune.py:976] (0/7) Epoch 18, batch 5550, loss[loss=0.1478, simple_loss=0.2311, pruned_loss=0.03227, over 4756.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2457, pruned_loss=0.05308, over 954306.24 frames. ], batch size: 28, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:30:36,208 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:30:39,503 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.559e+02 1.902e+02 2.213e+02 4.520e+02, threshold=3.805e+02, percent-clipped=3.0 +2023-03-26 22:30:50,053 INFO [finetune.py:976] (0/7) Epoch 18, batch 5600, loss[loss=0.1981, simple_loss=0.2809, pruned_loss=0.05761, over 4817.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2488, pruned_loss=0.0539, over 952225.65 frames. ], batch size: 40, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:31:25,746 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5753, 1.4192, 1.9834, 1.2395, 1.7745, 1.8634, 1.4081, 2.0449], + device='cuda:0'), covar=tensor([0.1337, 0.2378, 0.1173, 0.1602, 0.0916, 0.1299, 0.2967, 0.0718], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0204, 0.0190, 0.0190, 0.0176, 0.0215, 0.0218, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:31:42,254 INFO [finetune.py:976] (0/7) Epoch 18, batch 5650, loss[loss=0.166, simple_loss=0.2455, pruned_loss=0.0433, over 4922.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2517, pruned_loss=0.05466, over 951703.76 frames. ], batch size: 38, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:32:03,958 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1197, 1.2320, 1.5423, 1.0177, 1.2409, 1.3962, 1.2665, 1.4946], + device='cuda:0'), covar=tensor([0.0994, 0.1713, 0.1023, 0.1150, 0.0756, 0.0984, 0.2371, 0.0664], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0204, 0.0191, 0.0190, 0.0176, 0.0215, 0.0218, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:32:09,152 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.800e+01 1.506e+02 1.835e+02 2.191e+02 3.638e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-03-26 22:32:19,851 INFO [finetune.py:976] (0/7) Epoch 18, batch 5700, loss[loss=0.1265, simple_loss=0.1923, pruned_loss=0.03037, over 4353.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2472, pruned_loss=0.05409, over 933794.93 frames. ], batch size: 19, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:32:36,439 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-18.pt +2023-03-26 22:32:48,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:32:48,575 INFO [finetune.py:976] (0/7) Epoch 19, batch 0, loss[loss=0.1904, simple_loss=0.2602, pruned_loss=0.06028, over 4817.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2602, pruned_loss=0.06028, over 4817.00 frames. ], batch size: 47, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:32:48,576 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 22:33:03,100 INFO [finetune.py:1010] (0/7) Epoch 19, validation: loss=0.1586, simple_loss=0.2282, pruned_loss=0.04454, over 2265189.00 frames. +2023-03-26 22:33:03,101 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 22:33:25,842 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:33:32,573 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 22:33:38,061 INFO [finetune.py:976] (0/7) Epoch 19, batch 50, loss[loss=0.1967, simple_loss=0.2602, pruned_loss=0.06664, over 4737.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2565, pruned_loss=0.06013, over 214549.87 frames. ], batch size: 54, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:33:40,500 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.779e+01 1.456e+02 1.782e+02 2.150e+02 3.860e+02, threshold=3.565e+02, percent-clipped=1.0 +2023-03-26 22:33:41,937 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 22:33:44,738 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103159.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:07,245 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103193.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:11,663 INFO [finetune.py:976] (0/7) Epoch 19, batch 100, loss[loss=0.1725, simple_loss=0.232, pruned_loss=0.05652, over 4836.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2485, pruned_loss=0.05558, over 380782.58 frames. ], batch size: 47, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:34:17,530 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103207.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:40,705 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103243.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:45,882 INFO [finetune.py:976] (0/7) Epoch 19, batch 150, loss[loss=0.1519, simple_loss=0.2106, pruned_loss=0.04662, over 4835.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2443, pruned_loss=0.05403, over 508787.91 frames. ], batch size: 40, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:34:48,704 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.463e+02 1.787e+02 2.269e+02 3.542e+02, threshold=3.573e+02, percent-clipped=0.0 +2023-03-26 22:35:08,283 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0898, 1.8840, 2.2210, 2.2705, 2.1176, 4.6279, 1.7498, 1.9833], + device='cuda:0'), covar=tensor([0.0812, 0.1682, 0.1045, 0.0862, 0.1335, 0.0216, 0.1413, 0.1641], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:35:19,727 INFO [finetune.py:976] (0/7) Epoch 19, batch 200, loss[loss=0.1972, simple_loss=0.2727, pruned_loss=0.06088, over 4916.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2435, pruned_loss=0.05427, over 606499.65 frames. ], batch size: 36, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:35:22,739 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2160, 2.1913, 1.6383, 2.1868, 2.0635, 1.8535, 2.4162, 2.2291], + device='cuda:0'), covar=tensor([0.1313, 0.2040, 0.2866, 0.2682, 0.2610, 0.1744, 0.3023, 0.1823], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0188, 0.0235, 0.0254, 0.0246, 0.0203, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:35:42,743 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7495, 1.3258, 0.8294, 1.6059, 2.0205, 1.3018, 1.4862, 1.6788], + device='cuda:0'), covar=tensor([0.1456, 0.2033, 0.1985, 0.1167, 0.2032, 0.1915, 0.1456, 0.1870], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 22:35:53,176 INFO [finetune.py:976] (0/7) Epoch 19, batch 250, loss[loss=0.1437, simple_loss=0.2216, pruned_loss=0.03293, over 4901.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2458, pruned_loss=0.05411, over 684095.65 frames. ], batch size: 32, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:35:56,516 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.389e+01 1.572e+02 1.886e+02 2.263e+02 4.128e+02, threshold=3.772e+02, percent-clipped=1.0 +2023-03-26 22:36:05,017 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103366.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:36:06,879 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6390, 1.1464, 0.8991, 1.4287, 1.9796, 1.0377, 1.2876, 1.4811], + device='cuda:0'), covar=tensor([0.1543, 0.2207, 0.1939, 0.1267, 0.2034, 0.2108, 0.1608, 0.1977], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 22:36:16,991 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-03-26 22:36:25,723 INFO [finetune.py:976] (0/7) Epoch 19, batch 300, loss[loss=0.1922, simple_loss=0.2508, pruned_loss=0.06686, over 4768.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2493, pruned_loss=0.05461, over 743709.31 frames. ], batch size: 26, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:36:50,083 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0063, 1.8615, 1.6260, 1.7628, 1.7278, 1.7287, 1.8471, 2.4462], + device='cuda:0'), covar=tensor([0.3731, 0.4427, 0.3234, 0.3569, 0.4030, 0.2379, 0.3523, 0.1660], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0261, 0.0230, 0.0275, 0.0251, 0.0220, 0.0251, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:37:02,180 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103427.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:37:21,696 INFO [finetune.py:976] (0/7) Epoch 19, batch 350, loss[loss=0.1736, simple_loss=0.2326, pruned_loss=0.05732, over 4856.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2527, pruned_loss=0.05671, over 792267.30 frames. ], batch size: 31, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:37:28,081 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.554e+02 1.898e+02 2.403e+02 5.343e+02, threshold=3.796e+02, percent-clipped=4.0 +2023-03-26 22:37:29,299 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:03,040 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103488.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:03,125 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1050, 2.9871, 2.4471, 3.2834, 3.0114, 2.7825, 3.5484, 3.0967], + device='cuda:0'), covar=tensor([0.1152, 0.1935, 0.2844, 0.2113, 0.2222, 0.1432, 0.2401, 0.1576], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0187, 0.0234, 0.0252, 0.0244, 0.0202, 0.0213, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:38:10,181 INFO [finetune.py:976] (0/7) Epoch 19, batch 400, loss[loss=0.1576, simple_loss=0.2125, pruned_loss=0.05137, over 4245.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2536, pruned_loss=0.05716, over 827301.87 frames. ], batch size: 18, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:38:15,638 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:39,517 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:43,097 INFO [finetune.py:976] (0/7) Epoch 19, batch 450, loss[loss=0.1504, simple_loss=0.2128, pruned_loss=0.04399, over 4791.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2526, pruned_loss=0.05687, over 855460.85 frames. ], batch size: 51, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:38:45,991 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.312e+01 1.502e+02 1.696e+02 2.061e+02 2.854e+02, threshold=3.392e+02, percent-clipped=0.0 +2023-03-26 22:38:47,235 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:39:19,602 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:39:24,372 INFO [finetune.py:976] (0/7) Epoch 19, batch 500, loss[loss=0.193, simple_loss=0.2586, pruned_loss=0.06373, over 4900.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2505, pruned_loss=0.05622, over 876910.38 frames. ], batch size: 35, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:39:52,424 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 22:39:57,690 INFO [finetune.py:976] (0/7) Epoch 19, batch 550, loss[loss=0.1934, simple_loss=0.2564, pruned_loss=0.06518, over 4907.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2478, pruned_loss=0.05536, over 894158.29 frames. ], batch size: 43, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:40:00,581 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.543e+02 1.834e+02 2.179e+02 4.966e+02, threshold=3.668e+02, percent-clipped=2.0 +2023-03-26 22:40:12,635 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-03-26 22:40:31,016 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 22:40:31,340 INFO [finetune.py:976] (0/7) Epoch 19, batch 600, loss[loss=0.1881, simple_loss=0.2644, pruned_loss=0.05586, over 4745.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2488, pruned_loss=0.05602, over 907547.19 frames. ], batch size: 59, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:40:39,162 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0624, 1.9885, 1.9707, 1.3726, 1.9380, 2.0298, 2.1779, 1.6260], + device='cuda:0'), covar=tensor([0.0544, 0.0573, 0.0652, 0.0870, 0.0682, 0.0655, 0.0524, 0.1119], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0120, 0.0124, 0.0138, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:40:44,869 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4605, 1.2813, 1.2950, 1.4515, 1.6728, 1.5711, 1.3313, 1.3001], + device='cuda:0'), covar=tensor([0.0328, 0.0292, 0.0646, 0.0302, 0.0221, 0.0384, 0.0317, 0.0345], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0106, 0.0141, 0.0109, 0.0098, 0.0107, 0.0098, 0.0109], + device='cuda:0'), out_proj_covar=tensor([7.3230e-05, 8.1873e-05, 1.1085e-04, 8.3480e-05, 7.6208e-05, 7.8719e-05, + 7.2846e-05, 8.3005e-05], device='cuda:0') +2023-03-26 22:40:47,254 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103722.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:40:58,702 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:04,044 INFO [finetune.py:976] (0/7) Epoch 19, batch 650, loss[loss=0.1236, simple_loss=0.1913, pruned_loss=0.02797, over 4828.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2499, pruned_loss=0.05611, over 918760.78 frames. ], batch size: 25, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:41:06,466 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.539e+02 1.795e+02 2.169e+02 3.837e+02, threshold=3.591e+02, percent-clipped=1.0 +2023-03-26 22:41:07,218 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103754.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:15,342 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:30,823 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:37,406 INFO [finetune.py:976] (0/7) Epoch 19, batch 700, loss[loss=0.1442, simple_loss=0.216, pruned_loss=0.03615, over 4755.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2495, pruned_loss=0.0554, over 928849.31 frames. ], batch size: 27, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:41:38,886 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:39,432 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:02,001 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:13,866 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103836.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:23,983 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5111, 2.3725, 1.9649, 0.9737, 2.1758, 1.9159, 1.8123, 2.0909], + device='cuda:0'), covar=tensor([0.0872, 0.0815, 0.1572, 0.2025, 0.1365, 0.2090, 0.1786, 0.1032], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0200, 0.0183, 0.0211, 0.0208, 0.0222, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:42:26,321 INFO [finetune.py:976] (0/7) Epoch 19, batch 750, loss[loss=0.1551, simple_loss=0.2235, pruned_loss=0.04338, over 4883.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.25, pruned_loss=0.05523, over 932754.99 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:42:33,271 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.573e+02 1.871e+02 2.192e+02 5.260e+02, threshold=3.742e+02, percent-clipped=2.0 +2023-03-26 22:43:03,204 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103876.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:43:15,752 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 22:43:17,416 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9471, 1.8636, 1.5980, 1.6234, 1.6999, 1.6783, 1.7762, 2.3752], + device='cuda:0'), covar=tensor([0.3676, 0.3725, 0.3054, 0.3706, 0.3959, 0.2253, 0.3510, 0.1664], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0230, 0.0276, 0.0252, 0.0220, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:43:22,082 INFO [finetune.py:976] (0/7) Epoch 19, batch 800, loss[loss=0.1781, simple_loss=0.2473, pruned_loss=0.05442, over 4900.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2509, pruned_loss=0.05539, over 938497.35 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:43:48,646 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103937.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:43:55,832 INFO [finetune.py:976] (0/7) Epoch 19, batch 850, loss[loss=0.1793, simple_loss=0.2513, pruned_loss=0.05368, over 4859.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2489, pruned_loss=0.0551, over 938067.84 frames. ], batch size: 44, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:43:58,238 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.499e+02 1.798e+02 2.103e+02 3.961e+02, threshold=3.597e+02, percent-clipped=1.0 +2023-03-26 22:44:25,954 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103990.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:44:30,235 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:44:31,320 INFO [finetune.py:976] (0/7) Epoch 19, batch 900, loss[loss=0.1683, simple_loss=0.2351, pruned_loss=0.05079, over 4841.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2463, pruned_loss=0.05393, over 942712.37 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:44:32,108 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-104000.pt +2023-03-26 22:44:46,727 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104022.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:44:53,205 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.65 vs. limit=2.0 +2023-03-26 22:45:05,988 INFO [finetune.py:976] (0/7) Epoch 19, batch 950, loss[loss=0.204, simple_loss=0.258, pruned_loss=0.07499, over 4911.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2442, pruned_loss=0.05354, over 944889.38 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:45:06,107 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=104049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:07,328 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:08,386 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.542e+01 1.532e+02 1.748e+02 2.079e+02 4.067e+02, threshold=3.497e+02, percent-clipped=1.0 +2023-03-26 22:45:11,530 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:12,090 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8531, 4.0081, 3.7186, 2.0423, 4.0653, 3.0574, 0.9368, 2.8471], + device='cuda:0'), covar=tensor([0.2221, 0.1984, 0.1463, 0.3285, 0.1020, 0.0985, 0.4656, 0.1524], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0130, 0.0162, 0.0125, 0.0149, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 22:45:18,738 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104070.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:45:36,942 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:38,693 INFO [finetune.py:976] (0/7) Epoch 19, batch 1000, loss[loss=0.1355, simple_loss=0.216, pruned_loss=0.02754, over 4777.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2477, pruned_loss=0.0543, over 947436.23 frames. ], batch size: 26, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:45:45,414 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:52,079 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:46:12,343 INFO [finetune.py:976] (0/7) Epoch 19, batch 1050, loss[loss=0.1402, simple_loss=0.2107, pruned_loss=0.03483, over 4776.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2511, pruned_loss=0.05477, over 949723.48 frames. ], batch size: 26, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:46:12,573 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 22:46:14,761 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.591e+02 1.940e+02 2.273e+02 3.456e+02, threshold=3.881e+02, percent-clipped=0.0 +2023-03-26 22:46:53,935 INFO [finetune.py:976] (0/7) Epoch 19, batch 1100, loss[loss=0.189, simple_loss=0.2721, pruned_loss=0.05291, over 4897.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2537, pruned_loss=0.0558, over 951834.51 frames. ], batch size: 46, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:47:16,902 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104232.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:47:29,941 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3756, 2.5667, 2.3146, 1.6759, 2.3689, 2.4976, 2.6885, 2.0300], + device='cuda:0'), covar=tensor([0.0595, 0.0535, 0.0755, 0.0945, 0.0826, 0.0708, 0.0569, 0.1053], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0133, 0.0138, 0.0119, 0.0123, 0.0136, 0.0139, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:47:31,184 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6133, 1.5784, 1.5768, 1.6630, 1.0607, 3.5587, 1.3927, 1.7679], + device='cuda:0'), covar=tensor([0.3355, 0.2579, 0.2149, 0.2378, 0.1853, 0.0172, 0.2645, 0.1333], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0114, 0.0096, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:47:39,155 INFO [finetune.py:976] (0/7) Epoch 19, batch 1150, loss[loss=0.1778, simple_loss=0.2502, pruned_loss=0.0527, over 4811.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2537, pruned_loss=0.05548, over 953450.66 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:47:40,952 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5342, 1.4799, 2.2498, 3.5451, 2.3113, 2.4821, 1.0285, 2.8547], + device='cuda:0'), covar=tensor([0.2067, 0.1609, 0.1367, 0.0550, 0.0879, 0.1455, 0.2050, 0.0547], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0117, 0.0133, 0.0164, 0.0099, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:47:47,032 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.710e+02 1.992e+02 2.366e+02 4.129e+02, threshold=3.984e+02, percent-clipped=1.0 +2023-03-26 22:48:25,523 INFO [finetune.py:976] (0/7) Epoch 19, batch 1200, loss[loss=0.1481, simple_loss=0.2174, pruned_loss=0.03941, over 4848.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2512, pruned_loss=0.05483, over 953111.09 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:48:42,436 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7368, 1.3222, 0.7949, 1.6562, 2.0763, 1.5021, 1.6529, 1.7636], + device='cuda:0'), covar=tensor([0.1540, 0.2044, 0.2067, 0.1162, 0.1994, 0.1956, 0.1327, 0.1914], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0092, 0.0120, 0.0094, 0.0100, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 22:48:51,789 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-26 22:49:05,635 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:07,373 INFO [finetune.py:976] (0/7) Epoch 19, batch 1250, loss[loss=0.1755, simple_loss=0.2441, pruned_loss=0.05346, over 4825.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2481, pruned_loss=0.05367, over 953733.73 frames. ], batch size: 39, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:49:10,323 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.979e+01 1.472e+02 1.754e+02 2.218e+02 4.171e+02, threshold=3.509e+02, percent-clipped=1.0 +2023-03-26 22:49:10,408 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:13,414 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9399, 1.8173, 2.3057, 1.5807, 2.1145, 2.3602, 1.7863, 2.4502], + device='cuda:0'), covar=tensor([0.1370, 0.2000, 0.1585, 0.1898, 0.0948, 0.1333, 0.2709, 0.0867], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0188, 0.0175, 0.0213, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:49:15,252 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8535, 1.5591, 2.3273, 1.5547, 2.0824, 2.3136, 1.5577, 2.3819], + device='cuda:0'), covar=tensor([0.1364, 0.2412, 0.1324, 0.1922, 0.0960, 0.1232, 0.3116, 0.0854], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0188, 0.0175, 0.0213, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:49:39,161 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:41,423 INFO [finetune.py:976] (0/7) Epoch 19, batch 1300, loss[loss=0.1835, simple_loss=0.2445, pruned_loss=0.06128, over 4824.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2455, pruned_loss=0.05302, over 954449.66 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:49:45,738 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104405.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:56,435 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:10,343 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:14,236 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6552, 1.5179, 2.0610, 3.2800, 2.1505, 2.2586, 0.8772, 2.7958], + device='cuda:0'), covar=tensor([0.1566, 0.1383, 0.1227, 0.0546, 0.0779, 0.1654, 0.1756, 0.0419], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0116, 0.0132, 0.0163, 0.0098, 0.0135, 0.0122, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 22:50:14,760 INFO [finetune.py:976] (0/7) Epoch 19, batch 1350, loss[loss=0.2554, simple_loss=0.3197, pruned_loss=0.09555, over 4819.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2469, pruned_loss=0.054, over 957057.44 frames. ], batch size: 39, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:50:17,642 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.617e+02 1.931e+02 2.310e+02 3.973e+02, threshold=3.863e+02, percent-clipped=4.0 +2023-03-26 22:50:29,051 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104469.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:45,908 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7200, 1.7046, 1.4153, 1.7835, 2.2752, 1.8132, 1.7481, 1.3764], + device='cuda:0'), covar=tensor([0.2116, 0.1880, 0.1963, 0.1610, 0.1694, 0.1215, 0.2215, 0.1935], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0195, 0.0245, 0.0188, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:50:48,508 INFO [finetune.py:976] (0/7) Epoch 19, batch 1400, loss[loss=0.1829, simple_loss=0.2697, pruned_loss=0.04808, over 4935.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2514, pruned_loss=0.0551, over 955750.63 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:50:55,605 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8594, 3.8314, 3.7066, 1.8830, 4.0385, 3.1269, 0.8745, 2.8102], + device='cuda:0'), covar=tensor([0.2327, 0.2143, 0.1460, 0.3283, 0.0974, 0.0909, 0.4481, 0.1468], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0162, 0.0130, 0.0162, 0.0124, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 22:51:10,556 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104532.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:51:21,268 INFO [finetune.py:976] (0/7) Epoch 19, batch 1450, loss[loss=0.1915, simple_loss=0.2765, pruned_loss=0.05324, over 4828.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2525, pruned_loss=0.05503, over 955082.28 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:51:24,643 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.650e+02 1.913e+02 2.290e+02 4.485e+02, threshold=3.826e+02, percent-clipped=3.0 +2023-03-26 22:51:42,879 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104580.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:52:02,825 INFO [finetune.py:976] (0/7) Epoch 19, batch 1500, loss[loss=0.1993, simple_loss=0.2727, pruned_loss=0.06297, over 4894.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2533, pruned_loss=0.05536, over 955603.38 frames. ], batch size: 43, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:52:08,993 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 22:52:13,088 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7613, 1.0331, 1.7966, 1.7516, 1.5807, 1.5046, 1.6176, 1.6928], + device='cuda:0'), covar=tensor([0.3783, 0.3727, 0.3147, 0.3392, 0.4298, 0.3495, 0.4132, 0.2922], + device='cuda:0'), in_proj_covar=tensor([0.0251, 0.0242, 0.0260, 0.0278, 0.0276, 0.0250, 0.0286, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:52:17,810 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1218, 1.9321, 1.7882, 1.9732, 1.8416, 1.8495, 1.8917, 2.4750], + device='cuda:0'), covar=tensor([0.3534, 0.3917, 0.3123, 0.3454, 0.4186, 0.2334, 0.3643, 0.1658], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0230, 0.0275, 0.0251, 0.0221, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:52:30,346 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-03-26 22:52:33,841 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:52:35,532 INFO [finetune.py:976] (0/7) Epoch 19, batch 1550, loss[loss=0.2147, simple_loss=0.2795, pruned_loss=0.07491, over 4885.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2539, pruned_loss=0.05541, over 956004.48 frames. ], batch size: 43, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:52:40,198 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.878e+01 1.494e+02 1.864e+02 2.283e+02 3.386e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-26 22:52:40,300 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:26,172 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104694.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:29,188 INFO [finetune.py:976] (0/7) Epoch 19, batch 1600, loss[loss=0.1884, simple_loss=0.2577, pruned_loss=0.05954, over 4916.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2518, pruned_loss=0.05509, over 957985.58 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:53:35,244 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:38,264 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104705.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:46,633 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6996, 1.7104, 1.7683, 0.8838, 1.8870, 2.0437, 2.0162, 1.5337], + device='cuda:0'), covar=tensor([0.0984, 0.0624, 0.0401, 0.0567, 0.0374, 0.0514, 0.0333, 0.0635], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0151, 0.0125, 0.0126, 0.0132, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.1612e-05, 1.0971e-04, 8.9305e-05, 8.9674e-05, 9.2851e-05, 9.2611e-05, + 1.0181e-04, 1.0675e-04], device='cuda:0') +2023-03-26 22:54:01,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8931, 1.8335, 1.6904, 1.8776, 1.4934, 3.8142, 1.8274, 2.2825], + device='cuda:0'), covar=tensor([0.3734, 0.2785, 0.2280, 0.2712, 0.1609, 0.0248, 0.2275, 0.1039], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0114, 0.0096, 0.0096, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:54:11,339 INFO [finetune.py:976] (0/7) Epoch 19, batch 1650, loss[loss=0.1468, simple_loss=0.2206, pruned_loss=0.03649, over 4750.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.249, pruned_loss=0.05405, over 956532.06 frames. ], batch size: 27, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:54:13,774 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.513e+02 1.754e+02 2.121e+02 3.523e+02, threshold=3.508e+02, percent-clipped=0.0 +2023-03-26 22:54:13,847 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:54:19,816 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5996, 1.5402, 1.4906, 1.5121, 1.0404, 2.9945, 1.1398, 1.6037], + device='cuda:0'), covar=tensor([0.3358, 0.2564, 0.2153, 0.2483, 0.1896, 0.0256, 0.2773, 0.1305], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0114, 0.0096, 0.0096, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:54:33,281 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3006, 2.2101, 2.2944, 0.8547, 2.5737, 2.7835, 2.3480, 2.0391], + device='cuda:0'), covar=tensor([0.0948, 0.0700, 0.0442, 0.0731, 0.0483, 0.0590, 0.0422, 0.0716], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0151, 0.0125, 0.0126, 0.0132, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.1495e-05, 1.0960e-04, 8.9196e-05, 8.9379e-05, 9.2834e-05, 9.2410e-05, + 1.0153e-04, 1.0662e-04], device='cuda:0') +2023-03-26 22:54:36,311 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6744, 1.5980, 1.9797, 1.9249, 1.8213, 3.0293, 1.5715, 1.7225], + device='cuda:0'), covar=tensor([0.0834, 0.1618, 0.1335, 0.0849, 0.1236, 0.0311, 0.1257, 0.1362], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:54:41,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0946, 1.8946, 1.6386, 1.6531, 1.8331, 1.7842, 1.8283, 2.4714], + device='cuda:0'), covar=tensor([0.4068, 0.4419, 0.3511, 0.3983, 0.4010, 0.2614, 0.3760, 0.1894], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0260, 0.0230, 0.0274, 0.0250, 0.0220, 0.0251, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:54:44,698 INFO [finetune.py:976] (0/7) Epoch 19, batch 1700, loss[loss=0.1921, simple_loss=0.2606, pruned_loss=0.06178, over 4788.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2463, pruned_loss=0.05342, over 955497.29 frames. ], batch size: 26, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:55:17,897 INFO [finetune.py:976] (0/7) Epoch 19, batch 1750, loss[loss=0.1243, simple_loss=0.2062, pruned_loss=0.02117, over 4757.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2491, pruned_loss=0.05505, over 955461.21 frames. ], batch size: 27, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:55:20,304 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.630e+02 1.888e+02 2.368e+02 5.925e+02, threshold=3.776e+02, percent-clipped=5.0 +2023-03-26 22:55:48,653 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0017, 1.8608, 1.9726, 1.2751, 1.9376, 1.9769, 1.9763, 1.6182], + device='cuda:0'), covar=tensor([0.0508, 0.0653, 0.0643, 0.0857, 0.0781, 0.0608, 0.0556, 0.1192], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0132, 0.0137, 0.0118, 0.0123, 0.0135, 0.0137, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:55:51,603 INFO [finetune.py:976] (0/7) Epoch 19, batch 1800, loss[loss=0.1883, simple_loss=0.266, pruned_loss=0.05531, over 4811.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2491, pruned_loss=0.05395, over 955123.98 frames. ], batch size: 45, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:55:59,084 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-26 22:55:59,680 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-03-26 22:56:13,532 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7230, 1.5603, 1.4367, 0.8283, 1.6595, 1.7672, 1.7556, 1.3996], + device='cuda:0'), covar=tensor([0.0904, 0.0625, 0.0605, 0.0562, 0.0478, 0.0614, 0.0367, 0.0730], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0125, 0.0126, 0.0132, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.1746e-05, 1.1018e-04, 8.9406e-05, 8.9728e-05, 9.3072e-05, 9.2862e-05, + 1.0180e-04, 1.0728e-04], device='cuda:0') +2023-03-26 22:56:22,204 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8861, 1.4022, 0.8615, 1.6671, 2.0669, 1.3974, 1.5785, 1.7584], + device='cuda:0'), covar=tensor([0.1450, 0.2031, 0.2031, 0.1170, 0.1945, 0.1957, 0.1415, 0.1890], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-26 22:56:25,141 INFO [finetune.py:976] (0/7) Epoch 19, batch 1850, loss[loss=0.2009, simple_loss=0.2608, pruned_loss=0.07052, over 4768.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.252, pruned_loss=0.05587, over 954703.31 frames. ], batch size: 26, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:56:27,537 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.864e+01 1.560e+02 1.785e+02 2.312e+02 4.235e+02, threshold=3.569e+02, percent-clipped=1.0 +2023-03-26 22:57:00,525 INFO [finetune.py:976] (0/7) Epoch 19, batch 1900, loss[loss=0.2128, simple_loss=0.2848, pruned_loss=0.0704, over 4818.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2534, pruned_loss=0.05632, over 955719.10 frames. ], batch size: 33, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:57:22,448 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9073, 1.7513, 1.8261, 1.3492, 1.8382, 1.8893, 1.9496, 1.5069], + device='cuda:0'), covar=tensor([0.0477, 0.0623, 0.0610, 0.0747, 0.0717, 0.0571, 0.0484, 0.1084], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0132, 0.0137, 0.0118, 0.0123, 0.0136, 0.0137, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:57:27,677 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0517, 2.0939, 1.5959, 2.1484, 2.0586, 1.8029, 2.5157, 2.1192], + device='cuda:0'), covar=tensor([0.1403, 0.2107, 0.3170, 0.2462, 0.2652, 0.1717, 0.2978, 0.1821], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0188, 0.0236, 0.0253, 0.0246, 0.0203, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:57:42,256 INFO [finetune.py:976] (0/7) Epoch 19, batch 1950, loss[loss=0.1796, simple_loss=0.2507, pruned_loss=0.05421, over 4874.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2514, pruned_loss=0.05572, over 954877.00 frames. ], batch size: 34, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:57:44,664 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.430e+02 1.759e+02 2.099e+02 5.293e+02, threshold=3.517e+02, percent-clipped=3.0 +2023-03-26 22:57:51,470 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7817, 1.4823, 2.1251, 1.4408, 1.9530, 2.1135, 1.4944, 2.2035], + device='cuda:0'), covar=tensor([0.1226, 0.2048, 0.1172, 0.1676, 0.0807, 0.1216, 0.2770, 0.0795], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0203, 0.0190, 0.0187, 0.0174, 0.0212, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:58:31,188 INFO [finetune.py:976] (0/7) Epoch 19, batch 2000, loss[loss=0.1383, simple_loss=0.2097, pruned_loss=0.03341, over 4763.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2487, pruned_loss=0.05455, over 954715.58 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 22:59:17,044 INFO [finetune.py:976] (0/7) Epoch 19, batch 2050, loss[loss=0.193, simple_loss=0.2511, pruned_loss=0.06739, over 4914.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2463, pruned_loss=0.05425, over 956796.36 frames. ], batch size: 37, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 22:59:18,991 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5632, 2.4070, 1.8581, 0.8837, 2.0603, 1.9186, 1.8105, 2.0635], + device='cuda:0'), covar=tensor([0.0902, 0.0732, 0.1835, 0.2316, 0.1545, 0.2458, 0.2296, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0195, 0.0202, 0.0184, 0.0213, 0.0209, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 22:59:19,897 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.557e+02 1.803e+02 2.317e+02 4.729e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-26 22:59:20,010 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2868, 1.3646, 1.5725, 1.5644, 1.4681, 3.0286, 1.2361, 1.4208], + device='cuda:0'), covar=tensor([0.1224, 0.2432, 0.1129, 0.1051, 0.1907, 0.0297, 0.1962, 0.2297], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 22:59:29,328 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-26 22:59:50,002 INFO [finetune.py:976] (0/7) Epoch 19, batch 2100, loss[loss=0.2037, simple_loss=0.2611, pruned_loss=0.0732, over 4856.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2455, pruned_loss=0.05404, over 956601.31 frames. ], batch size: 44, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:00:23,800 INFO [finetune.py:976] (0/7) Epoch 19, batch 2150, loss[loss=0.1938, simple_loss=0.268, pruned_loss=0.05976, over 4903.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2489, pruned_loss=0.0551, over 956093.64 frames. ], batch size: 37, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:00:26,177 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0711, 0.9880, 0.9536, 0.4063, 0.9487, 1.1798, 1.2388, 0.9390], + device='cuda:0'), covar=tensor([0.0871, 0.0579, 0.0659, 0.0568, 0.0637, 0.0679, 0.0437, 0.0722], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0154, 0.0126, 0.0127, 0.0134, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([9.2206e-05, 1.1122e-04, 9.0479e-05, 9.0421e-05, 9.4227e-05, 9.3851e-05, + 1.0287e-04, 1.0740e-04], device='cuda:0') +2023-03-26 23:00:26,649 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.573e+02 1.937e+02 2.406e+02 5.182e+02, threshold=3.875e+02, percent-clipped=4.0 +2023-03-26 23:00:30,881 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4493, 1.4185, 1.9839, 2.8846, 1.9526, 2.2622, 0.9246, 2.4214], + device='cuda:0'), covar=tensor([0.1721, 0.1507, 0.1216, 0.0722, 0.0800, 0.1272, 0.1802, 0.0546], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0162, 0.0098, 0.0134, 0.0122, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:00:57,385 INFO [finetune.py:976] (0/7) Epoch 19, batch 2200, loss[loss=0.1886, simple_loss=0.2606, pruned_loss=0.05831, over 4796.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2513, pruned_loss=0.05627, over 955682.97 frames. ], batch size: 29, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:01:30,610 INFO [finetune.py:976] (0/7) Epoch 19, batch 2250, loss[loss=0.187, simple_loss=0.2612, pruned_loss=0.05639, over 4929.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2524, pruned_loss=0.0564, over 955013.98 frames. ], batch size: 33, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:01:33,460 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.629e+02 1.918e+02 2.372e+02 6.301e+02, threshold=3.835e+02, percent-clipped=3.0 +2023-03-26 23:01:56,208 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:02:03,221 INFO [finetune.py:976] (0/7) Epoch 19, batch 2300, loss[loss=0.2072, simple_loss=0.2794, pruned_loss=0.06751, over 4824.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2538, pruned_loss=0.05679, over 955259.95 frames. ], batch size: 47, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:02:30,761 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5071, 3.4107, 3.1751, 1.5036, 3.5028, 2.5037, 0.8517, 2.3273], + device='cuda:0'), covar=tensor([0.2337, 0.2043, 0.1665, 0.3423, 0.1261, 0.1170, 0.4259, 0.1564], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0177, 0.0160, 0.0129, 0.0160, 0.0123, 0.0147, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:02:45,812 INFO [finetune.py:976] (0/7) Epoch 19, batch 2350, loss[loss=0.1413, simple_loss=0.2213, pruned_loss=0.0306, over 4769.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2515, pruned_loss=0.05659, over 955375.84 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:02:45,946 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105449.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:02:48,224 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.493e+02 1.723e+02 2.054e+02 4.367e+02, threshold=3.447e+02, percent-clipped=1.0 +2023-03-26 23:03:07,759 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6370, 3.7627, 3.5199, 1.7753, 3.8642, 2.6772, 0.9630, 2.6304], + device='cuda:0'), covar=tensor([0.2227, 0.1918, 0.1533, 0.3369, 0.1077, 0.1118, 0.4283, 0.1471], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0177, 0.0160, 0.0129, 0.0160, 0.0123, 0.0147, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:03:10,199 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:03:19,408 INFO [finetune.py:976] (0/7) Epoch 19, batch 2400, loss[loss=0.2038, simple_loss=0.2617, pruned_loss=0.0729, over 4798.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2476, pruned_loss=0.05464, over 955633.31 frames. ], batch size: 51, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:03:23,363 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:04:14,404 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:04:15,991 INFO [finetune.py:976] (0/7) Epoch 19, batch 2450, loss[loss=0.1803, simple_loss=0.2439, pruned_loss=0.05831, over 4860.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2446, pruned_loss=0.05356, over 957574.55 frames. ], batch size: 34, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:04:18,403 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.415e+01 1.457e+02 1.742e+02 2.171e+02 6.143e+02, threshold=3.484e+02, percent-clipped=3.0 +2023-03-26 23:04:20,372 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0462, 1.3013, 1.2454, 1.2667, 1.3840, 2.4384, 1.2098, 1.4498], + device='cuda:0'), covar=tensor([0.1050, 0.1832, 0.1039, 0.0961, 0.1636, 0.0368, 0.1602, 0.1779], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:04:25,637 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:04:29,906 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2511, 2.0805, 1.9058, 2.0293, 2.0242, 2.0683, 2.0735, 2.7140], + device='cuda:0'), covar=tensor([0.3531, 0.4162, 0.3221, 0.3662, 0.3923, 0.2240, 0.3731, 0.1741], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0260, 0.0229, 0.0276, 0.0251, 0.0220, 0.0251, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:04:49,911 INFO [finetune.py:976] (0/7) Epoch 19, batch 2500, loss[loss=0.208, simple_loss=0.2793, pruned_loss=0.06836, over 4910.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2466, pruned_loss=0.05437, over 957690.32 frames. ], batch size: 36, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:05:12,352 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 23:05:23,461 INFO [finetune.py:976] (0/7) Epoch 19, batch 2550, loss[loss=0.1673, simple_loss=0.2449, pruned_loss=0.04482, over 4756.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2502, pruned_loss=0.05517, over 956347.94 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:05:26,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.575e+02 1.924e+02 2.412e+02 4.379e+02, threshold=3.848e+02, percent-clipped=4.0 +2023-03-26 23:05:56,905 INFO [finetune.py:976] (0/7) Epoch 19, batch 2600, loss[loss=0.1657, simple_loss=0.2373, pruned_loss=0.04704, over 4718.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2529, pruned_loss=0.0565, over 957035.61 frames. ], batch size: 23, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:06:04,017 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3234, 1.3966, 1.8704, 1.7325, 1.5395, 3.2996, 1.3323, 1.4676], + device='cuda:0'), covar=tensor([0.1157, 0.1943, 0.1291, 0.1082, 0.1691, 0.0306, 0.1702, 0.2047], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:06:08,110 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0721, 1.8954, 2.0628, 1.6717, 2.0446, 2.2488, 2.1758, 1.4764], + device='cuda:0'), covar=tensor([0.0686, 0.0780, 0.0721, 0.1038, 0.0905, 0.0599, 0.0599, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0133, 0.0138, 0.0119, 0.0123, 0.0136, 0.0138, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:06:12,866 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4570, 1.5424, 1.6030, 0.8734, 1.6082, 1.9262, 1.9171, 1.4235], + device='cuda:0'), covar=tensor([0.0939, 0.0620, 0.0494, 0.0556, 0.0420, 0.0537, 0.0294, 0.0658], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0125, 0.0127, 0.0132, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.1512e-05, 1.1024e-04, 8.9815e-05, 8.9903e-05, 9.3322e-05, 9.3072e-05, + 1.0224e-04, 1.0707e-04], device='cuda:0') +2023-03-26 23:06:27,121 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105744.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:06:30,086 INFO [finetune.py:976] (0/7) Epoch 19, batch 2650, loss[loss=0.2005, simple_loss=0.2744, pruned_loss=0.06337, over 4820.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2535, pruned_loss=0.05633, over 955044.15 frames. ], batch size: 38, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:06:32,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.545e+02 1.903e+02 2.181e+02 3.189e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 23:06:59,540 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105792.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:03,678 INFO [finetune.py:976] (0/7) Epoch 19, batch 2700, loss[loss=0.2042, simple_loss=0.2738, pruned_loss=0.06728, over 4821.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2522, pruned_loss=0.05583, over 955358.83 frames. ], batch size: 39, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:07:03,791 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:28,673 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9290, 1.8093, 1.6194, 1.4622, 1.9945, 1.7077, 1.8272, 1.9521], + device='cuda:0'), covar=tensor([0.1309, 0.1873, 0.2897, 0.2359, 0.2463, 0.1623, 0.2697, 0.1695], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0189, 0.0238, 0.0255, 0.0249, 0.0204, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:07:33,446 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:37,646 INFO [finetune.py:976] (0/7) Epoch 19, batch 2750, loss[loss=0.1801, simple_loss=0.2445, pruned_loss=0.0578, over 3944.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2496, pruned_loss=0.05517, over 952863.59 frames. ], batch size: 17, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:07:40,096 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.803e+01 1.395e+02 1.671e+02 1.966e+02 3.086e+02, threshold=3.343e+02, percent-clipped=0.0 +2023-03-26 23:07:40,223 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105853.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:43,729 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105858.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:45,028 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105860.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:08:22,328 INFO [finetune.py:976] (0/7) Epoch 19, batch 2800, loss[loss=0.1888, simple_loss=0.2503, pruned_loss=0.06365, over 4809.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2468, pruned_loss=0.05429, over 953775.25 frames. ], batch size: 51, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:08:30,237 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0356, 5.0282, 4.7645, 2.8506, 5.1601, 3.9843, 1.0646, 3.7148], + device='cuda:0'), covar=tensor([0.2222, 0.2069, 0.1298, 0.2994, 0.0808, 0.0873, 0.4659, 0.1240], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0161, 0.0130, 0.0162, 0.0124, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:08:35,020 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8474, 4.1168, 3.9345, 1.9035, 4.2067, 3.1602, 1.1047, 3.0649], + device='cuda:0'), covar=tensor([0.2304, 0.2160, 0.1406, 0.3414, 0.1029, 0.0999, 0.4415, 0.1344], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0178, 0.0161, 0.0130, 0.0162, 0.0124, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:09:03,614 INFO [finetune.py:976] (0/7) Epoch 19, batch 2850, loss[loss=0.1878, simple_loss=0.2429, pruned_loss=0.06638, over 4010.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2459, pruned_loss=0.0543, over 953469.61 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:09:10,699 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.025e+01 1.444e+02 1.775e+02 2.176e+02 4.047e+02, threshold=3.549e+02, percent-clipped=5.0 +2023-03-26 23:09:49,344 INFO [finetune.py:976] (0/7) Epoch 19, batch 2900, loss[loss=0.2205, simple_loss=0.289, pruned_loss=0.076, over 4895.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2472, pruned_loss=0.05471, over 953626.36 frames. ], batch size: 35, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:09:50,600 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-106000.pt +2023-03-26 23:10:21,008 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:24,380 INFO [finetune.py:976] (0/7) Epoch 19, batch 2950, loss[loss=0.1651, simple_loss=0.244, pruned_loss=0.04312, over 4874.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2511, pruned_loss=0.05552, over 954417.91 frames. ], batch size: 34, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:10:27,319 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.599e+02 1.865e+02 2.251e+02 4.962e+02, threshold=3.729e+02, percent-clipped=1.0 +2023-03-26 23:10:31,670 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0140, 0.9521, 0.9190, 1.1356, 1.2056, 1.1746, 0.9717, 0.9312], + device='cuda:0'), covar=tensor([0.0317, 0.0323, 0.0666, 0.0304, 0.0258, 0.0386, 0.0374, 0.0394], + device='cuda:0'), in_proj_covar=tensor([0.0096, 0.0107, 0.0143, 0.0110, 0.0099, 0.0110, 0.0100, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4459e-05, 8.2564e-05, 1.1225e-04, 8.4430e-05, 7.7197e-05, 8.0979e-05, + 7.4578e-05, 8.4335e-05], device='cuda:0') +2023-03-26 23:10:43,311 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106077.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:46,267 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1601, 2.0096, 2.0416, 1.3851, 2.0551, 2.0729, 2.0999, 1.7091], + device='cuda:0'), covar=tensor([0.0538, 0.0670, 0.0713, 0.0943, 0.0677, 0.0742, 0.0628, 0.1059], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0134, 0.0139, 0.0120, 0.0123, 0.0138, 0.0139, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:10:53,296 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:57,518 INFO [finetune.py:976] (0/7) Epoch 19, batch 3000, loss[loss=0.1712, simple_loss=0.2463, pruned_loss=0.04807, over 4819.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2524, pruned_loss=0.05556, over 956384.42 frames. ], batch size: 38, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:10:57,519 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 23:11:01,274 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8577, 1.7165, 1.6106, 1.9767, 2.1374, 1.8968, 1.4523, 1.6224], + device='cuda:0'), covar=tensor([0.2150, 0.1910, 0.1892, 0.1618, 0.1535, 0.1224, 0.2212, 0.1919], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0240, 0.0185, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:11:08,360 INFO [finetune.py:1010] (0/7) Epoch 19, validation: loss=0.1576, simple_loss=0.2259, pruned_loss=0.04462, over 2265189.00 frames. +2023-03-26 23:11:08,361 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 23:11:09,088 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9752, 2.3950, 2.2980, 1.2035, 2.5624, 2.0740, 1.7168, 2.1766], + device='cuda:0'), covar=tensor([0.0997, 0.1025, 0.1798, 0.2182, 0.1600, 0.2050, 0.2463, 0.1220], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0192, 0.0199, 0.0182, 0.0210, 0.0207, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:11:38,069 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.91 vs. limit=5.0 +2023-03-26 23:11:43,377 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106138.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:46,233 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:50,365 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:50,874 INFO [finetune.py:976] (0/7) Epoch 19, batch 3050, loss[loss=0.1746, simple_loss=0.2423, pruned_loss=0.05341, over 4749.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2519, pruned_loss=0.05513, over 955721.73 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:11:53,804 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.577e+02 1.927e+02 2.196e+02 3.458e+02, threshold=3.854e+02, percent-clipped=0.0 +2023-03-26 23:11:55,080 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106155.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:56,815 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4613, 1.4729, 1.9291, 1.7646, 1.5311, 3.3229, 1.3313, 1.5446], + device='cuda:0'), covar=tensor([0.0993, 0.1810, 0.1290, 0.0954, 0.1626, 0.0239, 0.1558, 0.1737], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:11:57,413 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:18,599 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:24,004 INFO [finetune.py:976] (0/7) Epoch 19, batch 3100, loss[loss=0.1514, simple_loss=0.223, pruned_loss=0.03988, over 4867.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2484, pruned_loss=0.05381, over 954475.97 frames. ], batch size: 34, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:12:29,245 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:57,684 INFO [finetune.py:976] (0/7) Epoch 19, batch 3150, loss[loss=0.1618, simple_loss=0.2223, pruned_loss=0.05059, over 4814.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2466, pruned_loss=0.0544, over 952308.14 frames. ], batch size: 25, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:13:00,116 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.828e+01 1.675e+02 1.879e+02 2.192e+02 3.916e+02, threshold=3.758e+02, percent-clipped=1.0 +2023-03-26 23:13:17,032 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:13:41,053 INFO [finetune.py:976] (0/7) Epoch 19, batch 3200, loss[loss=0.1652, simple_loss=0.231, pruned_loss=0.04968, over 4714.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2438, pruned_loss=0.05333, over 954903.28 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:13:58,434 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7583, 1.2380, 0.8526, 1.5561, 2.0519, 1.0471, 1.3344, 1.5542], + device='cuda:0'), covar=tensor([0.1330, 0.1947, 0.1827, 0.1097, 0.1913, 0.1957, 0.1470, 0.1769], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:14:00,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9046, 1.8294, 1.7670, 2.1255, 2.3785, 2.1390, 1.7036, 1.5540], + device='cuda:0'), covar=tensor([0.2042, 0.1937, 0.1736, 0.1481, 0.1637, 0.1071, 0.2241, 0.1912], + device='cuda:0'), in_proj_covar=tensor([0.0241, 0.0209, 0.0211, 0.0192, 0.0241, 0.0186, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:14:01,442 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:14:05,717 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:14:06,931 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3656, 1.6201, 0.8147, 2.0384, 2.5253, 1.8058, 1.7659, 2.0462], + device='cuda:0'), covar=tensor([0.1266, 0.1913, 0.2055, 0.1115, 0.1720, 0.1875, 0.1410, 0.1852], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0090, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:14:06,940 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6933, 1.5915, 2.1710, 1.9127, 1.8551, 4.2620, 1.5458, 1.7860], + device='cuda:0'), covar=tensor([0.0913, 0.1705, 0.1188, 0.0913, 0.1500, 0.0182, 0.1496, 0.1684], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:14:16,902 INFO [finetune.py:976] (0/7) Epoch 19, batch 3250, loss[loss=0.1366, simple_loss=0.1928, pruned_loss=0.04015, over 4183.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2442, pruned_loss=0.05391, over 953873.67 frames. ], batch size: 18, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:14:24,769 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.520e+02 1.839e+02 2.222e+02 4.428e+02, threshold=3.677e+02, percent-clipped=2.0 +2023-03-26 23:15:08,382 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:10,069 INFO [finetune.py:976] (0/7) Epoch 19, batch 3300, loss[loss=0.1875, simple_loss=0.2709, pruned_loss=0.05209, over 4851.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2496, pruned_loss=0.05526, over 954568.64 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:15:26,436 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 23:15:36,268 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1754, 2.1985, 1.7157, 2.1278, 2.0819, 1.8136, 2.4449, 2.1882], + device='cuda:0'), covar=tensor([0.1339, 0.1959, 0.2821, 0.2573, 0.2528, 0.1683, 0.2898, 0.1712], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0187, 0.0234, 0.0253, 0.0247, 0.0203, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:15:41,456 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106433.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:51,012 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:51,546 INFO [finetune.py:976] (0/7) Epoch 19, batch 3350, loss[loss=0.2058, simple_loss=0.2757, pruned_loss=0.06793, over 4903.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2514, pruned_loss=0.05553, over 954124.50 frames. ], batch size: 36, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:15:54,457 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.616e+02 1.883e+02 2.222e+02 4.657e+02, threshold=3.766e+02, percent-clipped=2.0 +2023-03-26 23:15:55,202 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:55,758 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:08,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5978, 3.6301, 3.4069, 1.6457, 3.8208, 2.8421, 0.7140, 2.5343], + device='cuda:0'), covar=tensor([0.2381, 0.2064, 0.1667, 0.3394, 0.1172, 0.0963, 0.4500, 0.1419], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0176, 0.0159, 0.0128, 0.0159, 0.0122, 0.0146, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:16:31,244 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106496.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:33,555 INFO [finetune.py:976] (0/7) Epoch 19, batch 3400, loss[loss=0.1949, simple_loss=0.2635, pruned_loss=0.06318, over 4817.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2519, pruned_loss=0.05564, over 954171.54 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:16:36,070 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106503.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:39,124 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-03-26 23:16:44,434 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:17:03,897 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2194, 2.0683, 1.7727, 1.9592, 1.9470, 1.9587, 2.0417, 2.6715], + device='cuda:0'), covar=tensor([0.3467, 0.4218, 0.3089, 0.3686, 0.3843, 0.2364, 0.3619, 0.1667], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0230, 0.0275, 0.0252, 0.0221, 0.0251, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:17:06,742 INFO [finetune.py:976] (0/7) Epoch 19, batch 3450, loss[loss=0.1744, simple_loss=0.2429, pruned_loss=0.05295, over 4897.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2511, pruned_loss=0.0552, over 954053.28 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:17:09,628 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.792e+01 1.525e+02 1.781e+02 2.060e+02 3.433e+02, threshold=3.562e+02, percent-clipped=0.0 +2023-03-26 23:17:13,284 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7028, 1.6633, 1.5432, 1.6800, 1.2623, 3.5190, 1.3842, 1.9168], + device='cuda:0'), covar=tensor([0.3250, 0.2411, 0.2092, 0.2349, 0.1666, 0.0200, 0.2448, 0.1181], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0114, 0.0119, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:17:40,385 INFO [finetune.py:976] (0/7) Epoch 19, batch 3500, loss[loss=0.2267, simple_loss=0.2681, pruned_loss=0.09258, over 4825.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.249, pruned_loss=0.05518, over 954763.80 frames. ], batch size: 25, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:17:57,142 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:18:14,144 INFO [finetune.py:976] (0/7) Epoch 19, batch 3550, loss[loss=0.1777, simple_loss=0.2347, pruned_loss=0.0604, over 4831.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2462, pruned_loss=0.05432, over 955001.67 frames. ], batch size: 33, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:18:16,539 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.567e+02 1.861e+02 2.307e+02 3.604e+02, threshold=3.722e+02, percent-clipped=2.0 +2023-03-26 23:18:51,970 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:18:55,732 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 23:18:56,156 INFO [finetune.py:976] (0/7) Epoch 19, batch 3600, loss[loss=0.2254, simple_loss=0.2906, pruned_loss=0.08015, over 4861.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2447, pruned_loss=0.05401, over 954527.93 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:19:19,195 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106733.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:19:29,858 INFO [finetune.py:976] (0/7) Epoch 19, batch 3650, loss[loss=0.2235, simple_loss=0.2987, pruned_loss=0.07418, over 4906.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2489, pruned_loss=0.05558, over 956214.80 frames. ], batch size: 35, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:19:34,982 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.609e+02 2.013e+02 2.438e+02 4.457e+02, threshold=4.025e+02, percent-clipped=1.0 +2023-03-26 23:20:03,525 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:20:23,369 INFO [finetune.py:976] (0/7) Epoch 19, batch 3700, loss[loss=0.2063, simple_loss=0.2764, pruned_loss=0.06811, over 4926.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2519, pruned_loss=0.05591, over 957182.19 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:20:32,841 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:20:48,432 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8789, 1.4624, 1.9342, 1.9536, 1.6696, 1.6485, 1.8713, 1.7384], + device='cuda:0'), covar=tensor([0.3825, 0.3710, 0.3238, 0.3480, 0.4692, 0.3737, 0.4256, 0.3114], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0239, 0.0259, 0.0277, 0.0275, 0.0249, 0.0284, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:20:59,510 INFO [finetune.py:976] (0/7) Epoch 19, batch 3750, loss[loss=0.1065, simple_loss=0.1673, pruned_loss=0.02279, over 4184.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2521, pruned_loss=0.05563, over 954348.16 frames. ], batch size: 17, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:21:06,524 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.604e+02 1.833e+02 2.350e+02 4.465e+02, threshold=3.666e+02, percent-clipped=2.0 +2023-03-26 23:21:33,691 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4673, 1.3139, 1.6851, 2.4735, 1.7268, 2.2135, 0.9258, 2.1579], + device='cuda:0'), covar=tensor([0.1631, 0.1454, 0.1082, 0.0636, 0.0858, 0.1133, 0.1553, 0.0570], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0133, 0.0163, 0.0099, 0.0135, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:21:48,083 INFO [finetune.py:976] (0/7) Epoch 19, batch 3800, loss[loss=0.1953, simple_loss=0.246, pruned_loss=0.07236, over 3992.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2527, pruned_loss=0.056, over 953481.44 frames. ], batch size: 17, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:22:02,503 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 23:22:07,681 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:16,657 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 23:22:23,050 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5834, 3.7936, 3.6919, 1.7748, 3.8692, 2.8969, 0.7296, 2.7111], + device='cuda:0'), covar=tensor([0.2429, 0.1838, 0.1377, 0.3201, 0.1090, 0.0992, 0.4528, 0.1296], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0179, 0.0162, 0.0131, 0.0162, 0.0124, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:22:24,702 INFO [finetune.py:976] (0/7) Epoch 19, batch 3850, loss[loss=0.1492, simple_loss=0.2181, pruned_loss=0.04018, over 4813.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2509, pruned_loss=0.05514, over 954620.26 frames. ], batch size: 40, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:22:27,157 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.623e+02 1.818e+02 2.255e+02 6.115e+02, threshold=3.637e+02, percent-clipped=1.0 +2023-03-26 23:22:39,183 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:52,753 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:57,317 INFO [finetune.py:976] (0/7) Epoch 19, batch 3900, loss[loss=0.2206, simple_loss=0.2742, pruned_loss=0.08352, over 4901.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.247, pruned_loss=0.05385, over 954288.77 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:23:24,072 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:23:24,084 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5922, 3.7164, 3.5489, 1.7737, 3.7818, 2.8223, 0.9747, 2.6497], + device='cuda:0'), covar=tensor([0.2870, 0.2151, 0.1495, 0.3513, 0.1110, 0.1130, 0.4599, 0.1527], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0161, 0.0130, 0.0162, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:23:29,941 INFO [finetune.py:976] (0/7) Epoch 19, batch 3950, loss[loss=0.1645, simple_loss=0.2363, pruned_loss=0.04635, over 4856.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2443, pruned_loss=0.05316, over 956835.59 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:23:35,024 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.519e+02 1.802e+02 2.250e+02 5.271e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-26 23:24:12,977 INFO [finetune.py:976] (0/7) Epoch 19, batch 4000, loss[loss=0.1796, simple_loss=0.2522, pruned_loss=0.05345, over 4917.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2457, pruned_loss=0.05433, over 955364.79 frames. ], batch size: 37, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:24:21,397 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:24:41,078 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107140.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:24:46,974 INFO [finetune.py:976] (0/7) Epoch 19, batch 4050, loss[loss=0.1728, simple_loss=0.2408, pruned_loss=0.05243, over 4760.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2496, pruned_loss=0.0556, over 955055.23 frames. ], batch size: 28, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:24:48,808 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107152.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:24:49,868 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.579e+02 1.895e+02 2.231e+02 3.900e+02, threshold=3.790e+02, percent-clipped=1.0 +2023-03-26 23:24:52,895 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:25:40,274 INFO [finetune.py:976] (0/7) Epoch 19, batch 4100, loss[loss=0.1457, simple_loss=0.2113, pruned_loss=0.03999, over 3797.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2509, pruned_loss=0.05544, over 952550.27 frames. ], batch size: 16, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:25:41,790 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107201.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:25:50,499 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107213.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:26:12,013 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 23:26:13,432 INFO [finetune.py:976] (0/7) Epoch 19, batch 4150, loss[loss=0.196, simple_loss=0.259, pruned_loss=0.06648, over 4799.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2513, pruned_loss=0.05542, over 953338.74 frames. ], batch size: 25, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:26:15,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8766, 1.6764, 2.2665, 1.3325, 2.0818, 2.1738, 1.5704, 2.2859], + device='cuda:0'), covar=tensor([0.1290, 0.1987, 0.1373, 0.1955, 0.0867, 0.1209, 0.2737, 0.0765], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0190, 0.0189, 0.0174, 0.0213, 0.0218, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:26:21,809 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.570e+02 1.970e+02 2.461e+02 5.293e+02, threshold=3.939e+02, percent-clipped=1.0 +2023-03-26 23:26:35,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2194, 2.1902, 1.6911, 2.2885, 2.1255, 1.8521, 2.5447, 2.2146], + device='cuda:0'), covar=tensor([0.1387, 0.2046, 0.2985, 0.2476, 0.2500, 0.1620, 0.2862, 0.1733], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0187, 0.0234, 0.0251, 0.0246, 0.0202, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:26:42,566 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8545, 1.8804, 1.6778, 1.9086, 1.5062, 4.6266, 1.8746, 2.4549], + device='cuda:0'), covar=tensor([0.3206, 0.2416, 0.2048, 0.2135, 0.1621, 0.0140, 0.2182, 0.1014], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0115, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:26:56,743 INFO [finetune.py:976] (0/7) Epoch 19, batch 4200, loss[loss=0.2451, simple_loss=0.2961, pruned_loss=0.09705, over 4146.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2517, pruned_loss=0.05503, over 952930.48 frames. ], batch size: 65, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:27:11,429 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3378, 1.2540, 1.8297, 2.7619, 1.7187, 2.1774, 0.8592, 2.3987], + device='cuda:0'), covar=tensor([0.2063, 0.1958, 0.1519, 0.1042, 0.1151, 0.1610, 0.2040, 0.0679], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0114, 0.0132, 0.0162, 0.0099, 0.0134, 0.0122, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:27:29,943 INFO [finetune.py:976] (0/7) Epoch 19, batch 4250, loss[loss=0.1919, simple_loss=0.263, pruned_loss=0.06038, over 4808.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2507, pruned_loss=0.05503, over 953578.36 frames. ], batch size: 51, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:27:33,460 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.503e+02 1.795e+02 2.146e+02 3.676e+02, threshold=3.590e+02, percent-clipped=0.0 +2023-03-26 23:28:03,397 INFO [finetune.py:976] (0/7) Epoch 19, batch 4300, loss[loss=0.1466, simple_loss=0.2274, pruned_loss=0.03288, over 4776.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2482, pruned_loss=0.05424, over 953474.13 frames. ], batch size: 26, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:28:21,360 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 23:28:30,302 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 23:28:34,471 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8747, 1.7663, 1.8998, 1.3545, 1.8374, 1.9218, 1.9008, 1.5347], + device='cuda:0'), covar=tensor([0.0459, 0.0518, 0.0553, 0.0778, 0.0820, 0.0493, 0.0449, 0.0980], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0121, 0.0125, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:28:36,190 INFO [finetune.py:976] (0/7) Epoch 19, batch 4350, loss[loss=0.1688, simple_loss=0.2438, pruned_loss=0.04694, over 4923.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2446, pruned_loss=0.05285, over 952651.69 frames. ], batch size: 38, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:28:40,179 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.532e+02 1.813e+02 2.231e+02 3.395e+02, threshold=3.625e+02, percent-clipped=1.0 +2023-03-26 23:29:00,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1522, 2.0465, 1.6383, 2.1365, 2.0421, 1.8493, 2.4005, 2.1338], + device='cuda:0'), covar=tensor([0.1379, 0.2108, 0.3045, 0.2500, 0.2599, 0.1679, 0.3323, 0.1854], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0189, 0.0236, 0.0254, 0.0248, 0.0203, 0.0216, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:29:21,298 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107496.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:29:23,019 INFO [finetune.py:976] (0/7) Epoch 19, batch 4400, loss[loss=0.2165, simple_loss=0.2808, pruned_loss=0.07608, over 4829.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2451, pruned_loss=0.05294, over 952615.25 frames. ], batch size: 33, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:29:29,605 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107508.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:29:31,354 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107510.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:29:56,824 INFO [finetune.py:976] (0/7) Epoch 19, batch 4450, loss[loss=0.1377, simple_loss=0.2041, pruned_loss=0.03562, over 4202.00 frames. ], tot_loss[loss=0.179, simple_loss=0.249, pruned_loss=0.05446, over 953070.59 frames. ], batch size: 18, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:29:59,905 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.983e+01 1.601e+02 1.972e+02 2.467e+02 3.942e+02, threshold=3.944e+02, percent-clipped=4.0 +2023-03-26 23:30:12,269 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:30:42,901 INFO [finetune.py:976] (0/7) Epoch 19, batch 4500, loss[loss=0.1578, simple_loss=0.2265, pruned_loss=0.04453, over 4901.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2508, pruned_loss=0.05469, over 952445.90 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:31:21,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5578, 1.6118, 1.2338, 1.4871, 1.8382, 1.8062, 1.4947, 1.4126], + device='cuda:0'), covar=tensor([0.0314, 0.0292, 0.0629, 0.0306, 0.0222, 0.0447, 0.0349, 0.0406], + device='cuda:0'), in_proj_covar=tensor([0.0096, 0.0106, 0.0143, 0.0111, 0.0099, 0.0109, 0.0099, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4500e-05, 8.1736e-05, 1.1264e-04, 8.5083e-05, 7.6743e-05, 8.0973e-05, + 7.3566e-05, 8.4085e-05], device='cuda:0') +2023-03-26 23:31:24,169 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3311, 1.9013, 2.1911, 2.1550, 1.9454, 1.9667, 2.1129, 2.0228], + device='cuda:0'), covar=tensor([0.4173, 0.4101, 0.3637, 0.4076, 0.5803, 0.4236, 0.5227, 0.3296], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0239, 0.0259, 0.0278, 0.0275, 0.0250, 0.0284, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:31:25,197 INFO [finetune.py:976] (0/7) Epoch 19, batch 4550, loss[loss=0.1538, simple_loss=0.234, pruned_loss=0.03678, over 4888.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2513, pruned_loss=0.05474, over 950805.82 frames. ], batch size: 32, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:31:28,198 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.552e+02 1.832e+02 2.186e+02 5.352e+02, threshold=3.664e+02, percent-clipped=1.0 +2023-03-26 23:31:31,368 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:12,109 INFO [finetune.py:976] (0/7) Epoch 19, batch 4600, loss[loss=0.1684, simple_loss=0.2391, pruned_loss=0.04881, over 4758.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2504, pruned_loss=0.05453, over 952391.56 frames. ], batch size: 26, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:32:26,322 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:29,334 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5679, 1.5321, 1.9475, 2.9250, 1.9583, 2.2883, 0.9567, 2.4976], + device='cuda:0'), covar=tensor([0.1652, 0.1322, 0.1239, 0.0592, 0.0859, 0.1227, 0.1770, 0.0501], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0162, 0.0099, 0.0135, 0.0123, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:32:37,553 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:45,689 INFO [finetune.py:976] (0/7) Epoch 19, batch 4650, loss[loss=0.1919, simple_loss=0.2564, pruned_loss=0.06373, over 4810.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2479, pruned_loss=0.05357, over 951996.72 frames. ], batch size: 45, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:32:48,738 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.855e+01 1.504e+02 1.713e+02 2.086e+02 4.043e+02, threshold=3.426e+02, percent-clipped=2.0 +2023-03-26 23:32:58,469 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8104, 1.7166, 2.2233, 1.5491, 2.0556, 2.1152, 1.6308, 2.3063], + device='cuda:0'), covar=tensor([0.1281, 0.1829, 0.1404, 0.1748, 0.0849, 0.1329, 0.2637, 0.0777], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0206, 0.0192, 0.0191, 0.0176, 0.0215, 0.0220, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:33:17,162 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107796.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:33:18,414 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107798.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:33:19,351 INFO [finetune.py:976] (0/7) Epoch 19, batch 4700, loss[loss=0.1959, simple_loss=0.2633, pruned_loss=0.06427, over 4773.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.245, pruned_loss=0.05308, over 952206.59 frames. ], batch size: 28, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:33:25,039 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107808.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:33:45,697 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8321, 1.2736, 0.9770, 1.7226, 2.0003, 1.5009, 1.4921, 1.7198], + device='cuda:0'), covar=tensor([0.1393, 0.1901, 0.1923, 0.1097, 0.2078, 0.2036, 0.1348, 0.1755], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0090, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:33:49,094 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107844.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:33:54,086 INFO [finetune.py:976] (0/7) Epoch 19, batch 4750, loss[loss=0.1395, simple_loss=0.2066, pruned_loss=0.03618, over 4298.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.243, pruned_loss=0.05265, over 952930.84 frames. ], batch size: 18, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:33:56,042 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2994, 2.2492, 1.8184, 2.4580, 2.2476, 2.0093, 2.7507, 2.4004], + device='cuda:0'), covar=tensor([0.1394, 0.2494, 0.3171, 0.2841, 0.2630, 0.1622, 0.3320, 0.1761], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0187, 0.0235, 0.0252, 0.0246, 0.0202, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:33:57,613 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.438e+02 1.688e+02 2.143e+02 3.806e+02, threshold=3.376e+02, percent-clipped=2.0 +2023-03-26 23:33:58,877 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107856.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:34:04,986 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107866.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:34:15,855 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-26 23:34:37,230 INFO [finetune.py:976] (0/7) Epoch 19, batch 4800, loss[loss=0.2298, simple_loss=0.3055, pruned_loss=0.07712, over 4822.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2471, pruned_loss=0.05401, over 952661.59 frames. ], batch size: 39, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:35:09,129 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2304, 2.8637, 2.6784, 1.4501, 2.7474, 2.2424, 2.2284, 2.5065], + device='cuda:0'), covar=tensor([0.0765, 0.0792, 0.1737, 0.2078, 0.1611, 0.2283, 0.1917, 0.1231], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0195, 0.0201, 0.0184, 0.0211, 0.0209, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:35:10,747 INFO [finetune.py:976] (0/7) Epoch 19, batch 4850, loss[loss=0.1965, simple_loss=0.2674, pruned_loss=0.06282, over 4920.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2503, pruned_loss=0.05446, over 954533.25 frames. ], batch size: 42, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:35:13,742 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.610e+02 1.895e+02 2.225e+02 4.035e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-03-26 23:35:17,877 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1066, 1.3070, 1.4609, 1.3026, 1.4305, 2.5661, 1.1380, 1.3702], + device='cuda:0'), covar=tensor([0.1242, 0.2289, 0.1208, 0.1089, 0.1931, 0.0415, 0.2021, 0.2197], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:35:45,817 INFO [finetune.py:976] (0/7) Epoch 19, batch 4900, loss[loss=0.1661, simple_loss=0.2467, pruned_loss=0.04277, over 4893.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2528, pruned_loss=0.05552, over 956147.86 frames. ], batch size: 43, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:35:46,590 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-108000.pt +2023-03-26 23:35:57,393 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108015.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:35:57,407 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6404, 1.1294, 0.9056, 1.5312, 1.9368, 1.0552, 1.3735, 1.5061], + device='cuda:0'), covar=tensor([0.1490, 0.2042, 0.1879, 0.1145, 0.2074, 0.1996, 0.1415, 0.1837], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0090, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:36:28,167 INFO [finetune.py:976] (0/7) Epoch 19, batch 4950, loss[loss=0.1452, simple_loss=0.2252, pruned_loss=0.03257, over 4334.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2514, pruned_loss=0.05425, over 953440.64 frames. ], batch size: 65, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:36:31,629 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.572e+02 1.807e+02 2.323e+02 4.539e+02, threshold=3.614e+02, percent-clipped=1.0 +2023-03-26 23:36:52,767 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.20 vs. limit=5.0 +2023-03-26 23:37:04,010 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108093.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:37:10,956 INFO [finetune.py:976] (0/7) Epoch 19, batch 5000, loss[loss=0.1549, simple_loss=0.2243, pruned_loss=0.04273, over 4897.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2505, pruned_loss=0.05366, over 955017.97 frames. ], batch size: 35, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:37:45,910 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1847, 1.9656, 1.6288, 1.7160, 1.9402, 1.9199, 1.9477, 2.5917], + device='cuda:0'), covar=tensor([0.3668, 0.3626, 0.3199, 0.3789, 0.4297, 0.2480, 0.3279, 0.2018], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0229, 0.0275, 0.0251, 0.0221, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:37:54,036 INFO [finetune.py:976] (0/7) Epoch 19, batch 5050, loss[loss=0.164, simple_loss=0.2347, pruned_loss=0.0467, over 4757.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2483, pruned_loss=0.05382, over 955124.98 frames. ], batch size: 26, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:37:55,870 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4736, 1.4460, 1.6152, 0.7459, 1.6664, 1.9814, 1.7854, 1.4551], + device='cuda:0'), covar=tensor([0.1043, 0.1025, 0.0672, 0.0708, 0.0515, 0.0647, 0.0427, 0.0886], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0124, 0.0124, 0.0131, 0.0128, 0.0141, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.0300e-05, 1.0819e-04, 8.8740e-05, 8.8141e-05, 9.2000e-05, 9.1518e-05, + 1.0124e-04, 1.0533e-04], device='cuda:0') +2023-03-26 23:37:57,572 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.577e+02 1.863e+02 2.132e+02 3.762e+02, threshold=3.725e+02, percent-clipped=1.0 +2023-03-26 23:38:05,857 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:38:27,769 INFO [finetune.py:976] (0/7) Epoch 19, batch 5100, loss[loss=0.1399, simple_loss=0.2133, pruned_loss=0.03327, over 4929.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2444, pruned_loss=0.05272, over 954105.39 frames. ], batch size: 38, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:38:32,756 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7937, 1.7578, 1.7949, 1.2971, 1.7971, 1.9712, 1.8746, 1.4998], + device='cuda:0'), covar=tensor([0.0427, 0.0520, 0.0579, 0.0797, 0.0952, 0.0400, 0.0458, 0.0948], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0120, 0.0124, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:38:38,021 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108214.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:38:38,092 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9891, 2.1450, 1.8905, 1.7407, 2.5708, 2.7691, 2.0528, 1.9991], + device='cuda:0'), covar=tensor([0.0401, 0.0392, 0.0537, 0.0390, 0.0233, 0.0437, 0.0362, 0.0466], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0108, 0.0145, 0.0112, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5299e-05, 8.3108e-05, 1.1403e-04, 8.5995e-05, 7.7695e-05, 8.2097e-05, + 7.4319e-05, 8.5362e-05], device='cuda:0') +2023-03-26 23:39:00,757 INFO [finetune.py:976] (0/7) Epoch 19, batch 5150, loss[loss=0.1736, simple_loss=0.2438, pruned_loss=0.05165, over 4724.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2448, pruned_loss=0.05293, over 953503.39 frames. ], batch size: 23, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:39:04,795 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.278e+01 1.451e+02 1.873e+02 2.231e+02 4.201e+02, threshold=3.747e+02, percent-clipped=0.0 +2023-03-26 23:39:23,383 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2000, 2.1426, 2.1725, 1.5772, 2.0855, 2.3333, 2.2562, 1.7422], + device='cuda:0'), covar=tensor([0.0563, 0.0647, 0.0688, 0.0855, 0.0694, 0.0671, 0.0565, 0.1195], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0134, 0.0139, 0.0120, 0.0125, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:39:23,956 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7203, 3.7621, 3.5101, 1.7257, 3.8778, 2.9802, 0.7985, 2.6868], + device='cuda:0'), covar=tensor([0.2120, 0.1983, 0.1517, 0.3432, 0.1019, 0.0927, 0.4432, 0.1389], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0174, 0.0157, 0.0128, 0.0158, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:39:39,542 INFO [finetune.py:976] (0/7) Epoch 19, batch 5200, loss[loss=0.2073, simple_loss=0.2937, pruned_loss=0.06048, over 4824.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2472, pruned_loss=0.05392, over 953142.69 frames. ], batch size: 39, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:39:53,922 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:01,013 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108325.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:13,097 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4927, 2.3256, 2.0331, 0.9242, 2.1457, 1.8719, 1.7838, 2.1419], + device='cuda:0'), covar=tensor([0.0959, 0.0892, 0.1605, 0.2115, 0.1454, 0.2338, 0.2260, 0.0989], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0195, 0.0201, 0.0184, 0.0211, 0.0209, 0.0224, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:40:16,476 INFO [finetune.py:976] (0/7) Epoch 19, batch 5250, loss[loss=0.1601, simple_loss=0.2346, pruned_loss=0.04282, over 4749.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2492, pruned_loss=0.05422, over 951954.81 frames. ], batch size: 54, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:40:19,068 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1026, 1.9242, 2.0251, 0.8880, 2.3759, 2.6121, 2.1564, 1.8286], + device='cuda:0'), covar=tensor([0.1127, 0.0842, 0.0601, 0.0829, 0.0678, 0.0749, 0.0538, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0125, 0.0125, 0.0132, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.1156e-05, 1.0897e-04, 8.9480e-05, 8.8795e-05, 9.3043e-05, 9.2359e-05, + 1.0242e-04, 1.0606e-04], device='cuda:0') +2023-03-26 23:40:19,993 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.566e+02 1.984e+02 2.332e+02 4.295e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-03-26 23:40:26,550 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108363.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:42,049 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:46,330 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:49,964 INFO [finetune.py:976] (0/7) Epoch 19, batch 5300, loss[loss=0.1756, simple_loss=0.256, pruned_loss=0.04762, over 4820.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2512, pruned_loss=0.05499, over 951187.70 frames. ], batch size: 47, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:41:22,534 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:41:32,019 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:41:36,861 INFO [finetune.py:976] (0/7) Epoch 19, batch 5350, loss[loss=0.1547, simple_loss=0.2337, pruned_loss=0.03781, over 4757.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2507, pruned_loss=0.05443, over 951683.46 frames. ], batch size: 27, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:41:39,886 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.546e+01 1.453e+02 1.815e+02 2.266e+02 3.194e+02, threshold=3.630e+02, percent-clipped=0.0 +2023-03-26 23:42:02,308 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3222, 1.4412, 1.4785, 0.8845, 1.4727, 1.7263, 1.7903, 1.3624], + device='cuda:0'), covar=tensor([0.0916, 0.0615, 0.0544, 0.0507, 0.0467, 0.0532, 0.0306, 0.0761], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0125, 0.0125, 0.0132, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0977e-05, 1.0885e-04, 8.9532e-05, 8.8752e-05, 9.2630e-05, 9.2217e-05, + 1.0230e-04, 1.0624e-04], device='cuda:0') +2023-03-26 23:42:09,100 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108493.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:42:09,703 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5622, 1.5773, 1.3454, 1.4908, 1.8343, 1.8099, 1.5289, 1.3682], + device='cuda:0'), covar=tensor([0.0351, 0.0342, 0.0611, 0.0347, 0.0223, 0.0473, 0.0363, 0.0458], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0108, 0.0146, 0.0112, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5820e-05, 8.3545e-05, 1.1461e-04, 8.6164e-05, 7.8144e-05, 8.2489e-05, + 7.4609e-05, 8.5666e-05], device='cuda:0') +2023-03-26 23:42:12,599 INFO [finetune.py:976] (0/7) Epoch 19, batch 5400, loss[loss=0.1502, simple_loss=0.2284, pruned_loss=0.03594, over 4793.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2485, pruned_loss=0.05394, over 951637.32 frames. ], batch size: 45, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:42:13,980 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4929, 1.5683, 1.5021, 0.9507, 1.5824, 1.8207, 1.8415, 1.4014], + device='cuda:0'), covar=tensor([0.1106, 0.0579, 0.0603, 0.0551, 0.0504, 0.0604, 0.0344, 0.0797], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0125, 0.0125, 0.0132, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.1089e-05, 1.0904e-04, 8.9712e-05, 8.8835e-05, 9.2651e-05, 9.2364e-05, + 1.0247e-04, 1.0647e-04], device='cuda:0') +2023-03-26 23:42:36,568 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 23:42:44,742 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7045, 1.6703, 1.4587, 1.6279, 2.0237, 2.0130, 1.7229, 1.5706], + device='cuda:0'), covar=tensor([0.0320, 0.0346, 0.0547, 0.0301, 0.0188, 0.0393, 0.0315, 0.0406], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0108, 0.0145, 0.0112, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5475e-05, 8.3177e-05, 1.1398e-04, 8.5721e-05, 7.7785e-05, 8.2043e-05, + 7.4283e-05, 8.5261e-05], device='cuda:0') +2023-03-26 23:42:47,707 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4490, 1.3732, 0.9313, 0.2347, 1.0853, 1.3303, 1.3868, 1.2785], + device='cuda:0'), covar=tensor([0.0984, 0.0785, 0.1278, 0.1810, 0.1337, 0.2097, 0.1979, 0.0781], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0200, 0.0183, 0.0211, 0.0209, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:42:58,626 INFO [finetune.py:976] (0/7) Epoch 19, batch 5450, loss[loss=0.1867, simple_loss=0.25, pruned_loss=0.06174, over 4855.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2472, pruned_loss=0.05417, over 952795.07 frames. ], batch size: 44, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:43:01,646 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.562e+02 1.816e+02 2.216e+02 4.232e+02, threshold=3.632e+02, percent-clipped=2.0 +2023-03-26 23:43:11,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2500, 2.9415, 2.6788, 1.4011, 2.8104, 2.3392, 2.2233, 2.4780], + device='cuda:0'), covar=tensor([0.0997, 0.0747, 0.1805, 0.2211, 0.1638, 0.2190, 0.2211, 0.1325], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0200, 0.0183, 0.0210, 0.0208, 0.0223, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:43:16,121 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-03-26 23:43:17,225 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108577.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:43:31,853 INFO [finetune.py:976] (0/7) Epoch 19, batch 5500, loss[loss=0.1298, simple_loss=0.1996, pruned_loss=0.03002, over 4788.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2433, pruned_loss=0.05279, over 954409.24 frames. ], batch size: 29, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:43:41,992 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-03-26 23:43:58,807 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108638.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:44:05,685 INFO [finetune.py:976] (0/7) Epoch 19, batch 5550, loss[loss=0.2042, simple_loss=0.2724, pruned_loss=0.06801, over 4927.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2453, pruned_loss=0.05355, over 956050.01 frames. ], batch size: 33, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:44:08,703 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.553e+02 1.822e+02 2.201e+02 3.552e+02, threshold=3.643e+02, percent-clipped=0.0 +2023-03-26 23:44:27,073 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108681.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:44:37,477 INFO [finetune.py:976] (0/7) Epoch 19, batch 5600, loss[loss=0.1865, simple_loss=0.273, pruned_loss=0.04996, over 4810.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2491, pruned_loss=0.05404, over 955808.06 frames. ], batch size: 45, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:44:46,268 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2296, 2.0023, 1.4652, 0.5931, 1.6665, 1.8672, 1.6821, 1.8345], + device='cuda:0'), covar=tensor([0.0819, 0.0832, 0.1597, 0.1976, 0.1377, 0.2277, 0.2365, 0.0923], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0201, 0.0183, 0.0211, 0.0209, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:45:09,469 INFO [finetune.py:976] (0/7) Epoch 19, batch 5650, loss[loss=0.1821, simple_loss=0.2577, pruned_loss=0.05328, over 4865.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2515, pruned_loss=0.05413, over 956874.42 frames. ], batch size: 31, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:45:12,316 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.584e+02 1.878e+02 2.184e+02 3.636e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-26 23:45:13,454 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6675, 3.3112, 3.5164, 3.3855, 3.2503, 3.2075, 3.7734, 1.1418], + device='cuda:0'), covar=tensor([0.1449, 0.1698, 0.1553, 0.2151, 0.2466, 0.2529, 0.1551, 0.8143], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0242, 0.0276, 0.0288, 0.0330, 0.0280, 0.0299, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:45:23,128 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-03-26 23:45:24,351 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-26 23:45:32,947 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108788.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:45:34,783 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108791.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:45:36,001 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7688, 1.5532, 1.4715, 1.2012, 1.5360, 1.6121, 1.5456, 1.9837], + device='cuda:0'), covar=tensor([0.3268, 0.3848, 0.2916, 0.3347, 0.3195, 0.2043, 0.3246, 0.1723], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0260, 0.0230, 0.0275, 0.0252, 0.0221, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:45:37,228 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 23:45:39,454 INFO [finetune.py:976] (0/7) Epoch 19, batch 5700, loss[loss=0.1701, simple_loss=0.2259, pruned_loss=0.0572, over 4121.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.246, pruned_loss=0.05342, over 933670.48 frames. ], batch size: 18, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:45:51,037 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7204, 2.5584, 2.2497, 2.7293, 2.5406, 2.4289, 2.9370, 2.7304], + device='cuda:0'), covar=tensor([0.1202, 0.1923, 0.2859, 0.2223, 0.2379, 0.1587, 0.2272, 0.1761], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0188, 0.0235, 0.0254, 0.0247, 0.0204, 0.0216, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:45:56,322 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-19.pt +2023-03-26 23:46:08,102 INFO [finetune.py:976] (0/7) Epoch 20, batch 0, loss[loss=0.1847, simple_loss=0.2573, pruned_loss=0.05601, over 4863.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2573, pruned_loss=0.05601, over 4863.00 frames. ], batch size: 31, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:46:08,103 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-26 23:46:15,849 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9634, 1.3258, 0.9760, 1.7093, 2.2214, 1.2766, 1.6736, 1.6393], + device='cuda:0'), covar=tensor([0.1321, 0.1827, 0.1664, 0.0983, 0.1680, 0.1876, 0.1158, 0.1750], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0091, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:46:17,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2218, 1.9914, 1.8121, 1.7759, 1.9129, 1.9438, 1.9853, 2.6145], + device='cuda:0'), covar=tensor([0.3863, 0.4745, 0.3482, 0.4069, 0.4077, 0.2657, 0.3916, 0.1856], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0230, 0.0275, 0.0251, 0.0221, 0.0252, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:46:24,542 INFO [finetune.py:1010] (0/7) Epoch 20, validation: loss=0.158, simple_loss=0.2276, pruned_loss=0.04423, over 2265189.00 frames. +2023-03-26 23:46:24,542 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-26 23:46:52,590 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:46:58,208 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.806e+01 1.424e+02 1.737e+02 2.098e+02 5.389e+02, threshold=3.475e+02, percent-clipped=2.0 +2023-03-26 23:47:01,504 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 23:47:17,654 INFO [finetune.py:976] (0/7) Epoch 20, batch 50, loss[loss=0.1816, simple_loss=0.2492, pruned_loss=0.05697, over 4847.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2546, pruned_loss=0.0561, over 216386.08 frames. ], batch size: 49, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:47:31,380 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2710, 2.8689, 2.7140, 1.2141, 2.9959, 2.1988, 0.7836, 1.8894], + device='cuda:0'), covar=tensor([0.2459, 0.2089, 0.1787, 0.3454, 0.1334, 0.1125, 0.3866, 0.1515], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0173, 0.0158, 0.0128, 0.0157, 0.0121, 0.0145, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-26 23:47:57,360 INFO [finetune.py:976] (0/7) Epoch 20, batch 100, loss[loss=0.1843, simple_loss=0.2529, pruned_loss=0.05786, over 4870.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2482, pruned_loss=0.05513, over 380727.48 frames. ], batch size: 34, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:48:06,339 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:48:07,713 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-03-26 23:48:17,425 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 23:48:23,147 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.362e+02 1.754e+02 2.070e+02 5.157e+02, threshold=3.508e+02, percent-clipped=1.0 +2023-03-26 23:48:23,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7015, 1.5818, 1.9463, 1.3095, 1.6497, 1.9265, 1.5610, 2.0449], + device='cuda:0'), covar=tensor([0.1119, 0.2124, 0.1246, 0.1730, 0.1007, 0.1213, 0.2962, 0.0745], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0203, 0.0190, 0.0188, 0.0173, 0.0211, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:48:38,582 INFO [finetune.py:976] (0/7) Epoch 20, batch 150, loss[loss=0.2549, simple_loss=0.3013, pruned_loss=0.1043, over 4830.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2437, pruned_loss=0.05401, over 505950.58 frames. ], batch size: 39, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:48:41,561 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:48:42,195 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5885, 1.6518, 1.3924, 1.6124, 1.9289, 1.8124, 1.5745, 1.4792], + device='cuda:0'), covar=tensor([0.0303, 0.0317, 0.0607, 0.0310, 0.0214, 0.0530, 0.0351, 0.0403], + device='cuda:0'), in_proj_covar=tensor([0.0096, 0.0106, 0.0142, 0.0110, 0.0098, 0.0109, 0.0098, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4134e-05, 8.1646e-05, 1.1191e-04, 8.4489e-05, 7.6606e-05, 8.0729e-05, + 7.3179e-05, 8.3893e-05], device='cuda:0') +2023-03-26 23:49:02,962 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 23:49:08,725 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 23:49:11,418 INFO [finetune.py:976] (0/7) Epoch 20, batch 200, loss[loss=0.1972, simple_loss=0.2694, pruned_loss=0.06249, over 4829.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.242, pruned_loss=0.05309, over 606563.65 frames. ], batch size: 47, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:49:13,181 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109029.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:49:19,019 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109037.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:49:26,817 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1043, 1.8124, 2.3214, 3.8930, 2.6977, 2.7748, 0.7456, 3.3230], + device='cuda:0'), covar=tensor([0.1570, 0.1493, 0.1528, 0.0600, 0.0766, 0.1663, 0.2217, 0.0391], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0133, 0.0163, 0.0100, 0.0135, 0.0124, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-26 23:49:29,131 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.519e+02 1.780e+02 2.129e+02 3.450e+02, threshold=3.561e+02, percent-clipped=0.0 +2023-03-26 23:49:31,572 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9150, 1.7408, 2.1309, 1.4186, 1.9771, 2.1711, 1.7129, 2.3292], + device='cuda:0'), covar=tensor([0.1195, 0.1937, 0.1297, 0.1794, 0.0837, 0.1363, 0.2612, 0.0770], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0192, 0.0190, 0.0175, 0.0213, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:49:44,471 INFO [finetune.py:976] (0/7) Epoch 20, batch 250, loss[loss=0.1835, simple_loss=0.2558, pruned_loss=0.0556, over 4824.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2439, pruned_loss=0.05274, over 682954.85 frames. ], batch size: 33, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:49:52,128 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:49:58,714 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109098.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:50:11,417 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0661, 1.9440, 1.6891, 1.9501, 1.8240, 1.8935, 1.8479, 2.5530], + device='cuda:0'), covar=tensor([0.3880, 0.4394, 0.3225, 0.4095, 0.4362, 0.2474, 0.4051, 0.1826], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0261, 0.0230, 0.0276, 0.0252, 0.0221, 0.0253, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:50:17,245 INFO [finetune.py:976] (0/7) Epoch 20, batch 300, loss[loss=0.1521, simple_loss=0.2208, pruned_loss=0.04172, over 4767.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2456, pruned_loss=0.05241, over 743634.67 frames. ], batch size: 28, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:50:23,612 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109136.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:50:31,197 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109147.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:50:35,398 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.518e+02 1.856e+02 2.256e+02 3.204e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-26 23:50:36,167 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7641, 1.7489, 1.4820, 1.8452, 2.2066, 1.9012, 1.6969, 1.4382], + device='cuda:0'), covar=tensor([0.2215, 0.1920, 0.1904, 0.1620, 0.1941, 0.1223, 0.2225, 0.1923], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0209, 0.0210, 0.0193, 0.0243, 0.0187, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:50:50,168 INFO [finetune.py:976] (0/7) Epoch 20, batch 350, loss[loss=0.1653, simple_loss=0.244, pruned_loss=0.04324, over 4798.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2476, pruned_loss=0.05291, over 791145.51 frames. ], batch size: 45, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:51:01,445 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:51:02,046 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5352, 2.4771, 2.4623, 1.9352, 2.4342, 2.8017, 2.6026, 2.1549], + device='cuda:0'), covar=tensor([0.0558, 0.0560, 0.0694, 0.0839, 0.0841, 0.0640, 0.0637, 0.1085], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0121, 0.0125, 0.0139, 0.0139, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:51:10,837 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1488, 1.9984, 1.6986, 2.0347, 1.8949, 1.9104, 1.9067, 2.6396], + device='cuda:0'), covar=tensor([0.3814, 0.4358, 0.3370, 0.3931, 0.4295, 0.2427, 0.3877, 0.1834], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0261, 0.0230, 0.0275, 0.0252, 0.0222, 0.0253, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:51:25,272 INFO [finetune.py:976] (0/7) Epoch 20, batch 400, loss[loss=0.1611, simple_loss=0.2435, pruned_loss=0.03933, over 4911.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2485, pruned_loss=0.05289, over 828093.08 frames. ], batch size: 33, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:51:34,315 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109233.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:03,601 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.657e+02 1.900e+02 2.185e+02 4.941e+02, threshold=3.801e+02, percent-clipped=3.0 +2023-03-26 23:52:04,363 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:26,284 INFO [finetune.py:976] (0/7) Epoch 20, batch 450, loss[loss=0.1717, simple_loss=0.2455, pruned_loss=0.04889, over 4877.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.248, pruned_loss=0.05253, over 854353.88 frames. ], batch size: 32, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:52:29,294 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:38,175 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:54,639 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109318.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:53:00,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.49 vs. limit=2.0 +2023-03-26 23:53:02,105 INFO [finetune.py:976] (0/7) Epoch 20, batch 500, loss[loss=0.2044, simple_loss=0.2653, pruned_loss=0.07181, over 4916.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2475, pruned_loss=0.05254, over 878318.45 frames. ], batch size: 36, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:53:34,823 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.492e+02 1.802e+02 2.178e+02 4.247e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-26 23:53:39,363 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:53:47,669 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4463, 1.5654, 1.5014, 0.8669, 1.6641, 1.7590, 1.7890, 1.3853], + device='cuda:0'), covar=tensor([0.1017, 0.0671, 0.0539, 0.0551, 0.0486, 0.0676, 0.0364, 0.0753], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0124, 0.0124, 0.0130, 0.0128, 0.0142, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.0243e-05, 1.0815e-04, 8.8500e-05, 8.7808e-05, 9.1689e-05, 9.1520e-05, + 1.0162e-04, 1.0523e-04], device='cuda:0') +2023-03-26 23:53:52,984 INFO [finetune.py:976] (0/7) Epoch 20, batch 550, loss[loss=0.2038, simple_loss=0.2611, pruned_loss=0.07319, over 4823.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2452, pruned_loss=0.05212, over 894620.71 frames. ], batch size: 38, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:53:53,720 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7984, 1.7869, 1.5545, 1.8080, 2.1730, 2.1294, 1.7421, 1.6107], + device='cuda:0'), covar=tensor([0.0301, 0.0330, 0.0625, 0.0311, 0.0209, 0.0404, 0.0342, 0.0422], + device='cuda:0'), in_proj_covar=tensor([0.0096, 0.0106, 0.0143, 0.0110, 0.0099, 0.0110, 0.0099, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.4665e-05, 8.1787e-05, 1.1256e-04, 8.4643e-05, 7.6756e-05, 8.1074e-05, + 7.3525e-05, 8.4114e-05], device='cuda:0') +2023-03-26 23:53:54,324 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109379.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:54:03,683 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109393.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:54:26,238 INFO [finetune.py:976] (0/7) Epoch 20, batch 600, loss[loss=0.2745, simple_loss=0.3279, pruned_loss=0.1105, over 4730.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2463, pruned_loss=0.05346, over 908389.72 frames. ], batch size: 59, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:54:39,802 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:54:42,864 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2733, 1.7265, 1.7278, 0.9935, 1.8827, 2.0251, 1.9327, 1.6925], + device='cuda:0'), covar=tensor([0.0835, 0.0762, 0.0591, 0.0646, 0.0694, 0.0755, 0.0503, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0125, 0.0125, 0.0131, 0.0129, 0.0143, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.1048e-05, 1.0913e-04, 8.9055e-05, 8.8457e-05, 9.2430e-05, 9.2308e-05, + 1.0237e-04, 1.0609e-04], device='cuda:0') +2023-03-26 23:54:44,548 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.417e+01 1.548e+02 1.729e+02 2.159e+02 3.434e+02, threshold=3.458e+02, percent-clipped=0.0 +2023-03-26 23:54:52,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7262, 1.7562, 1.5813, 1.7588, 1.5602, 4.3955, 1.6550, 2.1550], + device='cuda:0'), covar=tensor([0.3228, 0.2364, 0.2104, 0.2245, 0.1489, 0.0125, 0.2317, 0.1134], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0123, 0.0114, 0.0096, 0.0095, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-26 23:54:59,325 INFO [finetune.py:976] (0/7) Epoch 20, batch 650, loss[loss=0.1764, simple_loss=0.2461, pruned_loss=0.05332, over 4855.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2502, pruned_loss=0.05462, over 919532.97 frames. ], batch size: 25, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:55:01,892 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6332, 1.5491, 2.0060, 1.2718, 1.9400, 1.9195, 1.4870, 2.0533], + device='cuda:0'), covar=tensor([0.1447, 0.2218, 0.1700, 0.2046, 0.0860, 0.1444, 0.2753, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0202, 0.0189, 0.0186, 0.0171, 0.0210, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:55:10,858 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:55:16,089 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9669, 1.5244, 2.0553, 2.0175, 1.7774, 1.7457, 1.9567, 1.8676], + device='cuda:0'), covar=tensor([0.3742, 0.3778, 0.2902, 0.3363, 0.4414, 0.3451, 0.3973, 0.2759], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0240, 0.0261, 0.0280, 0.0278, 0.0252, 0.0288, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:55:33,023 INFO [finetune.py:976] (0/7) Epoch 20, batch 700, loss[loss=0.1948, simple_loss=0.2612, pruned_loss=0.06419, over 4911.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2519, pruned_loss=0.05531, over 926882.53 frames. ], batch size: 33, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:55:33,751 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3983, 2.2951, 2.4068, 1.7790, 2.3685, 2.4902, 2.4098, 2.0681], + device='cuda:0'), covar=tensor([0.0608, 0.0632, 0.0646, 0.0874, 0.0610, 0.0693, 0.0596, 0.0953], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0120, 0.0125, 0.0140, 0.0139, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-26 23:55:47,493 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:55:51,355 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.485e+02 1.783e+02 2.085e+02 4.380e+02, threshold=3.566e+02, percent-clipped=3.0 +2023-03-26 23:55:59,010 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0772, 1.9908, 2.0173, 0.7622, 2.2970, 2.5625, 2.0198, 1.8414], + device='cuda:0'), covar=tensor([0.0907, 0.0622, 0.0515, 0.0731, 0.0648, 0.0513, 0.0580, 0.0789], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0124, 0.0124, 0.0130, 0.0128, 0.0141, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.0285e-05, 1.0834e-04, 8.8403e-05, 8.7835e-05, 9.1608e-05, 9.1442e-05, + 1.0150e-04, 1.0520e-04], device='cuda:0') +2023-03-26 23:56:06,061 INFO [finetune.py:976] (0/7) Epoch 20, batch 750, loss[loss=0.1985, simple_loss=0.2637, pruned_loss=0.06663, over 4929.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2519, pruned_loss=0.05517, over 931881.55 frames. ], batch size: 42, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:56:07,739 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109579.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:39,568 INFO [finetune.py:976] (0/7) Epoch 20, batch 800, loss[loss=0.1695, simple_loss=0.2411, pruned_loss=0.04892, over 4185.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2517, pruned_loss=0.05472, over 935603.45 frames. ], batch size: 66, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:56:42,629 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 23:56:50,327 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:56,827 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:59,764 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.497e+02 1.774e+02 2.103e+02 3.199e+02, threshold=3.548e+02, percent-clipped=0.0 +2023-03-26 23:57:03,317 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:57:23,376 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109674.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:57:25,125 INFO [finetune.py:976] (0/7) Epoch 20, batch 850, loss[loss=0.1573, simple_loss=0.224, pruned_loss=0.04535, over 4834.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2494, pruned_loss=0.05419, over 938502.87 frames. ], batch size: 47, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:57:40,056 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109693.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:58:06,039 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:58:10,728 INFO [finetune.py:976] (0/7) Epoch 20, batch 900, loss[loss=0.1309, simple_loss=0.2018, pruned_loss=0.03003, over 4908.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2471, pruned_loss=0.05362, over 942962.84 frames. ], batch size: 46, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:58:22,308 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109741.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:58:40,325 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.530e+02 1.823e+02 2.181e+02 3.809e+02, threshold=3.647e+02, percent-clipped=1.0 +2023-03-26 23:59:03,840 INFO [finetune.py:976] (0/7) Epoch 20, batch 950, loss[loss=0.2096, simple_loss=0.2606, pruned_loss=0.07928, over 4174.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2459, pruned_loss=0.05375, over 944661.82 frames. ], batch size: 65, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:59:36,849 INFO [finetune.py:976] (0/7) Epoch 20, batch 1000, loss[loss=0.2472, simple_loss=0.321, pruned_loss=0.08675, over 4848.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2465, pruned_loss=0.05402, over 947338.27 frames. ], batch size: 44, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:59:38,325 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 23:59:52,402 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:59:55,343 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.630e+02 1.951e+02 2.313e+02 5.473e+02, threshold=3.903e+02, percent-clipped=2.0 +2023-03-27 00:00:00,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7761, 3.9415, 3.7194, 1.8069, 4.0191, 3.0695, 0.8241, 2.7466], + device='cuda:0'), covar=tensor([0.2131, 0.2370, 0.1538, 0.3574, 0.1036, 0.0938, 0.4621, 0.1556], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0131, 0.0161, 0.0123, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 00:00:08,418 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6181, 1.4639, 1.4311, 1.5056, 0.9775, 3.3419, 1.2947, 1.5705], + device='cuda:0'), covar=tensor([0.3155, 0.2374, 0.2137, 0.2322, 0.1951, 0.0210, 0.2906, 0.1341], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:00:10,589 INFO [finetune.py:976] (0/7) Epoch 20, batch 1050, loss[loss=0.1542, simple_loss=0.2344, pruned_loss=0.03707, over 4864.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2495, pruned_loss=0.05434, over 948189.52 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:00:10,798 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 00:00:24,789 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109898.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:00:43,711 INFO [finetune.py:976] (0/7) Epoch 20, batch 1100, loss[loss=0.1922, simple_loss=0.2618, pruned_loss=0.06123, over 4821.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2505, pruned_loss=0.054, over 950101.72 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:00:49,668 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:00:59,775 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109950.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:01,176 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 00:01:02,701 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.532e+02 1.890e+02 2.271e+02 3.423e+02, threshold=3.780e+02, percent-clipped=0.0 +2023-03-27 00:01:13,262 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-27 00:01:14,934 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5993, 0.7433, 1.6747, 1.5834, 1.4899, 1.4271, 1.5459, 1.5979], + device='cuda:0'), covar=tensor([0.3482, 0.3514, 0.2948, 0.3046, 0.4061, 0.3134, 0.3656, 0.2734], + device='cuda:0'), in_proj_covar=tensor([0.0253, 0.0241, 0.0262, 0.0281, 0.0279, 0.0254, 0.0290, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:01:15,512 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:17,220 INFO [finetune.py:976] (0/7) Epoch 20, batch 1150, loss[loss=0.1662, simple_loss=0.2358, pruned_loss=0.04829, over 4859.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2505, pruned_loss=0.05359, over 951289.38 frames. ], batch size: 31, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:01:31,926 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109998.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:33,712 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-110000.pt +2023-03-27 00:01:44,406 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110015.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:45,670 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110017.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:48,677 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110022.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:48,851 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-27 00:01:52,168 INFO [finetune.py:976] (0/7) Epoch 20, batch 1200, loss[loss=0.1754, simple_loss=0.2514, pruned_loss=0.04965, over 4886.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2506, pruned_loss=0.05418, over 951983.46 frames. ], batch size: 43, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:02:04,533 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110044.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:02:08,825 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 00:02:11,656 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.682e+01 1.527e+02 1.776e+02 2.166e+02 4.163e+02, threshold=3.552e+02, percent-clipped=2.0 +2023-03-27 00:02:32,493 INFO [finetune.py:976] (0/7) Epoch 20, batch 1250, loss[loss=0.1744, simple_loss=0.2441, pruned_loss=0.05233, over 4802.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2482, pruned_loss=0.05343, over 951774.09 frames. ], batch size: 29, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:02:33,235 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110078.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:02,202 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:09,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-27 00:03:09,987 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110110.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:13,058 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110115.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:23,864 INFO [finetune.py:976] (0/7) Epoch 20, batch 1300, loss[loss=0.1572, simple_loss=0.2318, pruned_loss=0.04127, over 4825.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2454, pruned_loss=0.05257, over 952081.26 frames. ], batch size: 40, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:03:24,671 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 00:03:33,290 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3451, 2.2215, 2.3987, 1.6877, 2.2300, 2.4710, 2.5038, 1.8250], + device='cuda:0'), covar=tensor([0.0526, 0.0564, 0.0598, 0.0809, 0.0687, 0.0560, 0.0473, 0.1065], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0135, 0.0141, 0.0121, 0.0125, 0.0140, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:03:45,426 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.532e+02 1.884e+02 2.318e+02 4.682e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-03-27 00:04:01,788 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7051, 1.5518, 1.6051, 1.6082, 1.1089, 3.3541, 1.3183, 1.7001], + device='cuda:0'), covar=tensor([0.3310, 0.2584, 0.2104, 0.2443, 0.1882, 0.0212, 0.2589, 0.1349], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0115, 0.0119, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:04:04,755 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:04:12,713 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110176.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:04:13,180 INFO [finetune.py:976] (0/7) Epoch 20, batch 1350, loss[loss=0.1633, simple_loss=0.2208, pruned_loss=0.05283, over 4211.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2454, pruned_loss=0.05313, over 950801.83 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:05:04,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6989, 2.4265, 1.7972, 0.9233, 2.1433, 2.1244, 1.9475, 2.2223], + device='cuda:0'), covar=tensor([0.0738, 0.0826, 0.1738, 0.2124, 0.1281, 0.1854, 0.2048, 0.0975], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0193, 0.0200, 0.0182, 0.0210, 0.0207, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:05:15,850 INFO [finetune.py:976] (0/7) Epoch 20, batch 1400, loss[loss=0.1496, simple_loss=0.2328, pruned_loss=0.03323, over 4813.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2483, pruned_loss=0.05378, over 951386.41 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:05:26,739 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110235.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:05:48,218 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.545e+02 1.824e+02 2.176e+02 3.637e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-27 00:05:56,187 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-03-27 00:06:02,045 INFO [finetune.py:976] (0/7) Epoch 20, batch 1450, loss[loss=0.1577, simple_loss=0.2346, pruned_loss=0.04039, over 4852.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2497, pruned_loss=0.05438, over 952147.49 frames. ], batch size: 49, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:06:06,231 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:06,292 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:10,365 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6768, 4.0435, 4.3122, 4.4679, 4.4315, 4.2008, 4.7940, 1.4561], + device='cuda:0'), covar=tensor([0.0755, 0.0823, 0.0758, 0.0958, 0.1197, 0.1345, 0.0540, 0.5701], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0280, 0.0292, 0.0333, 0.0283, 0.0302, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:06:20,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110303.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:28,602 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:30,444 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110318.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:35,827 INFO [finetune.py:976] (0/7) Epoch 20, batch 1500, loss[loss=0.1845, simple_loss=0.2597, pruned_loss=0.05464, over 4865.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2514, pruned_loss=0.05501, over 953208.42 frames. ], batch size: 34, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:06:47,649 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110344.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:54,080 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-27 00:06:55,109 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.590e+02 1.939e+02 2.244e+02 3.777e+02, threshold=3.878e+02, percent-clipped=2.0 +2023-03-27 00:07:00,470 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110363.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:01,159 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110364.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:07,069 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110373.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:09,448 INFO [finetune.py:976] (0/7) Epoch 20, batch 1550, loss[loss=0.1425, simple_loss=0.2309, pruned_loss=0.02708, over 4792.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2517, pruned_loss=0.05449, over 955457.58 frames. ], batch size: 26, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:07:10,794 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110379.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:21,934 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2000, 1.7415, 2.4693, 1.5078, 2.1840, 2.4368, 1.7442, 2.5555], + device='cuda:0'), covar=tensor([0.1308, 0.2277, 0.1471, 0.2241, 0.1074, 0.1441, 0.3157, 0.0980], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0206, 0.0192, 0.0191, 0.0176, 0.0215, 0.0220, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:07:25,519 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110400.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:38,433 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6228, 1.5326, 1.1472, 0.3023, 1.2728, 1.4332, 1.3532, 1.4923], + device='cuda:0'), covar=tensor([0.0907, 0.0654, 0.1232, 0.1730, 0.1258, 0.2088, 0.2164, 0.0782], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0191, 0.0199, 0.0181, 0.0208, 0.0206, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:07:42,115 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-27 00:07:43,127 INFO [finetune.py:976] (0/7) Epoch 20, batch 1600, loss[loss=0.1687, simple_loss=0.2395, pruned_loss=0.04892, over 4893.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2483, pruned_loss=0.05361, over 954738.86 frames. ], batch size: 37, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:08:13,512 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.758e+01 1.502e+02 1.787e+02 2.306e+02 4.709e+02, threshold=3.574e+02, percent-clipped=2.0 +2023-03-27 00:08:29,699 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110466.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:08:33,150 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:08:40,391 INFO [finetune.py:976] (0/7) Epoch 20, batch 1650, loss[loss=0.1423, simple_loss=0.2229, pruned_loss=0.03087, over 4782.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2463, pruned_loss=0.05287, over 953260.76 frames. ], batch size: 29, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:09:11,597 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2247, 1.8987, 2.5083, 4.2861, 2.9365, 3.0187, 0.9888, 3.6038], + device='cuda:0'), covar=tensor([0.1728, 0.1479, 0.1494, 0.0550, 0.0750, 0.1338, 0.2074, 0.0395], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0164, 0.0101, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:09:24,058 INFO [finetune.py:976] (0/7) Epoch 20, batch 1700, loss[loss=0.2486, simple_loss=0.3039, pruned_loss=0.09671, over 4742.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2454, pruned_loss=0.05353, over 954494.93 frames. ], batch size: 54, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:09:42,518 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.500e+02 1.773e+02 2.246e+02 3.830e+02, threshold=3.546e+02, percent-clipped=2.0 +2023-03-27 00:09:57,622 INFO [finetune.py:976] (0/7) Epoch 20, batch 1750, loss[loss=0.203, simple_loss=0.2741, pruned_loss=0.06599, over 4906.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2482, pruned_loss=0.05457, over 954531.09 frames. ], batch size: 36, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:10:18,013 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8017, 4.1345, 4.4197, 4.6263, 4.5526, 4.2539, 4.8878, 1.5246], + device='cuda:0'), covar=tensor([0.0739, 0.0869, 0.0793, 0.1027, 0.1105, 0.1595, 0.0504, 0.5896], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0277, 0.0291, 0.0331, 0.0282, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:10:20,407 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110611.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:10:27,762 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 00:10:30,497 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2832, 1.8965, 2.4603, 1.6451, 2.1972, 2.5078, 1.8038, 2.4679], + device='cuda:0'), covar=tensor([0.0981, 0.1871, 0.1165, 0.1693, 0.0828, 0.1238, 0.2674, 0.0813], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0189, 0.0174, 0.0213, 0.0218, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:10:30,986 INFO [finetune.py:976] (0/7) Epoch 20, batch 1800, loss[loss=0.1593, simple_loss=0.2125, pruned_loss=0.05301, over 4389.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2496, pruned_loss=0.05407, over 955717.50 frames. ], batch size: 19, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:10:38,892 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110639.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:10:47,404 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-27 00:10:53,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3806, 3.8014, 4.0184, 4.2418, 4.1576, 3.8185, 4.4362, 1.3101], + device='cuda:0'), covar=tensor([0.0867, 0.0880, 0.0915, 0.1037, 0.1268, 0.1732, 0.0787, 0.5876], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0246, 0.0278, 0.0293, 0.0332, 0.0284, 0.0303, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:10:55,666 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.630e+02 1.887e+02 2.220e+02 3.285e+02, threshold=3.774e+02, percent-clipped=0.0 +2023-03-27 00:10:55,835 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2584, 1.8317, 2.2186, 2.1713, 1.9607, 1.9343, 2.1820, 2.0045], + device='cuda:0'), covar=tensor([0.3810, 0.3996, 0.3222, 0.4056, 0.4906, 0.3898, 0.4712, 0.3188], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0240, 0.0262, 0.0281, 0.0278, 0.0253, 0.0289, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:11:01,585 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:03,344 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6854, 1.6931, 1.6372, 1.0597, 1.7826, 2.0909, 1.8842, 1.4888], + device='cuda:0'), covar=tensor([0.1217, 0.0707, 0.0509, 0.0633, 0.0422, 0.0541, 0.0427, 0.0751], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0125, 0.0125, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0930e-05, 1.0884e-04, 8.9388e-05, 8.8388e-05, 9.2352e-05, 9.2119e-05, + 1.0179e-04, 1.0589e-04], device='cuda:0') +2023-03-27 00:11:11,033 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110672.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:11,622 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110673.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:12,219 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110674.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:14,002 INFO [finetune.py:976] (0/7) Epoch 20, batch 1850, loss[loss=0.1446, simple_loss=0.2272, pruned_loss=0.03101, over 4743.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2492, pruned_loss=0.05408, over 953972.36 frames. ], batch size: 54, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:11:29,601 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110700.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:44,147 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110721.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:47,774 INFO [finetune.py:976] (0/7) Epoch 20, batch 1900, loss[loss=0.2101, simple_loss=0.2783, pruned_loss=0.07094, over 4813.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2509, pruned_loss=0.05479, over 952503.43 frames. ], batch size: 33, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:11:54,348 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110736.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:58,158 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-27 00:12:01,609 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110748.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:06,288 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.472e+02 1.826e+02 2.198e+02 3.929e+02, threshold=3.651e+02, percent-clipped=1.0 +2023-03-27 00:12:14,051 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:17,552 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110771.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:21,632 INFO [finetune.py:976] (0/7) Epoch 20, batch 1950, loss[loss=0.2134, simple_loss=0.2703, pruned_loss=0.0783, over 4809.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2493, pruned_loss=0.05397, over 953947.86 frames. ], batch size: 41, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:12:23,759 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 00:12:34,789 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110797.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:12:42,713 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2830, 2.2375, 1.7329, 2.2271, 2.1692, 1.8985, 2.5994, 2.3171], + device='cuda:0'), covar=tensor([0.1277, 0.2105, 0.2940, 0.2531, 0.2516, 0.1610, 0.3179, 0.1616], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0189, 0.0235, 0.0253, 0.0248, 0.0205, 0.0215, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:12:46,180 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110814.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:49,723 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110819.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:55,565 INFO [finetune.py:976] (0/7) Epoch 20, batch 2000, loss[loss=0.2007, simple_loss=0.2601, pruned_loss=0.07069, over 4814.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2463, pruned_loss=0.05285, over 952707.85 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:13:17,583 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.482e+02 1.740e+02 2.016e+02 2.901e+02, threshold=3.480e+02, percent-clipped=0.0 +2023-03-27 00:13:30,407 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110868.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:13:38,684 INFO [finetune.py:976] (0/7) Epoch 20, batch 2050, loss[loss=0.1946, simple_loss=0.2688, pruned_loss=0.06014, over 4848.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2446, pruned_loss=0.05292, over 950643.28 frames. ], batch size: 49, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:14:28,248 INFO [finetune.py:976] (0/7) Epoch 20, batch 2100, loss[loss=0.2302, simple_loss=0.2856, pruned_loss=0.08747, over 4863.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2442, pruned_loss=0.05334, over 950252.18 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:14:29,616 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:39,978 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110939.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:50,629 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.458e+01 1.520e+02 1.862e+02 2.217e+02 3.516e+02, threshold=3.725e+02, percent-clipped=1.0 +2023-03-27 00:14:52,593 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110958.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:53,162 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110959.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:58,415 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110967.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:03,707 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:05,917 INFO [finetune.py:976] (0/7) Epoch 20, batch 2150, loss[loss=0.1721, simple_loss=0.2438, pruned_loss=0.05016, over 4691.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2482, pruned_loss=0.05473, over 950240.85 frames. ], batch size: 23, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:15:12,588 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:25,801 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111007.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:33,622 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111019.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:35,877 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111022.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:37,895 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-27 00:15:38,858 INFO [finetune.py:976] (0/7) Epoch 20, batch 2200, loss[loss=0.181, simple_loss=0.2527, pruned_loss=0.05459, over 4819.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2515, pruned_loss=0.05558, over 949566.91 frames. ], batch size: 33, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:15:50,043 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6263, 3.2570, 2.8911, 1.6204, 3.0840, 2.5961, 2.4079, 2.8151], + device='cuda:0'), covar=tensor([0.0847, 0.0734, 0.1753, 0.2116, 0.1428, 0.1724, 0.1907, 0.1123], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0192, 0.0201, 0.0183, 0.0211, 0.0207, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:16:00,233 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.523e+02 1.854e+02 2.195e+02 4.707e+02, threshold=3.708e+02, percent-clipped=2.0 +2023-03-27 00:16:10,013 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6222, 1.5332, 1.3147, 1.6474, 1.9071, 1.7175, 1.2078, 1.3241], + device='cuda:0'), covar=tensor([0.2117, 0.1992, 0.1995, 0.1634, 0.1586, 0.1172, 0.2501, 0.1926], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0194, 0.0243, 0.0188, 0.0216, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:16:23,022 INFO [finetune.py:976] (0/7) Epoch 20, batch 2250, loss[loss=0.1929, simple_loss=0.2771, pruned_loss=0.05434, over 4820.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2521, pruned_loss=0.0553, over 951034.31 frames. ], batch size: 38, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:16:26,618 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5116, 1.0603, 0.8020, 1.3527, 1.8630, 0.8608, 1.2910, 1.4370], + device='cuda:0'), covar=tensor([0.1326, 0.1990, 0.1688, 0.1081, 0.1917, 0.1914, 0.1386, 0.1700], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:16:33,699 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111092.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:16:56,459 INFO [finetune.py:976] (0/7) Epoch 20, batch 2300, loss[loss=0.1406, simple_loss=0.21, pruned_loss=0.03558, over 4002.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2508, pruned_loss=0.05415, over 952083.61 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:17:15,912 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.510e+02 1.791e+02 2.077e+02 4.254e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-27 00:17:30,234 INFO [finetune.py:976] (0/7) Epoch 20, batch 2350, loss[loss=0.1426, simple_loss=0.2103, pruned_loss=0.03749, over 4713.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2483, pruned_loss=0.05296, over 952093.91 frames. ], batch size: 23, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:01,642 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:02,305 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:03,426 INFO [finetune.py:976] (0/7) Epoch 20, batch 2400, loss[loss=0.1387, simple_loss=0.2055, pruned_loss=0.03601, over 4723.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2461, pruned_loss=0.05268, over 949022.99 frames. ], batch size: 23, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:22,738 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.826e+01 1.548e+02 1.775e+02 2.063e+02 3.363e+02, threshold=3.550e+02, percent-clipped=0.0 +2023-03-27 00:18:30,978 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111267.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:36,919 INFO [finetune.py:976] (0/7) Epoch 20, batch 2450, loss[loss=0.2133, simple_loss=0.2636, pruned_loss=0.08153, over 4759.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.242, pruned_loss=0.05133, over 950927.34 frames. ], batch size: 54, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:50,052 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:21,774 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:22,345 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:33,327 INFO [finetune.py:976] (0/7) Epoch 20, batch 2500, loss[loss=0.2245, simple_loss=0.2921, pruned_loss=0.07844, over 4854.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2454, pruned_loss=0.05306, over 951381.61 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:19:38,994 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-27 00:19:39,522 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6663, 0.7106, 1.7205, 1.6360, 1.5484, 1.4573, 1.5246, 1.6228], + device='cuda:0'), covar=tensor([0.3185, 0.3445, 0.3006, 0.3051, 0.4143, 0.3338, 0.3708, 0.2834], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0243, 0.0264, 0.0283, 0.0281, 0.0255, 0.0291, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:19:47,677 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-27 00:20:03,164 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.591e+02 1.841e+02 2.119e+02 4.112e+02, threshold=3.683e+02, percent-clipped=1.0 +2023-03-27 00:20:17,417 INFO [finetune.py:976] (0/7) Epoch 20, batch 2550, loss[loss=0.1846, simple_loss=0.2616, pruned_loss=0.05382, over 4855.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2481, pruned_loss=0.05385, over 951264.18 frames. ], batch size: 31, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:20:27,519 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111392.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 00:20:51,266 INFO [finetune.py:976] (0/7) Epoch 20, batch 2600, loss[loss=0.1509, simple_loss=0.2238, pruned_loss=0.03901, over 4799.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2491, pruned_loss=0.05388, over 953484.49 frames. ], batch size: 40, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:20:59,724 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111440.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:21:10,161 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.233e+02 1.591e+02 1.878e+02 2.220e+02 5.233e+02, threshold=3.757e+02, percent-clipped=1.0 +2023-03-27 00:21:13,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8831, 1.5927, 1.5004, 1.2126, 1.6445, 1.6348, 1.6242, 2.2240], + device='cuda:0'), covar=tensor([0.3970, 0.4515, 0.3267, 0.3636, 0.3758, 0.2371, 0.3627, 0.1917], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0260, 0.0228, 0.0274, 0.0249, 0.0220, 0.0250, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:21:20,889 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:21:21,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4448, 1.5345, 1.8339, 1.6807, 1.7547, 3.0561, 1.4046, 1.6791], + device='cuda:0'), covar=tensor([0.1127, 0.1892, 0.1099, 0.1006, 0.1558, 0.0345, 0.1567, 0.1738], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0077, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:21:31,397 INFO [finetune.py:976] (0/7) Epoch 20, batch 2650, loss[loss=0.1994, simple_loss=0.2755, pruned_loss=0.06164, over 4804.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2513, pruned_loss=0.05449, over 953133.33 frames. ], batch size: 40, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:21:32,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4167, 3.8257, 4.0589, 4.2439, 4.1718, 3.9015, 4.5178, 1.5855], + device='cuda:0'), covar=tensor([0.0834, 0.0836, 0.0809, 0.1036, 0.1183, 0.1559, 0.0677, 0.5490], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0278, 0.0292, 0.0334, 0.0283, 0.0303, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:21:59,074 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 00:22:06,274 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111524.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:07,975 INFO [finetune.py:976] (0/7) Epoch 20, batch 2700, loss[loss=0.169, simple_loss=0.2466, pruned_loss=0.04567, over 4924.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2507, pruned_loss=0.0541, over 954637.13 frames. ], batch size: 33, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:22:09,763 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:27,296 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.581e+02 1.832e+02 2.307e+02 3.346e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 00:22:38,119 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:41,153 INFO [finetune.py:976] (0/7) Epoch 20, batch 2750, loss[loss=0.1554, simple_loss=0.2248, pruned_loss=0.04298, over 4805.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2482, pruned_loss=0.05376, over 954120.74 frames. ], batch size: 25, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:22:44,074 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111581.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:23:06,504 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111614.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:23:14,366 INFO [finetune.py:976] (0/7) Epoch 20, batch 2800, loss[loss=0.1617, simple_loss=0.2385, pruned_loss=0.04249, over 4760.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2456, pruned_loss=0.05297, over 956519.45 frames. ], batch size: 27, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:23:25,756 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8779, 1.4472, 2.0377, 1.8641, 1.7079, 1.6847, 1.8588, 1.8884], + device='cuda:0'), covar=tensor([0.3871, 0.3938, 0.2849, 0.3606, 0.4346, 0.3660, 0.4165, 0.2768], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0241, 0.0262, 0.0280, 0.0277, 0.0252, 0.0288, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:23:32,759 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.029e+01 1.565e+02 1.748e+02 2.177e+02 3.583e+02, threshold=3.496e+02, percent-clipped=0.0 +2023-03-27 00:23:38,054 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111662.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:23:48,042 INFO [finetune.py:976] (0/7) Epoch 20, batch 2850, loss[loss=0.1945, simple_loss=0.2586, pruned_loss=0.06519, over 4911.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2441, pruned_loss=0.05257, over 956679.20 frames. ], batch size: 37, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:23:53,596 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6841, 1.6092, 2.0569, 1.8819, 1.7759, 3.6029, 1.4819, 1.7238], + device='cuda:0'), covar=tensor([0.1039, 0.1960, 0.1116, 0.1088, 0.1621, 0.0305, 0.1699, 0.1921], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:24:41,627 INFO [finetune.py:976] (0/7) Epoch 20, batch 2900, loss[loss=0.1737, simple_loss=0.2359, pruned_loss=0.05578, over 4162.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2469, pruned_loss=0.05332, over 956447.97 frames. ], batch size: 18, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:25:12,983 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.542e+02 1.820e+02 2.254e+02 5.949e+02, threshold=3.641e+02, percent-clipped=1.0 +2023-03-27 00:25:19,958 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.01 vs. limit=5.0 +2023-03-27 00:25:31,431 INFO [finetune.py:976] (0/7) Epoch 20, batch 2950, loss[loss=0.1609, simple_loss=0.2299, pruned_loss=0.0459, over 4901.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2501, pruned_loss=0.05434, over 956532.06 frames. ], batch size: 32, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:25:40,063 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111791.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:25:49,657 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 00:26:03,392 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:26:05,125 INFO [finetune.py:976] (0/7) Epoch 20, batch 3000, loss[loss=0.1793, simple_loss=0.2488, pruned_loss=0.0549, over 4925.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2514, pruned_loss=0.05495, over 955788.52 frames. ], batch size: 33, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:26:05,127 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 00:26:06,788 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4475, 1.3371, 1.3075, 1.4159, 1.6989, 1.6417, 1.4517, 1.2812], + device='cuda:0'), covar=tensor([0.0381, 0.0316, 0.0665, 0.0313, 0.0263, 0.0409, 0.0331, 0.0428], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0107, 0.0145, 0.0111, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.4838e-05, 8.2350e-05, 1.1394e-04, 8.5056e-05, 7.7760e-05, 8.1940e-05, + 7.4232e-05, 8.5598e-05], device='cuda:0') +2023-03-27 00:26:10,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9184, 3.4626, 3.6271, 3.8381, 3.6264, 3.5037, 3.9835, 1.3644], + device='cuda:0'), covar=tensor([0.0831, 0.0864, 0.0907, 0.0887, 0.1401, 0.1555, 0.0787, 0.5326], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0278, 0.0292, 0.0333, 0.0283, 0.0304, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:26:11,196 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4532, 1.3136, 1.2925, 1.4014, 1.6942, 1.6199, 1.4028, 1.2172], + device='cuda:0'), covar=tensor([0.0365, 0.0325, 0.0640, 0.0318, 0.0244, 0.0423, 0.0342, 0.0465], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0107, 0.0145, 0.0111, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.4838e-05, 8.2350e-05, 1.1394e-04, 8.5056e-05, 7.7760e-05, 8.1940e-05, + 7.4232e-05, 8.5598e-05], device='cuda:0') +2023-03-27 00:26:20,306 INFO [finetune.py:1010] (0/7) Epoch 20, validation: loss=0.1563, simple_loss=0.2257, pruned_loss=0.04344, over 2265189.00 frames. +2023-03-27 00:26:20,306 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 00:26:29,881 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4446, 1.5044, 1.7088, 1.7298, 1.5094, 3.2325, 1.4931, 1.5172], + device='cuda:0'), covar=tensor([0.0888, 0.1699, 0.1122, 0.0888, 0.1573, 0.0246, 0.1343, 0.1769], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:26:40,763 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111852.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 00:26:46,423 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7740, 2.6217, 2.1461, 1.1189, 2.3102, 2.0877, 2.0223, 2.3513], + device='cuda:0'), covar=tensor([0.0730, 0.0729, 0.1592, 0.1984, 0.1572, 0.2059, 0.1820, 0.0906], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0193, 0.0200, 0.0183, 0.0212, 0.0207, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:26:48,441 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.618e+02 1.875e+02 2.230e+02 3.575e+02, threshold=3.749e+02, percent-clipped=0.0 +2023-03-27 00:26:57,657 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4505, 1.5079, 1.6979, 1.7811, 1.4710, 3.2735, 1.5624, 1.4914], + device='cuda:0'), covar=tensor([0.1005, 0.1859, 0.1145, 0.0912, 0.1708, 0.0254, 0.1325, 0.1812], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:27:11,158 INFO [finetune.py:976] (0/7) Epoch 20, batch 3050, loss[loss=0.1866, simple_loss=0.2493, pruned_loss=0.06201, over 4867.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2531, pruned_loss=0.05519, over 958072.19 frames. ], batch size: 31, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:27:18,720 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:27:46,882 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9393, 1.4565, 2.0051, 1.8929, 1.7036, 1.6738, 1.8822, 1.8518], + device='cuda:0'), covar=tensor([0.3594, 0.3791, 0.3032, 0.3631, 0.4410, 0.3748, 0.4353, 0.2868], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0241, 0.0262, 0.0280, 0.0279, 0.0253, 0.0289, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:28:06,554 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3749, 3.7443, 4.0067, 4.2069, 4.1527, 3.8888, 4.4906, 1.4749], + device='cuda:0'), covar=tensor([0.0638, 0.0803, 0.0797, 0.0909, 0.0977, 0.1312, 0.0582, 0.5471], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0278, 0.0292, 0.0334, 0.0283, 0.0304, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:28:08,479 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 00:28:14,418 INFO [finetune.py:976] (0/7) Epoch 20, batch 3100, loss[loss=0.1348, simple_loss=0.2096, pruned_loss=0.03007, over 4768.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2516, pruned_loss=0.05484, over 957190.35 frames. ], batch size: 28, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:28:15,693 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:28:17,590 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7519, 1.6483, 1.6159, 0.9047, 1.7998, 1.9449, 1.9064, 1.4356], + device='cuda:0'), covar=tensor([0.0848, 0.0588, 0.0472, 0.0517, 0.0389, 0.0553, 0.0311, 0.0611], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0124, 0.0124, 0.0130, 0.0128, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([9.0354e-05, 1.0831e-04, 8.8573e-05, 8.7698e-05, 9.1283e-05, 9.1835e-05, + 1.0071e-04, 1.0540e-04], device='cuda:0') +2023-03-27 00:28:32,998 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.507e+02 1.737e+02 2.301e+02 5.151e+02, threshold=3.474e+02, percent-clipped=3.0 +2023-03-27 00:28:47,145 INFO [finetune.py:976] (0/7) Epoch 20, batch 3150, loss[loss=0.172, simple_loss=0.2419, pruned_loss=0.05105, over 4768.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2481, pruned_loss=0.05371, over 958122.25 frames. ], batch size: 28, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:29:02,214 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-112000.pt +2023-03-27 00:29:23,071 INFO [finetune.py:976] (0/7) Epoch 20, batch 3200, loss[loss=0.2306, simple_loss=0.2742, pruned_loss=0.09352, over 4260.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2444, pruned_loss=0.05275, over 957696.16 frames. ], batch size: 65, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:29:33,200 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9035, 1.7509, 1.8907, 1.4069, 1.8617, 1.9194, 1.9473, 1.5054], + device='cuda:0'), covar=tensor([0.0501, 0.0676, 0.0570, 0.0737, 0.0730, 0.0601, 0.0518, 0.1088], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0135, 0.0140, 0.0120, 0.0124, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:29:42,594 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1833, 1.6080, 1.8606, 0.6964, 2.1027, 2.4199, 1.8460, 1.6682], + device='cuda:0'), covar=tensor([0.1272, 0.1264, 0.0686, 0.0938, 0.0689, 0.0786, 0.0638, 0.0988], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0124, 0.0124, 0.0129, 0.0128, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.9932e-05, 1.0797e-04, 8.8239e-05, 8.7379e-05, 9.0762e-05, 9.1645e-05, + 1.0047e-04, 1.0530e-04], device='cuda:0') +2023-03-27 00:29:53,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6688, 1.5466, 2.1298, 3.3745, 2.2640, 2.3729, 1.0428, 2.7011], + device='cuda:0'), covar=tensor([0.1742, 0.1467, 0.1340, 0.0508, 0.0794, 0.1623, 0.1889, 0.0447], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0164, 0.0100, 0.0135, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:29:56,972 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.705e+01 1.565e+02 1.739e+02 2.228e+02 3.922e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 00:30:06,300 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112066.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:30:13,834 INFO [finetune.py:976] (0/7) Epoch 20, batch 3250, loss[loss=0.1564, simple_loss=0.2318, pruned_loss=0.04048, over 4909.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2441, pruned_loss=0.05284, over 958624.53 frames. ], batch size: 35, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:30:15,758 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4029, 2.2034, 1.9925, 2.2037, 2.1176, 2.1561, 2.1681, 2.8279], + device='cuda:0'), covar=tensor([0.3697, 0.4543, 0.3380, 0.3653, 0.3984, 0.2374, 0.3524, 0.1742], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0261, 0.0230, 0.0275, 0.0251, 0.0221, 0.0250, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:30:48,091 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3161, 1.2349, 1.2141, 0.7182, 1.2254, 1.4115, 1.4742, 1.1639], + device='cuda:0'), covar=tensor([0.0801, 0.0490, 0.0491, 0.0477, 0.0485, 0.0508, 0.0275, 0.0528], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0125, 0.0125, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0881e-05, 1.0893e-04, 8.9349e-05, 8.8287e-05, 9.2029e-05, 9.2592e-05, + 1.0155e-04, 1.0634e-04], device='cuda:0') +2023-03-27 00:30:54,493 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112124.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:30:56,194 INFO [finetune.py:976] (0/7) Epoch 20, batch 3300, loss[loss=0.1748, simple_loss=0.2477, pruned_loss=0.05091, over 4765.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2471, pruned_loss=0.05355, over 957989.26 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:30:56,311 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:31:09,777 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112147.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:31:16,114 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.319e+01 1.653e+02 1.982e+02 2.472e+02 3.934e+02, threshold=3.965e+02, percent-clipped=2.0 +2023-03-27 00:31:26,466 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:31:29,977 INFO [finetune.py:976] (0/7) Epoch 20, batch 3350, loss[loss=0.1857, simple_loss=0.2611, pruned_loss=0.05516, over 4902.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2503, pruned_loss=0.05459, over 959211.90 frames. ], batch size: 36, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:32:11,816 INFO [finetune.py:976] (0/7) Epoch 20, batch 3400, loss[loss=0.1925, simple_loss=0.25, pruned_loss=0.06753, over 4123.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2525, pruned_loss=0.05535, over 957096.48 frames. ], batch size: 65, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:32:12,000 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 00:32:31,200 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.676e+02 1.964e+02 2.392e+02 4.564e+02, threshold=3.928e+02, percent-clipped=2.0 +2023-03-27 00:32:44,378 INFO [finetune.py:976] (0/7) Epoch 20, batch 3450, loss[loss=0.1926, simple_loss=0.2603, pruned_loss=0.06251, over 4828.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2533, pruned_loss=0.05564, over 956201.57 frames. ], batch size: 49, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:32:51,093 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0831, 1.8870, 1.6458, 1.5799, 1.8422, 1.8068, 1.8835, 2.5005], + device='cuda:0'), covar=tensor([0.3413, 0.3970, 0.3212, 0.3833, 0.3647, 0.2346, 0.3489, 0.1748], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0262, 0.0231, 0.0277, 0.0251, 0.0222, 0.0251, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:32:52,841 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9481, 1.8666, 1.6785, 2.0792, 2.5318, 2.1148, 1.7767, 1.5491], + device='cuda:0'), covar=tensor([0.2144, 0.1903, 0.1803, 0.1509, 0.1547, 0.1059, 0.2178, 0.1799], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0194, 0.0243, 0.0189, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:33:19,432 INFO [finetune.py:976] (0/7) Epoch 20, batch 3500, loss[loss=0.1677, simple_loss=0.2306, pruned_loss=0.05242, over 4824.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2506, pruned_loss=0.05509, over 956339.69 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:33:56,443 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.856e+01 1.454e+02 1.799e+02 2.150e+02 5.052e+02, threshold=3.598e+02, percent-clipped=2.0 +2023-03-27 00:34:18,956 INFO [finetune.py:976] (0/7) Epoch 20, batch 3550, loss[loss=0.1885, simple_loss=0.2464, pruned_loss=0.06526, over 4868.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2475, pruned_loss=0.05386, over 957244.60 frames. ], batch size: 34, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:34:28,793 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5188, 1.4573, 1.8026, 1.8100, 1.6479, 3.3547, 1.4820, 1.6242], + device='cuda:0'), covar=tensor([0.0977, 0.1900, 0.1058, 0.0959, 0.1604, 0.0246, 0.1507, 0.1859], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:34:36,595 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112389.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:17,935 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112422.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:20,888 INFO [finetune.py:976] (0/7) Epoch 20, batch 3600, loss[loss=0.1383, simple_loss=0.2028, pruned_loss=0.0369, over 4717.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2435, pruned_loss=0.0521, over 956763.55 frames. ], batch size: 23, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:35:32,157 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.68 vs. limit=2.0 +2023-03-27 00:35:37,948 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112447.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:35:39,827 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:44,342 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.527e+02 1.833e+02 2.137e+02 4.874e+02, threshold=3.667e+02, percent-clipped=1.0 +2023-03-27 00:36:08,335 INFO [finetune.py:976] (0/7) Epoch 20, batch 3650, loss[loss=0.1985, simple_loss=0.2755, pruned_loss=0.06071, over 4810.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2466, pruned_loss=0.05359, over 954586.13 frames. ], batch size: 40, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:36:20,408 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112495.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:36:41,250 INFO [finetune.py:976] (0/7) Epoch 20, batch 3700, loss[loss=0.2231, simple_loss=0.2829, pruned_loss=0.08163, over 4217.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2504, pruned_loss=0.05453, over 952692.52 frames. ], batch size: 66, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:37:01,298 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.568e+02 1.941e+02 2.323e+02 3.962e+02, threshold=3.882e+02, percent-clipped=1.0 +2023-03-27 00:37:03,794 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3047, 1.3372, 1.4604, 0.7554, 1.4574, 1.7272, 1.7028, 1.3297], + device='cuda:0'), covar=tensor([0.1082, 0.0926, 0.0609, 0.0624, 0.0611, 0.0599, 0.0483, 0.0876], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0125, 0.0124, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0439e-05, 1.0861e-04, 8.9138e-05, 8.7751e-05, 9.1464e-05, 9.2601e-05, + 1.0136e-04, 1.0592e-04], device='cuda:0') +2023-03-27 00:37:05,562 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112562.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:10,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5255, 1.8016, 1.5063, 1.5964, 2.2298, 2.0208, 1.8631, 1.8044], + device='cuda:0'), covar=tensor([0.0492, 0.0329, 0.0536, 0.0334, 0.0257, 0.0648, 0.0330, 0.0390], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0099, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5094e-05, 8.2293e-05, 1.1336e-04, 8.5197e-05, 7.7940e-05, 8.2365e-05, + 7.4049e-05, 8.5713e-05], device='cuda:0') +2023-03-27 00:37:15,914 INFO [finetune.py:976] (0/7) Epoch 20, batch 3750, loss[loss=0.2183, simple_loss=0.2859, pruned_loss=0.07536, over 4921.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2515, pruned_loss=0.05502, over 952899.60 frames. ], batch size: 38, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:37:16,030 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:50,115 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112617.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:50,247 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-03-27 00:37:54,229 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112623.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:56,877 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4722, 3.3849, 3.3288, 1.5733, 3.5479, 2.6669, 1.0494, 2.5066], + device='cuda:0'), covar=tensor([0.2716, 0.1677, 0.1556, 0.3329, 0.1111, 0.1031, 0.4055, 0.1400], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0177, 0.0160, 0.0129, 0.0162, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 00:37:57,418 INFO [finetune.py:976] (0/7) Epoch 20, batch 3800, loss[loss=0.1819, simple_loss=0.2521, pruned_loss=0.05589, over 4743.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2526, pruned_loss=0.05496, over 953397.58 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:38:03,712 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 00:38:04,231 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112638.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:38:16,043 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.515e+02 1.800e+02 2.233e+02 3.828e+02, threshold=3.600e+02, percent-clipped=0.0 +2023-03-27 00:38:30,670 INFO [finetune.py:976] (0/7) Epoch 20, batch 3850, loss[loss=0.1931, simple_loss=0.2549, pruned_loss=0.06558, over 4802.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2508, pruned_loss=0.05401, over 953581.68 frames. ], batch size: 45, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:38:31,388 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112678.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:38:59,763 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112722.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:38:59,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4244, 1.6291, 1.7031, 0.8567, 1.6407, 1.8746, 1.8675, 1.5022], + device='cuda:0'), covar=tensor([0.0977, 0.0611, 0.0474, 0.0603, 0.0418, 0.0546, 0.0342, 0.0702], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0125, 0.0124, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0292e-05, 1.0846e-04, 8.9019e-05, 8.7496e-05, 9.1263e-05, 9.2405e-05, + 1.0119e-04, 1.0580e-04], device='cuda:0') +2023-03-27 00:39:03,155 INFO [finetune.py:976] (0/7) Epoch 20, batch 3900, loss[loss=0.1757, simple_loss=0.2354, pruned_loss=0.05806, over 4820.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2486, pruned_loss=0.05391, over 954415.39 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:39:15,037 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112745.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:39:21,554 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.403e+01 1.612e+02 1.959e+02 2.410e+02 4.123e+02, threshold=3.918e+02, percent-clipped=1.0 +2023-03-27 00:39:32,078 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112770.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:39:36,268 INFO [finetune.py:976] (0/7) Epoch 20, batch 3950, loss[loss=0.1883, simple_loss=0.2317, pruned_loss=0.07243, over 4302.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2449, pruned_loss=0.0528, over 954096.87 frames. ], batch size: 18, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:39:52,595 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5984, 2.2907, 2.8590, 1.7673, 2.6358, 2.8730, 2.0909, 3.1353], + device='cuda:0'), covar=tensor([0.1558, 0.2003, 0.1675, 0.2428, 0.1138, 0.1447, 0.2924, 0.0836], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0191, 0.0190, 0.0175, 0.0213, 0.0219, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:39:59,318 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4505, 1.0854, 0.7272, 1.3115, 1.9297, 0.7162, 1.1786, 1.3274], + device='cuda:0'), covar=tensor([0.1523, 0.2088, 0.1867, 0.1266, 0.1952, 0.2036, 0.1595, 0.1940], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:40:19,040 INFO [finetune.py:976] (0/7) Epoch 20, batch 4000, loss[loss=0.1528, simple_loss=0.2301, pruned_loss=0.03777, over 4907.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2443, pruned_loss=0.05244, over 950162.63 frames. ], batch size: 43, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:40:48,437 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.638e+02 1.904e+02 2.320e+02 3.891e+02, threshold=3.808e+02, percent-clipped=0.0 +2023-03-27 00:40:50,431 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1441, 2.0686, 1.7087, 2.0507, 1.9671, 1.9389, 2.0372, 2.6706], + device='cuda:0'), covar=tensor([0.3856, 0.4907, 0.3510, 0.4094, 0.4295, 0.2702, 0.3970, 0.1806], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0263, 0.0231, 0.0276, 0.0253, 0.0222, 0.0251, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:41:04,956 INFO [finetune.py:976] (0/7) Epoch 20, batch 4050, loss[loss=0.2075, simple_loss=0.2751, pruned_loss=0.0699, over 4831.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2476, pruned_loss=0.05388, over 950690.32 frames. ], batch size: 49, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:41:41,254 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:41:43,105 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112921.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:41:47,212 INFO [finetune.py:976] (0/7) Epoch 20, batch 4100, loss[loss=0.1545, simple_loss=0.2348, pruned_loss=0.03717, over 4860.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2505, pruned_loss=0.05442, over 951886.47 frames. ], batch size: 34, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:41:51,372 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112933.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:42:06,078 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.570e+02 1.869e+02 2.355e+02 4.214e+02, threshold=3.739e+02, percent-clipped=0.0 +2023-03-27 00:42:07,464 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7430, 1.7718, 1.5076, 1.9421, 2.3212, 1.8767, 1.5927, 1.4313], + device='cuda:0'), covar=tensor([0.2247, 0.1904, 0.1895, 0.1503, 0.1540, 0.1226, 0.2214, 0.1880], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0209, 0.0211, 0.0193, 0.0241, 0.0187, 0.0216, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:42:17,469 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112973.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:42:19,818 INFO [finetune.py:976] (0/7) Epoch 20, batch 4150, loss[loss=0.1864, simple_loss=0.2518, pruned_loss=0.06046, over 4825.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2522, pruned_loss=0.05521, over 952605.00 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:42:23,996 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:43:03,966 INFO [finetune.py:976] (0/7) Epoch 20, batch 4200, loss[loss=0.2003, simple_loss=0.2762, pruned_loss=0.06221, over 4803.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2523, pruned_loss=0.05499, over 953534.55 frames. ], batch size: 40, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:43:16,404 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113045.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:43:23,432 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.578e+02 1.900e+02 2.258e+02 5.235e+02, threshold=3.800e+02, percent-clipped=4.0 +2023-03-27 00:43:36,991 INFO [finetune.py:976] (0/7) Epoch 20, batch 4250, loss[loss=0.1794, simple_loss=0.2427, pruned_loss=0.05803, over 4908.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.25, pruned_loss=0.05439, over 953879.19 frames. ], batch size: 46, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:43:47,712 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113093.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:44:10,286 INFO [finetune.py:976] (0/7) Epoch 20, batch 4300, loss[loss=0.1524, simple_loss=0.2285, pruned_loss=0.03814, over 4818.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2481, pruned_loss=0.05397, over 955340.14 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:44:30,859 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.679e+01 1.396e+02 1.741e+02 2.023e+02 4.034e+02, threshold=3.482e+02, percent-clipped=1.0 +2023-03-27 00:44:43,636 INFO [finetune.py:976] (0/7) Epoch 20, batch 4350, loss[loss=0.1224, simple_loss=0.1898, pruned_loss=0.02751, over 4728.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2442, pruned_loss=0.05256, over 954691.78 frames. ], batch size: 23, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:45:12,210 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:17,616 INFO [finetune.py:976] (0/7) Epoch 20, batch 4400, loss[loss=0.2021, simple_loss=0.2801, pruned_loss=0.06199, over 4845.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2456, pruned_loss=0.05342, over 954125.70 frames. ], batch size: 47, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:45:21,834 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:22,444 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113234.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:32,837 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 00:45:49,232 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.729e+01 1.603e+02 1.846e+02 2.173e+02 5.642e+02, threshold=3.692e+02, percent-clipped=4.0 +2023-03-27 00:46:00,743 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113266.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:08,333 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113273.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:46:10,701 INFO [finetune.py:976] (0/7) Epoch 20, batch 4450, loss[loss=0.1758, simple_loss=0.2564, pruned_loss=0.04763, over 4759.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2483, pruned_loss=0.05406, over 954122.42 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:46:10,769 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:13,203 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113281.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:23,174 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113295.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:32,652 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5633, 1.5616, 1.8664, 1.2146, 1.6446, 1.8154, 1.4426, 2.0406], + device='cuda:0'), covar=tensor([0.1358, 0.2155, 0.1220, 0.1844, 0.0984, 0.1255, 0.3064, 0.0884], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0205, 0.0190, 0.0190, 0.0175, 0.0213, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:46:35,142 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3873, 2.9665, 2.7361, 1.4730, 2.9649, 2.1864, 2.2592, 2.6950], + device='cuda:0'), covar=tensor([0.0753, 0.0857, 0.1650, 0.2099, 0.1330, 0.2198, 0.1974, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0192, 0.0200, 0.0182, 0.0210, 0.0209, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:46:50,260 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:53,840 INFO [finetune.py:976] (0/7) Epoch 20, batch 4500, loss[loss=0.1638, simple_loss=0.241, pruned_loss=0.04328, over 4893.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2499, pruned_loss=0.05426, over 952675.32 frames. ], batch size: 36, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:46:56,403 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5176, 3.2488, 3.0401, 1.3982, 3.3781, 2.6632, 1.0696, 2.3143], + device='cuda:0'), covar=tensor([0.2942, 0.2008, 0.1943, 0.3652, 0.1288, 0.1001, 0.4014, 0.1564], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0178, 0.0161, 0.0130, 0.0162, 0.0124, 0.0148, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 00:47:13,426 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.617e+02 2.027e+02 2.372e+02 5.258e+02, threshold=4.055e+02, percent-clipped=2.0 +2023-03-27 00:47:27,582 INFO [finetune.py:976] (0/7) Epoch 20, batch 4550, loss[loss=0.2, simple_loss=0.2679, pruned_loss=0.06608, over 4916.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.252, pruned_loss=0.05528, over 953049.81 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:47:40,393 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-27 00:47:47,434 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0789, 1.9475, 1.8982, 0.9376, 2.2471, 2.3840, 2.1100, 1.7421], + device='cuda:0'), covar=tensor([0.1006, 0.0663, 0.0542, 0.0664, 0.0427, 0.0714, 0.0415, 0.0858], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0124, 0.0123, 0.0129, 0.0128, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.9582e-05, 1.0771e-04, 8.8539e-05, 8.6679e-05, 9.0437e-05, 9.1342e-05, + 1.0022e-04, 1.0523e-04], device='cuda:0') +2023-03-27 00:48:03,324 INFO [finetune.py:976] (0/7) Epoch 20, batch 4600, loss[loss=0.1385, simple_loss=0.2215, pruned_loss=0.02774, over 4745.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2515, pruned_loss=0.05496, over 953465.57 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:48:04,058 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8719, 1.6373, 2.1585, 1.3770, 1.9697, 2.1675, 1.5527, 2.3108], + device='cuda:0'), covar=tensor([0.1698, 0.2205, 0.1490, 0.2241, 0.1058, 0.1581, 0.3195, 0.0830], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0207, 0.0192, 0.0191, 0.0176, 0.0214, 0.0221, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:48:31,340 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.757e+01 1.614e+02 1.832e+02 2.145e+02 3.668e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 00:48:45,584 INFO [finetune.py:976] (0/7) Epoch 20, batch 4650, loss[loss=0.1352, simple_loss=0.2086, pruned_loss=0.0309, over 4769.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2484, pruned_loss=0.05405, over 954176.95 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:48:54,232 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 00:49:10,345 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3303, 1.9044, 2.3634, 2.2608, 2.0425, 2.0649, 2.2697, 2.1736], + device='cuda:0'), covar=tensor([0.4274, 0.4507, 0.3380, 0.3904, 0.5055, 0.4063, 0.5005, 0.3338], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0242, 0.0263, 0.0281, 0.0279, 0.0254, 0.0289, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:49:19,010 INFO [finetune.py:976] (0/7) Epoch 20, batch 4700, loss[loss=0.1404, simple_loss=0.2201, pruned_loss=0.03039, over 4919.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2445, pruned_loss=0.05231, over 955059.72 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:49:33,199 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0481, 1.7248, 2.3381, 1.5964, 2.1160, 2.3339, 1.6901, 2.4546], + device='cuda:0'), covar=tensor([0.1364, 0.2114, 0.1398, 0.2039, 0.0875, 0.1359, 0.2743, 0.0794], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0207, 0.0192, 0.0191, 0.0176, 0.0214, 0.0221, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:49:34,413 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 00:49:37,205 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.713e+01 1.518e+02 1.819e+02 2.172e+02 5.123e+02, threshold=3.639e+02, percent-clipped=2.0 +2023-03-27 00:49:38,078 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-27 00:49:50,421 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2191, 2.0657, 1.7617, 2.0268, 1.9294, 1.9502, 1.9961, 2.7042], + device='cuda:0'), covar=tensor([0.3725, 0.4134, 0.3372, 0.3727, 0.3963, 0.2456, 0.3984, 0.1726], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0264, 0.0232, 0.0278, 0.0254, 0.0223, 0.0253, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:49:51,492 INFO [finetune.py:976] (0/7) Epoch 20, batch 4750, loss[loss=0.1477, simple_loss=0.2266, pruned_loss=0.03438, over 4923.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2433, pruned_loss=0.05211, over 955203.18 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:49:51,579 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:00,066 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113590.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:03,712 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9227, 1.6797, 2.1409, 1.3758, 1.9130, 2.1071, 1.6172, 2.3425], + device='cuda:0'), covar=tensor([0.1379, 0.2230, 0.1340, 0.2038, 0.0987, 0.1311, 0.2960, 0.0772], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0206, 0.0191, 0.0191, 0.0176, 0.0213, 0.0220, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:50:10,337 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-27 00:50:18,359 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113617.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:23,692 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113625.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:24,845 INFO [finetune.py:976] (0/7) Epoch 20, batch 4800, loss[loss=0.2098, simple_loss=0.2876, pruned_loss=0.066, over 4856.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2471, pruned_loss=0.05394, over 953775.18 frames. ], batch size: 44, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:50:37,462 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:38,077 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8624, 1.5076, 2.0032, 1.3813, 1.9424, 2.0871, 1.4665, 2.2106], + device='cuda:0'), covar=tensor([0.1109, 0.2140, 0.1320, 0.1833, 0.0811, 0.1140, 0.2869, 0.0721], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0207, 0.0192, 0.0192, 0.0176, 0.0214, 0.0221, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:50:43,821 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.583e+02 1.918e+02 2.188e+02 4.674e+02, threshold=3.836e+02, percent-clipped=2.0 +2023-03-27 00:50:49,916 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1816, 2.1196, 1.6823, 2.2400, 2.1347, 1.8697, 2.4316, 2.2419], + device='cuda:0'), covar=tensor([0.1243, 0.1918, 0.2928, 0.2387, 0.2444, 0.1712, 0.2695, 0.1608], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0186, 0.0233, 0.0251, 0.0245, 0.0203, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:51:01,329 INFO [finetune.py:976] (0/7) Epoch 20, batch 4850, loss[loss=0.1687, simple_loss=0.2424, pruned_loss=0.04749, over 4759.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2503, pruned_loss=0.05511, over 952917.69 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:51:02,044 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113678.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:51:34,440 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113707.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:51:34,558 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 00:51:41,833 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7602, 1.2288, 0.8667, 1.7248, 2.0145, 1.5873, 1.4449, 1.7857], + device='cuda:0'), covar=tensor([0.1441, 0.1926, 0.1909, 0.1076, 0.1921, 0.1889, 0.1437, 0.1760], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0118, 0.0093, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:51:57,945 INFO [finetune.py:976] (0/7) Epoch 20, batch 4900, loss[loss=0.1954, simple_loss=0.2743, pruned_loss=0.05828, over 4830.00 frames. ], tot_loss[loss=0.181, simple_loss=0.252, pruned_loss=0.05504, over 952058.63 frames. ], batch size: 39, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:52:20,127 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.649e+02 2.053e+02 2.498e+02 4.513e+02, threshold=4.105e+02, percent-clipped=3.0 +2023-03-27 00:52:26,751 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1446, 2.1531, 1.3751, 2.3245, 2.2130, 1.7430, 3.0867, 2.2079], + device='cuda:0'), covar=tensor([0.1434, 0.2199, 0.3496, 0.3163, 0.2712, 0.1846, 0.2207, 0.1827], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0187, 0.0233, 0.0251, 0.0245, 0.0203, 0.0213, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:52:34,813 INFO [finetune.py:976] (0/7) Epoch 20, batch 4950, loss[loss=0.1886, simple_loss=0.2647, pruned_loss=0.0563, over 4839.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2526, pruned_loss=0.05527, over 952000.08 frames. ], batch size: 47, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:53:07,553 INFO [finetune.py:976] (0/7) Epoch 20, batch 5000, loss[loss=0.1965, simple_loss=0.2555, pruned_loss=0.06869, over 4927.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2506, pruned_loss=0.05413, over 953419.20 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:53:26,538 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.516e+02 1.779e+02 2.023e+02 5.156e+02, threshold=3.559e+02, percent-clipped=2.0 +2023-03-27 00:53:39,278 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4052, 1.3748, 1.5774, 1.0779, 1.3604, 1.5152, 1.3232, 1.6746], + device='cuda:0'), covar=tensor([0.1163, 0.2084, 0.1218, 0.1483, 0.0995, 0.1224, 0.3104, 0.0876], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0206, 0.0190, 0.0189, 0.0175, 0.0213, 0.0219, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:53:42,068 INFO [finetune.py:976] (0/7) Epoch 20, batch 5050, loss[loss=0.1559, simple_loss=0.2225, pruned_loss=0.0447, over 4846.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2475, pruned_loss=0.05338, over 954999.89 frames. ], batch size: 49, lr: 3.22e-03, grad_scale: 64.0 +2023-03-27 00:53:48,119 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7160, 1.2634, 0.6153, 1.6242, 2.1821, 1.5151, 1.4335, 1.7394], + device='cuda:0'), covar=tensor([0.1978, 0.2782, 0.2636, 0.1530, 0.2106, 0.2384, 0.2093, 0.2717], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0094, 0.0110, 0.0091, 0.0118, 0.0092, 0.0096, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:53:51,029 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113890.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:05,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7694, 2.6898, 2.3472, 2.8682, 2.6520, 2.6189, 2.5993, 3.4814], + device='cuda:0'), covar=tensor([0.3531, 0.4082, 0.3256, 0.3669, 0.3691, 0.2405, 0.3886, 0.1612], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0261, 0.0231, 0.0276, 0.0251, 0.0222, 0.0251, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:54:14,770 INFO [finetune.py:976] (0/7) Epoch 20, batch 5100, loss[loss=0.1759, simple_loss=0.2539, pruned_loss=0.04898, over 4901.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2437, pruned_loss=0.05198, over 954832.34 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:54:23,006 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:26,650 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 00:54:31,740 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-03-27 00:54:35,763 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.567e+02 1.826e+02 2.193e+02 3.507e+02, threshold=3.652e+02, percent-clipped=0.0 +2023-03-27 00:54:46,071 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113973.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:48,407 INFO [finetune.py:976] (0/7) Epoch 20, batch 5150, loss[loss=0.1407, simple_loss=0.2203, pruned_loss=0.03053, over 4760.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.244, pruned_loss=0.0525, over 955820.09 frames. ], batch size: 27, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:55:05,353 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-114000.pt +2023-03-27 00:55:07,779 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:55:07,861 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8857, 1.5207, 1.9551, 1.8022, 1.6751, 1.6285, 1.7805, 1.8175], + device='cuda:0'), covar=tensor([0.3069, 0.3160, 0.2617, 0.3215, 0.3763, 0.3249, 0.3447, 0.2448], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0243, 0.0263, 0.0282, 0.0279, 0.0256, 0.0290, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:55:23,297 INFO [finetune.py:976] (0/7) Epoch 20, batch 5200, loss[loss=0.2212, simple_loss=0.2814, pruned_loss=0.08046, over 4782.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2485, pruned_loss=0.05444, over 953614.35 frames. ], batch size: 29, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:55:43,820 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.581e+02 1.814e+02 2.243e+02 3.815e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-03-27 00:55:56,417 INFO [finetune.py:976] (0/7) Epoch 20, batch 5250, loss[loss=0.2271, simple_loss=0.2888, pruned_loss=0.08271, over 4845.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2499, pruned_loss=0.05469, over 952683.19 frames. ], batch size: 44, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:56:36,233 INFO [finetune.py:976] (0/7) Epoch 20, batch 5300, loss[loss=0.1352, simple_loss=0.2083, pruned_loss=0.03103, over 4829.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2503, pruned_loss=0.05433, over 953319.32 frames. ], batch size: 25, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:56:36,337 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:57:13,340 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.589e+01 1.501e+02 1.834e+02 2.338e+02 4.244e+02, threshold=3.669e+02, percent-clipped=1.0 +2023-03-27 00:57:30,109 INFO [finetune.py:976] (0/7) Epoch 20, batch 5350, loss[loss=0.1543, simple_loss=0.2319, pruned_loss=0.03837, over 4708.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2488, pruned_loss=0.05297, over 950727.10 frames. ], batch size: 23, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:57:37,347 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114188.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:58:01,148 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-27 00:58:03,280 INFO [finetune.py:976] (0/7) Epoch 20, batch 5400, loss[loss=0.1758, simple_loss=0.2417, pruned_loss=0.05497, over 4767.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2469, pruned_loss=0.05242, over 952232.90 frames. ], batch size: 27, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:58:17,591 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6955, 1.5794, 2.1693, 2.8904, 1.9622, 2.2246, 1.3779, 2.4645], + device='cuda:0'), covar=tensor([0.1597, 0.1282, 0.1014, 0.0559, 0.0834, 0.1324, 0.1504, 0.0510], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0136, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 00:58:23,337 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.275e+01 1.521e+02 1.786e+02 2.103e+02 5.074e+02, threshold=3.573e+02, percent-clipped=1.0 +2023-03-27 00:58:33,067 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5190, 1.6169, 1.9771, 1.7918, 1.6283, 3.5308, 1.5118, 1.7024], + device='cuda:0'), covar=tensor([0.0938, 0.1624, 0.1088, 0.0894, 0.1511, 0.0195, 0.1305, 0.1619], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 00:58:33,632 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:58:33,659 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0389, 1.9577, 2.0767, 1.5256, 1.9789, 2.1942, 2.1703, 1.6211], + device='cuda:0'), covar=tensor([0.0447, 0.0517, 0.0540, 0.0761, 0.0927, 0.0423, 0.0405, 0.0932], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0121, 0.0125, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:58:35,983 INFO [finetune.py:976] (0/7) Epoch 20, batch 5450, loss[loss=0.1677, simple_loss=0.2457, pruned_loss=0.04487, over 4764.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2448, pruned_loss=0.05204, over 952835.46 frames. ], batch size: 59, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:58:51,630 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114302.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:59:05,064 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:59:08,645 INFO [finetune.py:976] (0/7) Epoch 20, batch 5500, loss[loss=0.138, simple_loss=0.2094, pruned_loss=0.03329, over 4761.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.243, pruned_loss=0.05173, over 953244.87 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:59:09,355 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3437, 3.7375, 3.9397, 4.1501, 4.0949, 3.8516, 4.4472, 1.2950], + device='cuda:0'), covar=tensor([0.0837, 0.0967, 0.0867, 0.1004, 0.1296, 0.1553, 0.0695, 0.5843], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0243, 0.0278, 0.0291, 0.0331, 0.0285, 0.0303, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 00:59:23,088 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:59:27,721 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.524e+02 1.755e+02 2.254e+02 4.886e+02, threshold=3.510e+02, percent-clipped=3.0 +2023-03-27 00:59:42,371 INFO [finetune.py:976] (0/7) Epoch 20, batch 5550, loss[loss=0.1954, simple_loss=0.2765, pruned_loss=0.05721, over 4837.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2453, pruned_loss=0.05267, over 951260.92 frames. ], batch size: 47, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:59:49,159 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7605, 4.0272, 3.8197, 1.8843, 4.1426, 3.2163, 0.8263, 2.8880], + device='cuda:0'), covar=tensor([0.2500, 0.1897, 0.1399, 0.3529, 0.0881, 0.0916, 0.4726, 0.1458], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0130, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 01:00:14,060 INFO [finetune.py:976] (0/7) Epoch 20, batch 5600, loss[loss=0.1893, simple_loss=0.2685, pruned_loss=0.05503, over 4924.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2487, pruned_loss=0.05367, over 950433.86 frames. ], batch size: 42, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:00:20,976 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4297, 2.2931, 2.3371, 1.6469, 2.2955, 2.5087, 2.5498, 1.9066], + device='cuda:0'), covar=tensor([0.0520, 0.0575, 0.0625, 0.0887, 0.0595, 0.0619, 0.0521, 0.1025], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0139, 0.0121, 0.0124, 0.0139, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:00:22,691 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8410, 4.1896, 4.4717, 4.7306, 4.5980, 4.2974, 4.9753, 1.4852], + device='cuda:0'), covar=tensor([0.0700, 0.0806, 0.0838, 0.0768, 0.1161, 0.1549, 0.0526, 0.5723], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0242, 0.0277, 0.0290, 0.0330, 0.0284, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:00:31,887 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.587e+02 1.911e+02 2.256e+02 4.682e+02, threshold=3.822e+02, percent-clipped=4.0 +2023-03-27 01:00:43,489 INFO [finetune.py:976] (0/7) Epoch 20, batch 5650, loss[loss=0.1734, simple_loss=0.2422, pruned_loss=0.05234, over 4854.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2505, pruned_loss=0.0539, over 951946.06 frames. ], batch size: 31, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:00:46,978 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:01:13,265 INFO [finetune.py:976] (0/7) Epoch 20, batch 5700, loss[loss=0.1521, simple_loss=0.2182, pruned_loss=0.04296, over 4189.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2476, pruned_loss=0.05374, over 939128.58 frames. ], batch size: 18, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:01:26,934 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9716, 1.8636, 1.8138, 1.8996, 1.7415, 3.7959, 1.9369, 2.5044], + device='cuda:0'), covar=tensor([0.3693, 0.2962, 0.2195, 0.2831, 0.1537, 0.0245, 0.2217, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:01:29,353 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-20.pt +2023-03-27 01:01:39,095 INFO [finetune.py:976] (0/7) Epoch 21, batch 0, loss[loss=0.1492, simple_loss=0.2212, pruned_loss=0.03863, over 4701.00 frames. ], tot_loss[loss=0.1492, simple_loss=0.2212, pruned_loss=0.03863, over 4701.00 frames. ], batch size: 23, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:01:39,096 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 01:01:46,275 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8469, 3.4572, 3.5589, 3.7574, 3.5798, 3.4374, 3.9202, 1.3262], + device='cuda:0'), covar=tensor([0.0839, 0.0759, 0.0872, 0.0853, 0.1420, 0.1595, 0.0756, 0.5157], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0241, 0.0277, 0.0289, 0.0330, 0.0283, 0.0301, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:01:55,194 INFO [finetune.py:1010] (0/7) Epoch 21, validation: loss=0.1598, simple_loss=0.2277, pruned_loss=0.0459, over 2265189.00 frames. +2023-03-27 01:01:55,195 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 01:01:56,946 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.573e+01 1.356e+02 1.658e+02 2.014e+02 3.472e+02, threshold=3.316e+02, percent-clipped=0.0 +2023-03-27 01:02:47,783 INFO [finetune.py:976] (0/7) Epoch 21, batch 50, loss[loss=0.1537, simple_loss=0.2231, pruned_loss=0.04217, over 4809.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2511, pruned_loss=0.05588, over 214631.19 frames. ], batch size: 41, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:03:04,146 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1268, 1.9928, 1.7107, 1.8443, 2.0951, 1.7722, 2.2560, 2.1450], + device='cuda:0'), covar=tensor([0.1411, 0.2016, 0.2988, 0.2587, 0.2723, 0.1752, 0.3039, 0.1915], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0236, 0.0255, 0.0248, 0.0205, 0.0216, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:03:08,921 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6073, 4.0725, 3.7969, 1.8368, 4.0780, 3.0577, 1.1182, 2.9569], + device='cuda:0'), covar=tensor([0.2539, 0.1694, 0.1492, 0.3146, 0.0881, 0.0939, 0.4000, 0.1362], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0129, 0.0161, 0.0122, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 01:03:21,578 INFO [finetune.py:976] (0/7) Epoch 21, batch 100, loss[loss=0.1944, simple_loss=0.25, pruned_loss=0.06941, over 4379.00 frames. ], tot_loss[loss=0.174, simple_loss=0.243, pruned_loss=0.05254, over 379379.83 frames. ], batch size: 19, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:03:23,376 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.865e+01 1.560e+02 1.971e+02 2.354e+02 5.080e+02, threshold=3.943e+02, percent-clipped=2.0 +2023-03-27 01:03:54,252 INFO [finetune.py:976] (0/7) Epoch 21, batch 150, loss[loss=0.1953, simple_loss=0.2572, pruned_loss=0.06669, over 4917.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2399, pruned_loss=0.05286, over 505662.38 frames. ], batch size: 36, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:04:02,505 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:04:06,870 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-27 01:04:26,908 INFO [finetune.py:976] (0/7) Epoch 21, batch 200, loss[loss=0.1879, simple_loss=0.2595, pruned_loss=0.05813, over 4818.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2386, pruned_loss=0.05173, over 605934.36 frames. ], batch size: 39, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:04:29,191 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.562e+02 1.886e+02 2.300e+02 5.249e+02, threshold=3.772e+02, percent-clipped=1.0 +2023-03-27 01:04:32,218 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4256, 1.5102, 1.2710, 1.4085, 1.9099, 1.7031, 1.5465, 1.4107], + device='cuda:0'), covar=tensor([0.0363, 0.0281, 0.0568, 0.0314, 0.0178, 0.0508, 0.0321, 0.0346], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0108, 0.0145, 0.0112, 0.0101, 0.0112, 0.0101, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.5859e-05, 8.3034e-05, 1.1403e-04, 8.5990e-05, 7.8267e-05, 8.2721e-05, + 7.5029e-05, 8.6315e-05], device='cuda:0') +2023-03-27 01:04:42,946 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:04:46,572 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:05:00,770 INFO [finetune.py:976] (0/7) Epoch 21, batch 250, loss[loss=0.1809, simple_loss=0.2517, pruned_loss=0.05505, over 4830.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.244, pruned_loss=0.05292, over 685150.88 frames. ], batch size: 30, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:05:16,394 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0588, 1.0321, 1.0235, 0.5371, 0.9346, 1.1914, 1.1895, 1.0101], + device='cuda:0'), covar=tensor([0.0823, 0.0547, 0.0544, 0.0482, 0.0559, 0.0623, 0.0405, 0.0640], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0122, 0.0129, 0.0128, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9384e-05, 1.0775e-04, 8.9071e-05, 8.6334e-05, 9.0617e-05, 9.1642e-05, + 1.0147e-04, 1.0593e-04], device='cuda:0') +2023-03-27 01:05:19,226 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114831.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:05:33,146 INFO [finetune.py:976] (0/7) Epoch 21, batch 300, loss[loss=0.2096, simple_loss=0.2867, pruned_loss=0.06626, over 4829.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.247, pruned_loss=0.05319, over 746588.87 frames. ], batch size: 47, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:05:36,362 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.500e+02 1.787e+02 2.137e+02 3.935e+02, threshold=3.575e+02, percent-clipped=3.0 +2023-03-27 01:05:46,346 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114871.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:06:06,528 INFO [finetune.py:976] (0/7) Epoch 21, batch 350, loss[loss=0.1669, simple_loss=0.2479, pruned_loss=0.04295, over 4918.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2495, pruned_loss=0.05432, over 787666.38 frames. ], batch size: 38, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:06:26,471 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114932.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:06:36,553 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9157, 2.4316, 3.0528, 2.0037, 2.6657, 3.0612, 2.2574, 3.2422], + device='cuda:0'), covar=tensor([0.1339, 0.2066, 0.1514, 0.2261, 0.1264, 0.1441, 0.2661, 0.0779], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0206, 0.0190, 0.0189, 0.0174, 0.0213, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:06:40,072 INFO [finetune.py:976] (0/7) Epoch 21, batch 400, loss[loss=0.1596, simple_loss=0.2351, pruned_loss=0.04207, over 4809.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2514, pruned_loss=0.05375, over 827469.47 frames. ], batch size: 25, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:06:41,872 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.627e+02 1.950e+02 2.383e+02 4.205e+02, threshold=3.900e+02, percent-clipped=3.0 +2023-03-27 01:07:20,653 INFO [finetune.py:976] (0/7) Epoch 21, batch 450, loss[loss=0.1482, simple_loss=0.2106, pruned_loss=0.04291, over 4821.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2504, pruned_loss=0.05356, over 855334.75 frames. ], batch size: 33, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:07:40,719 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-27 01:08:11,192 INFO [finetune.py:976] (0/7) Epoch 21, batch 500, loss[loss=0.1639, simple_loss=0.2312, pruned_loss=0.04835, over 4786.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2483, pruned_loss=0.05297, over 879618.21 frames. ], batch size: 29, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:08:13,015 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.445e+02 1.728e+02 2.124e+02 2.919e+02, threshold=3.456e+02, percent-clipped=0.0 +2023-03-27 01:08:15,066 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 01:08:20,679 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1352, 3.6106, 3.7962, 4.0034, 3.9398, 3.6824, 4.2334, 1.3170], + device='cuda:0'), covar=tensor([0.0837, 0.0850, 0.0890, 0.1060, 0.1122, 0.1545, 0.0674, 0.5493], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0241, 0.0277, 0.0290, 0.0329, 0.0281, 0.0302, 0.0296], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:08:24,204 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:08:33,763 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115086.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:08:45,009 INFO [finetune.py:976] (0/7) Epoch 21, batch 550, loss[loss=0.1573, simple_loss=0.2179, pruned_loss=0.0484, over 4796.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2458, pruned_loss=0.05278, over 897029.14 frames. ], batch size: 25, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:08:57,064 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-27 01:09:14,145 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:09:18,263 INFO [finetune.py:976] (0/7) Epoch 21, batch 600, loss[loss=0.1944, simple_loss=0.2589, pruned_loss=0.06498, over 4088.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.247, pruned_loss=0.0534, over 908834.79 frames. ], batch size: 65, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:09:20,110 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.558e+02 1.835e+02 2.263e+02 4.639e+02, threshold=3.670e+02, percent-clipped=5.0 +2023-03-27 01:09:44,670 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1242, 2.0195, 1.6530, 2.0039, 1.9062, 1.8879, 1.9636, 2.6015], + device='cuda:0'), covar=tensor([0.3895, 0.4277, 0.3495, 0.4076, 0.4309, 0.2518, 0.4048, 0.1779], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0262, 0.0230, 0.0275, 0.0252, 0.0222, 0.0251, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:09:51,868 INFO [finetune.py:976] (0/7) Epoch 21, batch 650, loss[loss=0.1389, simple_loss=0.2113, pruned_loss=0.03322, over 4784.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2492, pruned_loss=0.05362, over 920538.76 frames. ], batch size: 26, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:09:58,641 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6992, 1.6057, 1.3894, 1.7779, 2.2483, 1.8293, 1.5909, 1.3637], + device='cuda:0'), covar=tensor([0.2177, 0.1996, 0.1921, 0.1602, 0.1637, 0.1220, 0.2363, 0.1848], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0213, 0.0195, 0.0243, 0.0188, 0.0218, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:10:07,425 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115227.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:10:25,138 INFO [finetune.py:976] (0/7) Epoch 21, batch 700, loss[loss=0.2251, simple_loss=0.293, pruned_loss=0.07863, over 4919.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2499, pruned_loss=0.05351, over 927478.18 frames. ], batch size: 38, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:10:26,912 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.664e+02 1.957e+02 2.283e+02 3.730e+02, threshold=3.914e+02, percent-clipped=1.0 +2023-03-27 01:10:41,426 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.30 vs. limit=5.0 +2023-03-27 01:10:47,082 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-03-27 01:10:58,919 INFO [finetune.py:976] (0/7) Epoch 21, batch 750, loss[loss=0.2045, simple_loss=0.2892, pruned_loss=0.05983, over 4810.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2499, pruned_loss=0.05318, over 933216.81 frames. ], batch size: 45, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:11:24,718 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2455, 2.2298, 2.2804, 1.5278, 2.1777, 2.3546, 2.3537, 1.8339], + device='cuda:0'), covar=tensor([0.0633, 0.0661, 0.0684, 0.0953, 0.0693, 0.0714, 0.0626, 0.1153], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0137, 0.0141, 0.0121, 0.0125, 0.0140, 0.0141, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:11:31,745 INFO [finetune.py:976] (0/7) Epoch 21, batch 800, loss[loss=0.1481, simple_loss=0.2226, pruned_loss=0.03674, over 4784.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2478, pruned_loss=0.05227, over 936295.86 frames. ], batch size: 25, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:11:33,561 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.426e+02 1.720e+02 2.044e+02 3.360e+02, threshold=3.441e+02, percent-clipped=0.0 +2023-03-27 01:11:41,495 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115370.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:11:42,120 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:11:42,711 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115372.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:11:56,034 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4294, 1.3996, 1.5744, 2.4786, 1.7170, 2.2506, 0.9087, 2.1275], + device='cuda:0'), covar=tensor([0.1848, 0.1420, 0.1249, 0.0787, 0.0889, 0.1125, 0.1696, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0100, 0.0135, 0.0124, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 01:12:00,994 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 01:12:04,583 INFO [finetune.py:976] (0/7) Epoch 21, batch 850, loss[loss=0.1797, simple_loss=0.2422, pruned_loss=0.05864, over 4794.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2458, pruned_loss=0.05168, over 939833.06 frames. ], batch size: 51, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:12:14,848 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:24,430 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115431.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:25,033 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:41,966 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115442.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:52,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7334, 1.2909, 0.9292, 1.6980, 2.1167, 1.4527, 1.6094, 1.6775], + device='cuda:0'), covar=tensor([0.1390, 0.1951, 0.1956, 0.1091, 0.1803, 0.2116, 0.1264, 0.1834], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0093, 0.0097, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 01:12:54,669 INFO [finetune.py:976] (0/7) Epoch 21, batch 900, loss[loss=0.1437, simple_loss=0.2165, pruned_loss=0.03545, over 4835.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2426, pruned_loss=0.05045, over 943453.24 frames. ], batch size: 33, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:13:00,781 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.494e+02 1.776e+02 2.140e+02 4.219e+02, threshold=3.551e+02, percent-clipped=3.0 +2023-03-27 01:13:17,556 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7462, 1.6302, 1.5718, 1.6748, 1.6117, 4.4188, 1.7561, 2.1065], + device='cuda:0'), covar=tensor([0.3255, 0.2412, 0.2137, 0.2373, 0.1541, 0.0118, 0.2374, 0.1221], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:13:37,347 INFO [finetune.py:976] (0/7) Epoch 21, batch 950, loss[loss=0.1426, simple_loss=0.2258, pruned_loss=0.02967, over 4813.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2428, pruned_loss=0.05136, over 946225.66 frames. ], batch size: 45, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:13:42,865 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8775, 2.0107, 1.3613, 2.0662, 1.9444, 1.5878, 2.6934, 1.9718], + device='cuda:0'), covar=tensor([0.1452, 0.1878, 0.3231, 0.2872, 0.2651, 0.1655, 0.2088, 0.1828], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0188, 0.0236, 0.0254, 0.0247, 0.0204, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:13:51,930 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115527.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:14:01,362 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8448, 1.7692, 1.5387, 1.9866, 2.2174, 1.9951, 1.6949, 1.4539], + device='cuda:0'), covar=tensor([0.2217, 0.2022, 0.2067, 0.1602, 0.1914, 0.1154, 0.2397, 0.1898], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0194, 0.0243, 0.0188, 0.0217, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:14:11,292 INFO [finetune.py:976] (0/7) Epoch 21, batch 1000, loss[loss=0.2188, simple_loss=0.2875, pruned_loss=0.07505, over 4819.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2454, pruned_loss=0.05232, over 948739.50 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:14:13,112 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.550e+02 1.849e+02 2.159e+02 3.452e+02, threshold=3.698e+02, percent-clipped=0.0 +2023-03-27 01:14:15,061 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6405, 1.2652, 0.9107, 1.4991, 1.9756, 1.1687, 1.4384, 1.5029], + device='cuda:0'), covar=tensor([0.1529, 0.2089, 0.1890, 0.1221, 0.2096, 0.2193, 0.1503, 0.2110], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 01:14:24,460 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:14:44,023 INFO [finetune.py:976] (0/7) Epoch 21, batch 1050, loss[loss=0.1313, simple_loss=0.2007, pruned_loss=0.03099, over 4781.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2485, pruned_loss=0.05344, over 951104.43 frames. ], batch size: 26, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:14:47,763 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3553, 1.4141, 1.4937, 0.7076, 1.4706, 1.7309, 1.7134, 1.3204], + device='cuda:0'), covar=tensor([0.0921, 0.0642, 0.0467, 0.0579, 0.0423, 0.0615, 0.0363, 0.0664], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0126, 0.0124, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0346e-05, 1.0847e-04, 9.0029e-05, 8.7604e-05, 9.1774e-05, 9.2256e-05, + 1.0227e-04, 1.0661e-04], device='cuda:0') +2023-03-27 01:15:15,579 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8955, 4.2604, 4.4574, 4.6777, 4.6052, 4.3693, 5.0091, 1.5494], + device='cuda:0'), covar=tensor([0.0694, 0.0743, 0.0751, 0.0965, 0.1122, 0.1492, 0.0527, 0.5614], + device='cuda:0'), in_proj_covar=tensor([0.0343, 0.0237, 0.0274, 0.0287, 0.0325, 0.0278, 0.0299, 0.0294], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:15:16,670 INFO [finetune.py:976] (0/7) Epoch 21, batch 1100, loss[loss=0.196, simple_loss=0.2649, pruned_loss=0.06352, over 4839.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2491, pruned_loss=0.05354, over 951025.69 frames. ], batch size: 49, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:15:19,443 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.582e+02 1.822e+02 2.328e+02 4.675e+02, threshold=3.643e+02, percent-clipped=4.0 +2023-03-27 01:15:31,170 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 01:15:50,437 INFO [finetune.py:976] (0/7) Epoch 21, batch 1150, loss[loss=0.1259, simple_loss=0.186, pruned_loss=0.03296, over 4315.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.249, pruned_loss=0.05335, over 950638.62 frames. ], batch size: 19, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:16:05,308 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:05,453 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-27 01:16:05,915 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115727.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:14,918 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115742.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:24,029 INFO [finetune.py:976] (0/7) Epoch 21, batch 1200, loss[loss=0.1844, simple_loss=0.2588, pruned_loss=0.05498, over 4891.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2474, pruned_loss=0.0526, over 951489.72 frames. ], batch size: 36, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:16:25,821 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.465e+02 1.737e+02 2.048e+02 4.574e+02, threshold=3.475e+02, percent-clipped=2.0 +2023-03-27 01:16:35,971 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0324, 1.5562, 2.3731, 1.5561, 2.0684, 2.3015, 1.5307, 2.3881], + device='cuda:0'), covar=tensor([0.1328, 0.2353, 0.1292, 0.2029, 0.0974, 0.1252, 0.3265, 0.0773], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0203, 0.0190, 0.0189, 0.0174, 0.0212, 0.0217, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:16:47,357 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115790.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:56,822 INFO [finetune.py:976] (0/7) Epoch 21, batch 1250, loss[loss=0.1338, simple_loss=0.2098, pruned_loss=0.0289, over 4833.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2453, pruned_loss=0.05198, over 952401.20 frames. ], batch size: 30, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:17:29,532 INFO [finetune.py:976] (0/7) Epoch 21, batch 1300, loss[loss=0.1629, simple_loss=0.2383, pruned_loss=0.04371, over 4711.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2434, pruned_loss=0.05145, over 952515.19 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:17:32,370 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.563e+02 1.759e+02 2.181e+02 4.124e+02, threshold=3.519e+02, percent-clipped=1.0 +2023-03-27 01:18:12,455 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115892.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:18:23,979 INFO [finetune.py:976] (0/7) Epoch 21, batch 1350, loss[loss=0.1813, simple_loss=0.245, pruned_loss=0.05876, over 4779.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2437, pruned_loss=0.05121, over 953383.84 frames. ], batch size: 26, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:18:56,354 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1077, 2.1931, 1.7153, 2.0456, 2.0223, 2.0403, 2.0409, 2.9707], + device='cuda:0'), covar=tensor([0.3788, 0.4384, 0.3653, 0.4424, 0.4798, 0.2489, 0.4188, 0.1537], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0261, 0.0231, 0.0274, 0.0251, 0.0221, 0.0251, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:19:00,568 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115953.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:19:01,043 INFO [finetune.py:976] (0/7) Epoch 21, batch 1400, loss[loss=0.179, simple_loss=0.26, pruned_loss=0.049, over 4818.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2468, pruned_loss=0.05205, over 953689.55 frames. ], batch size: 39, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:19:02,861 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.541e+02 1.818e+02 2.071e+02 3.575e+02, threshold=3.635e+02, percent-clipped=1.0 +2023-03-27 01:19:31,954 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-116000.pt +2023-03-27 01:19:35,504 INFO [finetune.py:976] (0/7) Epoch 21, batch 1450, loss[loss=0.1567, simple_loss=0.2307, pruned_loss=0.0414, over 4835.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2495, pruned_loss=0.05326, over 953283.05 frames. ], batch size: 30, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:19:35,669 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 01:19:43,002 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:19:51,746 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:19:52,334 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:09,068 INFO [finetune.py:976] (0/7) Epoch 21, batch 1500, loss[loss=0.1484, simple_loss=0.2221, pruned_loss=0.03729, over 4751.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2504, pruned_loss=0.05335, over 954540.41 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:20:10,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.719e+02 2.058e+02 2.312e+02 4.180e+02, threshold=4.116e+02, percent-clipped=2.0 +2023-03-27 01:20:23,162 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116074.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:23,792 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:23,853 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116075.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:20:42,654 INFO [finetune.py:976] (0/7) Epoch 21, batch 1550, loss[loss=0.1671, simple_loss=0.2302, pruned_loss=0.05198, over 4823.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2496, pruned_loss=0.0532, over 953531.74 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:20:43,364 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:50,860 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116116.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:15,951 INFO [finetune.py:976] (0/7) Epoch 21, batch 1600, loss[loss=0.1577, simple_loss=0.234, pruned_loss=0.04075, over 4923.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2475, pruned_loss=0.05239, over 955405.74 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:21:17,777 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.061e+01 1.555e+02 1.841e+02 2.223e+02 4.654e+02, threshold=3.683e+02, percent-clipped=1.0 +2023-03-27 01:21:23,354 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116166.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:31,967 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116177.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:36,052 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8255, 1.1837, 0.8930, 1.6091, 2.1357, 1.3618, 1.4551, 1.5909], + device='cuda:0'), covar=tensor([0.1415, 0.2045, 0.1830, 0.1125, 0.1863, 0.1987, 0.1440, 0.1888], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0120, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 01:21:49,856 INFO [finetune.py:976] (0/7) Epoch 21, batch 1650, loss[loss=0.2062, simple_loss=0.2586, pruned_loss=0.07687, over 4718.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2445, pruned_loss=0.05127, over 955444.29 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:22:19,459 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116248.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:22:23,485 INFO [finetune.py:976] (0/7) Epoch 21, batch 1700, loss[loss=0.1993, simple_loss=0.2689, pruned_loss=0.06483, over 4917.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2421, pruned_loss=0.05069, over 956694.55 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:22:25,324 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.524e+01 1.502e+02 1.774e+02 2.116e+02 3.203e+02, threshold=3.548e+02, percent-clipped=0.0 +2023-03-27 01:22:29,054 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6209, 1.4880, 2.3018, 1.9775, 1.8117, 4.1634, 1.4205, 1.7461], + device='cuda:0'), covar=tensor([0.0940, 0.1828, 0.1193, 0.0929, 0.1607, 0.0236, 0.1603, 0.1787], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0074, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:22:59,246 INFO [finetune.py:976] (0/7) Epoch 21, batch 1750, loss[loss=0.1541, simple_loss=0.2351, pruned_loss=0.0366, over 4861.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2434, pruned_loss=0.05087, over 955805.90 frames. ], batch size: 31, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:23:19,669 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116323.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:23:43,284 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8260, 2.4555, 2.3256, 2.7432, 2.5205, 2.5456, 2.4662, 3.4349], + device='cuda:0'), covar=tensor([0.3830, 0.5360, 0.3572, 0.4287, 0.4407, 0.2623, 0.4432, 0.1661], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0262, 0.0232, 0.0276, 0.0253, 0.0222, 0.0252, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:23:58,748 INFO [finetune.py:976] (0/7) Epoch 21, batch 1800, loss[loss=0.1867, simple_loss=0.2561, pruned_loss=0.05861, over 4822.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2471, pruned_loss=0.05267, over 956084.56 frames. ], batch size: 40, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:24:00,588 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.627e+02 1.938e+02 2.363e+02 5.057e+02, threshold=3.876e+02, percent-clipped=3.0 +2023-03-27 01:24:08,546 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116370.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:24:18,595 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:24:31,788 INFO [finetune.py:976] (0/7) Epoch 21, batch 1850, loss[loss=0.1324, simple_loss=0.1971, pruned_loss=0.03385, over 4296.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2479, pruned_loss=0.05283, over 955532.19 frames. ], batch size: 18, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:00,257 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:05,404 INFO [finetune.py:976] (0/7) Epoch 21, batch 1900, loss[loss=0.1905, simple_loss=0.2593, pruned_loss=0.06087, over 4884.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2487, pruned_loss=0.05286, over 956727.61 frames. ], batch size: 32, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:07,228 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.572e+02 1.880e+02 2.123e+02 3.861e+02, threshold=3.760e+02, percent-clipped=0.0 +2023-03-27 01:25:10,213 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116461.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:16,977 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:38,739 INFO [finetune.py:976] (0/7) Epoch 21, batch 1950, loss[loss=0.1494, simple_loss=0.2275, pruned_loss=0.03559, over 4883.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2484, pruned_loss=0.05217, over 959015.42 frames. ], batch size: 35, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:39,475 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:40,113 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8970, 1.2672, 2.0225, 1.9156, 1.7537, 1.7244, 1.8522, 1.8434], + device='cuda:0'), covar=tensor([0.3859, 0.3911, 0.3174, 0.3662, 0.4636, 0.3602, 0.4290, 0.3093], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0242, 0.0263, 0.0281, 0.0280, 0.0256, 0.0290, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:25:40,698 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116507.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:26:07,964 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116548.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:26:11,459 INFO [finetune.py:976] (0/7) Epoch 21, batch 2000, loss[loss=0.1633, simple_loss=0.2251, pruned_loss=0.05075, over 4853.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2463, pruned_loss=0.05181, over 957468.85 frames. ], batch size: 44, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:26:13,787 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.123e+01 1.433e+02 1.690e+02 2.067e+02 3.885e+02, threshold=3.380e+02, percent-clipped=2.0 +2023-03-27 01:26:19,735 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:26:25,393 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 01:26:39,350 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116596.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:26:44,681 INFO [finetune.py:976] (0/7) Epoch 21, batch 2050, loss[loss=0.1416, simple_loss=0.2148, pruned_loss=0.03416, over 4755.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2426, pruned_loss=0.05047, over 958326.24 frames. ], batch size: 27, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:26:54,204 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116618.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:05,167 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2965, 2.8167, 2.7044, 1.2296, 2.9481, 2.1732, 0.6710, 1.8558], + device='cuda:0'), covar=tensor([0.2009, 0.2226, 0.1745, 0.3564, 0.1359, 0.1154, 0.4225, 0.1622], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0179, 0.0159, 0.0130, 0.0162, 0.0123, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 01:27:07,672 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6390, 1.6557, 1.4137, 1.5420, 1.9602, 1.8527, 1.6583, 1.4414], + device='cuda:0'), covar=tensor([0.0326, 0.0292, 0.0596, 0.0329, 0.0236, 0.0551, 0.0286, 0.0398], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0108, 0.0146, 0.0113, 0.0101, 0.0112, 0.0101, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.6436e-05, 8.2925e-05, 1.1475e-04, 8.6878e-05, 7.8489e-05, 8.2656e-05, + 7.4816e-05, 8.7013e-05], device='cuda:0') +2023-03-27 01:27:08,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0224, 2.0255, 1.9699, 0.9138, 2.3119, 2.4882, 2.0700, 1.8469], + device='cuda:0'), covar=tensor([0.1012, 0.0641, 0.0564, 0.0688, 0.0418, 0.0640, 0.0456, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0125, 0.0123, 0.0130, 0.0128, 0.0141, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.9530e-05, 1.0783e-04, 8.9361e-05, 8.7244e-05, 9.1267e-05, 9.1622e-05, + 1.0122e-04, 1.0574e-04], device='cuda:0') +2023-03-27 01:27:14,580 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-27 01:27:18,446 INFO [finetune.py:976] (0/7) Epoch 21, batch 2100, loss[loss=0.1513, simple_loss=0.2259, pruned_loss=0.03831, over 4760.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2433, pruned_loss=0.05161, over 956156.99 frames. ], batch size: 28, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:27:20,846 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.617e+02 1.783e+02 2.198e+02 6.495e+02, threshold=3.567e+02, percent-clipped=4.0 +2023-03-27 01:27:24,970 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6649, 2.2541, 1.9653, 0.9819, 2.2915, 1.9301, 1.6271, 2.0712], + device='cuda:0'), covar=tensor([0.1150, 0.1282, 0.2531, 0.2735, 0.1669, 0.2613, 0.2969, 0.1350], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0191, 0.0197, 0.0181, 0.0207, 0.0207, 0.0221, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:27:29,075 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116670.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 01:27:34,422 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:34,479 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:51,925 INFO [finetune.py:976] (0/7) Epoch 21, batch 2150, loss[loss=0.2079, simple_loss=0.2721, pruned_loss=0.07191, over 4904.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.248, pruned_loss=0.05351, over 956644.87 frames. ], batch size: 36, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:27:52,053 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:00,967 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:01,139 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 01:28:26,697 INFO [finetune.py:976] (0/7) Epoch 21, batch 2200, loss[loss=0.1919, simple_loss=0.2524, pruned_loss=0.06568, over 4076.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2485, pruned_loss=0.05353, over 955605.93 frames. ], batch size: 65, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:28:30,714 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.636e+02 2.054e+02 2.505e+02 6.138e+02, threshold=4.108e+02, percent-clipped=5.0 +2023-03-27 01:28:32,654 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116761.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:39,671 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116765.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:44,489 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116772.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:19,238 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116802.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 01:29:19,264 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116802.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:20,383 INFO [finetune.py:976] (0/7) Epoch 21, batch 2250, loss[loss=0.2122, simple_loss=0.2795, pruned_loss=0.07238, over 4893.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2503, pruned_loss=0.05419, over 956044.55 frames. ], batch size: 43, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:29:28,967 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116809.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:31,497 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-27 01:29:39,991 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:30:01,468 INFO [finetune.py:976] (0/7) Epoch 21, batch 2300, loss[loss=0.1671, simple_loss=0.2422, pruned_loss=0.04599, over 4911.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2505, pruned_loss=0.0539, over 956482.08 frames. ], batch size: 37, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:30:04,886 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.477e+02 1.740e+02 2.172e+02 4.454e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 01:30:06,812 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116861.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:30:08,061 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116863.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:30:34,138 INFO [finetune.py:976] (0/7) Epoch 21, batch 2350, loss[loss=0.163, simple_loss=0.2343, pruned_loss=0.0458, over 4766.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2477, pruned_loss=0.05292, over 956396.04 frames. ], batch size: 26, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:31:07,399 INFO [finetune.py:976] (0/7) Epoch 21, batch 2400, loss[loss=0.1736, simple_loss=0.2439, pruned_loss=0.05165, over 4821.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2456, pruned_loss=0.05235, over 957130.43 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:31:08,649 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6465, 1.6501, 2.0469, 3.3323, 2.3260, 2.3283, 0.7771, 2.8808], + device='cuda:0'), covar=tensor([0.1706, 0.1357, 0.1328, 0.0536, 0.0757, 0.1573, 0.1941, 0.0425], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0100, 0.0137, 0.0124, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 01:31:09,766 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.566e+02 1.863e+02 2.219e+02 3.648e+02, threshold=3.726e+02, percent-clipped=1.0 +2023-03-27 01:31:21,633 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:31:25,170 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116979.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:31:40,995 INFO [finetune.py:976] (0/7) Epoch 21, batch 2450, loss[loss=0.1773, simple_loss=0.2547, pruned_loss=0.04995, over 4821.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2423, pruned_loss=0.05118, over 955831.54 frames. ], batch size: 41, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:31:44,279 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-27 01:31:55,655 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7555, 2.7521, 2.5429, 1.9673, 2.7884, 2.9736, 2.9920, 2.3965], + device='cuda:0'), covar=tensor([0.0634, 0.0683, 0.0782, 0.0883, 0.0650, 0.0713, 0.0648, 0.1123], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0135, 0.0139, 0.0120, 0.0125, 0.0138, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:31:57,424 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:32:06,894 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4895, 2.3707, 2.1136, 1.0194, 2.2948, 1.8668, 1.7569, 2.1742], + device='cuda:0'), covar=tensor([0.1028, 0.0850, 0.1799, 0.2308, 0.1511, 0.2444, 0.2354, 0.1083], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0191, 0.0198, 0.0182, 0.0208, 0.0207, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:32:14,703 INFO [finetune.py:976] (0/7) Epoch 21, batch 2500, loss[loss=0.1933, simple_loss=0.2803, pruned_loss=0.05318, over 4821.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2438, pruned_loss=0.05193, over 955768.59 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:32:17,113 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.510e+02 1.797e+02 2.340e+02 3.968e+02, threshold=3.593e+02, percent-clipped=1.0 +2023-03-27 01:32:18,381 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:32:46,782 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:32:47,570 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 01:32:47,906 INFO [finetune.py:976] (0/7) Epoch 21, batch 2550, loss[loss=0.1264, simple_loss=0.191, pruned_loss=0.03092, over 4745.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2478, pruned_loss=0.05289, over 955333.79 frames. ], batch size: 23, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:33:02,844 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6906, 2.7716, 2.5632, 1.8613, 2.5830, 2.7565, 2.8272, 2.4508], + device='cuda:0'), covar=tensor([0.0610, 0.0579, 0.0667, 0.0860, 0.0609, 0.0691, 0.0567, 0.0896], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0134, 0.0138, 0.0119, 0.0124, 0.0138, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:33:19,398 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117150.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:33:21,774 INFO [finetune.py:976] (0/7) Epoch 21, batch 2600, loss[loss=0.2076, simple_loss=0.2717, pruned_loss=0.07173, over 4901.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2495, pruned_loss=0.05338, over 956523.46 frames. ], batch size: 35, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:33:24,216 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.533e+02 1.830e+02 2.226e+02 4.351e+02, threshold=3.661e+02, percent-clipped=3.0 +2023-03-27 01:33:24,296 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117158.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:33:26,116 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:33:38,468 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-27 01:33:38,962 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2207, 2.6691, 2.5427, 1.2775, 2.8032, 2.2667, 2.2284, 2.4588], + device='cuda:0'), covar=tensor([0.0978, 0.0839, 0.1686, 0.2208, 0.1455, 0.2372, 0.2051, 0.1045], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0190, 0.0196, 0.0181, 0.0207, 0.0207, 0.0220, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:34:06,271 INFO [finetune.py:976] (0/7) Epoch 21, batch 2650, loss[loss=0.1412, simple_loss=0.2039, pruned_loss=0.03921, over 4771.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2493, pruned_loss=0.05328, over 955542.87 frames. ], batch size: 26, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:34:09,386 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117209.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:34:26,095 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:35:03,766 INFO [finetune.py:976] (0/7) Epoch 21, batch 2700, loss[loss=0.1773, simple_loss=0.2386, pruned_loss=0.05798, over 4822.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2486, pruned_loss=0.05258, over 952925.33 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:35:06,200 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.522e+02 1.732e+02 2.127e+02 4.053e+02, threshold=3.464e+02, percent-clipped=3.0 +2023-03-27 01:35:23,783 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117274.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:35:30,659 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:35:41,273 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3538, 2.2656, 1.8939, 2.3558, 2.1466, 2.1542, 2.1178, 3.0201], + device='cuda:0'), covar=tensor([0.3517, 0.4687, 0.3310, 0.4169, 0.4379, 0.2392, 0.4391, 0.1608], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0261, 0.0232, 0.0276, 0.0252, 0.0222, 0.0251, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:35:45,276 INFO [finetune.py:976] (0/7) Epoch 21, batch 2750, loss[loss=0.1938, simple_loss=0.2471, pruned_loss=0.07028, over 4822.00 frames. ], tot_loss[loss=0.175, simple_loss=0.246, pruned_loss=0.052, over 954660.04 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:35:54,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4913, 2.4310, 2.3850, 1.8340, 2.3522, 2.6492, 2.5864, 2.0523], + device='cuda:0'), covar=tensor([0.0574, 0.0654, 0.0634, 0.0830, 0.0955, 0.0636, 0.0530, 0.1078], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0121, 0.0125, 0.0139, 0.0141, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:35:56,237 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117322.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:01,581 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:18,643 INFO [finetune.py:976] (0/7) Epoch 21, batch 2800, loss[loss=0.1728, simple_loss=0.2348, pruned_loss=0.05535, over 4795.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2431, pruned_loss=0.05114, over 954519.48 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:36:21,562 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.495e+02 1.752e+02 2.115e+02 2.888e+02, threshold=3.503e+02, percent-clipped=0.0 +2023-03-27 01:36:22,905 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117360.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:33,229 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2940, 2.1272, 2.2163, 1.6215, 2.1444, 2.2581, 2.3182, 1.7652], + device='cuda:0'), covar=tensor([0.0561, 0.0607, 0.0693, 0.0873, 0.0643, 0.0711, 0.0544, 0.1206], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0121, 0.0126, 0.0140, 0.0141, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:36:41,133 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5905, 1.5276, 1.5457, 0.8140, 1.6627, 1.7679, 1.7782, 1.3806], + device='cuda:0'), covar=tensor([0.0892, 0.0667, 0.0512, 0.0520, 0.0442, 0.0657, 0.0388, 0.0738], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0126, 0.0124, 0.0131, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.0390e-05, 1.0882e-04, 9.0068e-05, 8.7706e-05, 9.1954e-05, 9.1912e-05, + 1.0195e-04, 1.0661e-04], device='cuda:0') +2023-03-27 01:36:42,959 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:47,685 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1679, 1.8855, 1.8636, 0.9307, 2.2302, 2.2792, 2.1023, 1.7934], + device='cuda:0'), covar=tensor([0.0843, 0.0671, 0.0565, 0.0591, 0.0549, 0.0616, 0.0496, 0.0695], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0126, 0.0124, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0246e-05, 1.0870e-04, 8.9987e-05, 8.7600e-05, 9.1822e-05, 9.1823e-05, + 1.0174e-04, 1.0651e-04], device='cuda:0') +2023-03-27 01:36:52,478 INFO [finetune.py:976] (0/7) Epoch 21, batch 2850, loss[loss=0.1514, simple_loss=0.2384, pruned_loss=0.03217, over 4908.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2413, pruned_loss=0.05051, over 955027.89 frames. ], batch size: 32, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:36:54,968 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117408.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:37:15,674 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 01:37:25,543 INFO [finetune.py:976] (0/7) Epoch 21, batch 2900, loss[loss=0.1486, simple_loss=0.2233, pruned_loss=0.03694, over 4722.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2433, pruned_loss=0.0513, over 955434.88 frames. ], batch size: 23, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:37:28,395 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.554e+02 1.875e+02 2.295e+02 6.888e+02, threshold=3.749e+02, percent-clipped=2.0 +2023-03-27 01:37:28,495 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117458.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:37:42,392 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0812, 0.9767, 0.9486, 1.1461, 1.2078, 1.2055, 0.9986, 0.9610], + device='cuda:0'), covar=tensor([0.0340, 0.0318, 0.0629, 0.0320, 0.0289, 0.0396, 0.0329, 0.0423], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0106, 0.0143, 0.0111, 0.0099, 0.0110, 0.0099, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5387e-05, 8.1691e-05, 1.1275e-04, 8.5494e-05, 7.6779e-05, 8.1050e-05, + 7.4074e-05, 8.5425e-05], device='cuda:0') +2023-03-27 01:37:59,199 INFO [finetune.py:976] (0/7) Epoch 21, batch 2950, loss[loss=0.1735, simple_loss=0.2425, pruned_loss=0.05223, over 4823.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2457, pruned_loss=0.0521, over 955771.10 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:38:00,488 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:38:15,406 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0583, 2.0647, 1.7982, 2.3960, 2.1577, 1.8999, 2.5259, 2.1640], + device='cuda:0'), covar=tensor([0.1323, 0.2123, 0.2645, 0.1998, 0.2205, 0.1475, 0.2480, 0.1679], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0187, 0.0235, 0.0253, 0.0247, 0.0203, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:38:32,247 INFO [finetune.py:976] (0/7) Epoch 21, batch 3000, loss[loss=0.2046, simple_loss=0.2823, pruned_loss=0.06341, over 4736.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2476, pruned_loss=0.05295, over 953404.60 frames. ], batch size: 54, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:38:32,248 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 01:38:34,031 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8515, 1.3436, 0.8898, 1.6770, 2.1237, 1.1415, 1.5674, 1.5534], + device='cuda:0'), covar=tensor([0.1306, 0.1830, 0.1714, 0.1118, 0.1797, 0.1921, 0.1275, 0.1882], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 01:38:42,801 INFO [finetune.py:1010] (0/7) Epoch 21, validation: loss=0.1567, simple_loss=0.2253, pruned_loss=0.04408, over 2265189.00 frames. +2023-03-27 01:38:42,801 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 01:38:45,675 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.534e+02 1.924e+02 2.362e+02 3.621e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-27 01:39:00,137 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:39:06,491 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 01:39:17,642 INFO [finetune.py:976] (0/7) Epoch 21, batch 3050, loss[loss=0.1829, simple_loss=0.252, pruned_loss=0.05687, over 4908.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2479, pruned_loss=0.0526, over 954094.05 frames. ], batch size: 37, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:39:56,379 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3899, 1.2731, 1.5496, 1.6047, 1.4919, 3.0026, 1.1354, 1.4036], + device='cuda:0'), covar=tensor([0.1101, 0.2443, 0.1215, 0.1072, 0.1943, 0.0319, 0.2144, 0.2365], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0074, 0.0076, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:40:13,670 INFO [finetune.py:976] (0/7) Epoch 21, batch 3100, loss[loss=0.1261, simple_loss=0.1973, pruned_loss=0.02749, over 4827.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2467, pruned_loss=0.05221, over 955318.51 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:40:19,669 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.482e+02 1.759e+02 2.208e+02 4.258e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-27 01:40:46,800 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:40:58,311 INFO [finetune.py:976] (0/7) Epoch 21, batch 3150, loss[loss=0.2002, simple_loss=0.2632, pruned_loss=0.06858, over 4734.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.245, pruned_loss=0.05197, over 954205.76 frames. ], batch size: 59, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:41:31,632 INFO [finetune.py:976] (0/7) Epoch 21, batch 3200, loss[loss=0.2351, simple_loss=0.2973, pruned_loss=0.08645, over 4799.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2422, pruned_loss=0.05099, over 953274.72 frames. ], batch size: 45, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:41:34,036 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.569e+02 1.801e+02 2.101e+02 4.822e+02, threshold=3.602e+02, percent-clipped=2.0 +2023-03-27 01:42:05,172 INFO [finetune.py:976] (0/7) Epoch 21, batch 3250, loss[loss=0.1641, simple_loss=0.2345, pruned_loss=0.04688, over 4928.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2422, pruned_loss=0.05115, over 952023.08 frames. ], batch size: 42, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:42:13,085 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-27 01:42:38,404 INFO [finetune.py:976] (0/7) Epoch 21, batch 3300, loss[loss=0.1573, simple_loss=0.2339, pruned_loss=0.0403, over 4795.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2461, pruned_loss=0.05282, over 953796.23 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:42:40,847 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.638e+02 1.917e+02 2.241e+02 9.038e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-27 01:42:54,847 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117877.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:43:11,570 INFO [finetune.py:976] (0/7) Epoch 21, batch 3350, loss[loss=0.1549, simple_loss=0.2261, pruned_loss=0.04187, over 4734.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2489, pruned_loss=0.05358, over 953386.54 frames. ], batch size: 23, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:43:25,771 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:43:45,043 INFO [finetune.py:976] (0/7) Epoch 21, batch 3400, loss[loss=0.1404, simple_loss=0.2266, pruned_loss=0.02712, over 4766.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2508, pruned_loss=0.05447, over 955516.98 frames. ], batch size: 28, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:43:47,450 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.619e+02 1.880e+02 2.233e+02 5.629e+02, threshold=3.761e+02, percent-clipped=2.0 +2023-03-27 01:44:05,851 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:16,066 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-118000.pt +2023-03-27 01:44:19,699 INFO [finetune.py:976] (0/7) Epoch 21, batch 3450, loss[loss=0.1862, simple_loss=0.2485, pruned_loss=0.06192, over 4801.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2512, pruned_loss=0.05437, over 955119.07 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:44:24,685 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6004, 1.5728, 1.3106, 1.5498, 2.0072, 1.9755, 1.5586, 1.4189], + device='cuda:0'), covar=tensor([0.0356, 0.0367, 0.0720, 0.0371, 0.0207, 0.0426, 0.0347, 0.0489], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0107, 0.0144, 0.0112, 0.0099, 0.0110, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.5750e-05, 8.1931e-05, 1.1312e-04, 8.5737e-05, 7.7186e-05, 8.1068e-05, + 7.4765e-05, 8.5812e-05], device='cuda:0') +2023-03-27 01:44:39,327 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:44,144 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:59,765 INFO [finetune.py:976] (0/7) Epoch 21, batch 3500, loss[loss=0.1542, simple_loss=0.2285, pruned_loss=0.03994, over 4764.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.248, pruned_loss=0.05309, over 954521.64 frames. ], batch size: 26, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:45:02,212 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.127e+01 1.501e+02 1.833e+02 2.184e+02 3.839e+02, threshold=3.666e+02, percent-clipped=2.0 +2023-03-27 01:45:52,328 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118101.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:45:57,919 INFO [finetune.py:976] (0/7) Epoch 21, batch 3550, loss[loss=0.1646, simple_loss=0.2327, pruned_loss=0.0482, over 4925.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2448, pruned_loss=0.05182, over 954157.10 frames. ], batch size: 42, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:46:30,096 INFO [finetune.py:976] (0/7) Epoch 21, batch 3600, loss[loss=0.1766, simple_loss=0.2372, pruned_loss=0.05796, over 4906.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2427, pruned_loss=0.05155, over 953204.74 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:46:33,043 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.558e+02 1.902e+02 2.180e+02 3.976e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-27 01:46:39,822 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118168.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:47:03,624 INFO [finetune.py:976] (0/7) Epoch 21, batch 3650, loss[loss=0.1645, simple_loss=0.2452, pruned_loss=0.04187, over 4757.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2438, pruned_loss=0.05162, over 955648.93 frames. ], batch size: 28, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:47:19,814 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118229.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:47:36,721 INFO [finetune.py:976] (0/7) Epoch 21, batch 3700, loss[loss=0.1749, simple_loss=0.2416, pruned_loss=0.05407, over 4837.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2467, pruned_loss=0.05256, over 956934.51 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:47:39,056 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.607e+02 1.941e+02 2.377e+02 3.454e+02, threshold=3.882e+02, percent-clipped=0.0 +2023-03-27 01:47:46,805 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:47:59,984 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118289.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:10,305 INFO [finetune.py:976] (0/7) Epoch 21, batch 3750, loss[loss=0.2107, simple_loss=0.2864, pruned_loss=0.06755, over 4890.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2499, pruned_loss=0.05444, over 954828.54 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:48:10,479 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 01:48:26,320 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:26,955 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118330.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:40,932 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:43,713 INFO [finetune.py:976] (0/7) Epoch 21, batch 3800, loss[loss=0.1584, simple_loss=0.2388, pruned_loss=0.03901, over 4799.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2504, pruned_loss=0.05404, over 955217.27 frames. ], batch size: 25, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:48:46,091 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.364e+01 1.561e+02 1.815e+02 2.293e+02 4.441e+02, threshold=3.631e+02, percent-clipped=1.0 +2023-03-27 01:48:49,243 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6427, 1.5459, 1.4992, 1.5152, 1.3593, 3.5398, 1.5286, 1.9763], + device='cuda:0'), covar=tensor([0.3154, 0.2354, 0.2063, 0.2382, 0.1593, 0.0189, 0.2668, 0.1187], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:49:06,611 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:49:11,213 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118396.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:49:17,068 INFO [finetune.py:976] (0/7) Epoch 21, batch 3850, loss[loss=0.1845, simple_loss=0.2506, pruned_loss=0.05919, over 4823.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2491, pruned_loss=0.0531, over 956027.42 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:49:18,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7794, 4.0899, 3.8269, 1.8681, 4.2524, 3.2547, 1.0781, 3.0041], + device='cuda:0'), covar=tensor([0.2302, 0.2160, 0.1545, 0.3392, 0.0948, 0.0851, 0.4242, 0.1303], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0178, 0.0158, 0.0130, 0.0161, 0.0122, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 01:49:24,431 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7071, 1.5746, 1.4980, 1.5762, 1.0060, 3.5418, 1.3918, 1.8817], + device='cuda:0'), covar=tensor([0.3276, 0.2559, 0.2124, 0.2333, 0.1804, 0.0196, 0.2624, 0.1198], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:49:50,287 INFO [finetune.py:976] (0/7) Epoch 21, batch 3900, loss[loss=0.1776, simple_loss=0.2441, pruned_loss=0.05556, over 4815.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2472, pruned_loss=0.05327, over 955959.82 frames. ], batch size: 39, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:49:52,685 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.530e+02 1.819e+02 2.327e+02 4.856e+02, threshold=3.639e+02, percent-clipped=2.0 +2023-03-27 01:50:25,012 INFO [finetune.py:976] (0/7) Epoch 21, batch 3950, loss[loss=0.1528, simple_loss=0.2241, pruned_loss=0.04075, over 4876.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2445, pruned_loss=0.05228, over 957171.72 frames. ], batch size: 34, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:50:43,589 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118524.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:50:52,785 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4565, 1.5383, 1.5214, 0.8253, 1.6121, 1.8553, 1.8396, 1.4296], + device='cuda:0'), covar=tensor([0.0921, 0.0602, 0.0632, 0.0572, 0.0521, 0.0559, 0.0321, 0.0694], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0126, 0.0123, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([9.0104e-05, 1.0829e-04, 9.0367e-05, 8.7164e-05, 9.2108e-05, 9.1682e-05, + 1.0131e-04, 1.0617e-04], device='cuda:0') +2023-03-27 01:51:09,161 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 01:51:19,654 INFO [finetune.py:976] (0/7) Epoch 21, batch 4000, loss[loss=0.1252, simple_loss=0.1859, pruned_loss=0.03228, over 4829.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2438, pruned_loss=0.05248, over 956782.54 frames. ], batch size: 25, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:51:26,592 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.742e+01 1.543e+02 1.809e+02 2.201e+02 4.154e+02, threshold=3.618e+02, percent-clipped=2.0 +2023-03-27 01:51:48,821 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9352, 1.8364, 1.5757, 1.6080, 1.7736, 1.7364, 1.8132, 2.4233], + device='cuda:0'), covar=tensor([0.3884, 0.4221, 0.3176, 0.3788, 0.3948, 0.2537, 0.3692, 0.1872], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0233, 0.0277, 0.0254, 0.0224, 0.0253, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:51:56,640 INFO [finetune.py:976] (0/7) Epoch 21, batch 4050, loss[loss=0.1779, simple_loss=0.2343, pruned_loss=0.0608, over 4657.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2462, pruned_loss=0.05328, over 955763.34 frames. ], batch size: 23, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:52:07,345 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0304, 1.0135, 0.9332, 1.1099, 1.1908, 1.1447, 0.9738, 0.9440], + device='cuda:0'), covar=tensor([0.0367, 0.0279, 0.0647, 0.0300, 0.0277, 0.0443, 0.0342, 0.0424], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0106, 0.0143, 0.0111, 0.0098, 0.0109, 0.0100, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.5458e-05, 8.1033e-05, 1.1226e-04, 8.5005e-05, 7.6499e-05, 8.0939e-05, + 7.4150e-05, 8.4839e-05], device='cuda:0') +2023-03-27 01:52:11,484 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118625.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:15,690 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118631.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:24,640 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118645.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:30,074 INFO [finetune.py:976] (0/7) Epoch 21, batch 4100, loss[loss=0.1543, simple_loss=0.2328, pruned_loss=0.03795, over 4840.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2495, pruned_loss=0.05438, over 955872.77 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:52:33,512 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.828e+01 1.601e+02 1.824e+02 2.338e+02 3.980e+02, threshold=3.647e+02, percent-clipped=1.0 +2023-03-27 01:52:51,553 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:56,336 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:58,685 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118696.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:53:03,520 INFO [finetune.py:976] (0/7) Epoch 21, batch 4150, loss[loss=0.1787, simple_loss=0.2563, pruned_loss=0.0506, over 4776.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2509, pruned_loss=0.0547, over 954389.87 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:53:31,225 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118744.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:53:37,411 INFO [finetune.py:976] (0/7) Epoch 21, batch 4200, loss[loss=0.1569, simple_loss=0.2307, pruned_loss=0.04155, over 4898.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2505, pruned_loss=0.05377, over 955143.94 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:53:39,820 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.932e+01 1.492e+02 1.761e+02 2.066e+02 3.902e+02, threshold=3.521e+02, percent-clipped=1.0 +2023-03-27 01:54:11,363 INFO [finetune.py:976] (0/7) Epoch 21, batch 4250, loss[loss=0.1921, simple_loss=0.2538, pruned_loss=0.06518, over 4900.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2481, pruned_loss=0.05326, over 954393.15 frames. ], batch size: 32, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:54:25,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:54:38,754 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.85 vs. limit=5.0 +2023-03-27 01:54:45,138 INFO [finetune.py:976] (0/7) Epoch 21, batch 4300, loss[loss=0.2107, simple_loss=0.2788, pruned_loss=0.0713, over 4795.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2458, pruned_loss=0.0529, over 951836.20 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:54:47,580 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.563e+02 1.855e+02 2.179e+02 3.656e+02, threshold=3.709e+02, percent-clipped=1.0 +2023-03-27 01:54:57,550 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:18,872 INFO [finetune.py:976] (0/7) Epoch 21, batch 4350, loss[loss=0.1524, simple_loss=0.2221, pruned_loss=0.04133, over 4715.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2421, pruned_loss=0.05114, over 954329.61 frames. ], batch size: 59, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:55:33,242 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:48,468 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:59,298 INFO [finetune.py:976] (0/7) Epoch 21, batch 4400, loss[loss=0.2227, simple_loss=0.2818, pruned_loss=0.08179, over 4903.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2431, pruned_loss=0.05155, over 953702.21 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:56:01,710 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.753e+01 1.466e+02 1.745e+02 2.136e+02 3.634e+02, threshold=3.490e+02, percent-clipped=0.0 +2023-03-27 01:56:17,888 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118973.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:31,293 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:36,932 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:39,988 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8908, 2.5103, 2.4006, 1.3028, 2.6614, 2.0350, 1.6536, 2.3485], + device='cuda:0'), covar=tensor([0.1041, 0.1049, 0.1724, 0.2098, 0.1535, 0.2261, 0.2611, 0.1152], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0191, 0.0199, 0.0182, 0.0210, 0.0209, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:56:40,524 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118993.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:51,290 INFO [finetune.py:976] (0/7) Epoch 21, batch 4450, loss[loss=0.1957, simple_loss=0.2755, pruned_loss=0.05795, over 4907.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2464, pruned_loss=0.05232, over 952415.26 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:57:11,404 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:25,001 INFO [finetune.py:976] (0/7) Epoch 21, batch 4500, loss[loss=0.2152, simple_loss=0.2876, pruned_loss=0.07139, over 4891.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2478, pruned_loss=0.05313, over 951707.90 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:57:27,416 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.548e+02 1.910e+02 2.429e+02 4.520e+02, threshold=3.820e+02, percent-clipped=3.0 +2023-03-27 01:57:38,318 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:51,498 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119093.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:58,451 INFO [finetune.py:976] (0/7) Epoch 21, batch 4550, loss[loss=0.1927, simple_loss=0.2675, pruned_loss=0.05892, over 4935.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2493, pruned_loss=0.0536, over 951472.49 frames. ], batch size: 41, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:58:19,746 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:58:27,418 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5721, 1.4252, 1.4233, 1.5095, 1.0045, 2.9757, 1.1585, 1.5129], + device='cuda:0'), covar=tensor([0.3269, 0.2526, 0.2243, 0.2492, 0.1837, 0.0264, 0.2691, 0.1265], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0113, 0.0097, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:58:31,551 INFO [finetune.py:976] (0/7) Epoch 21, batch 4600, loss[loss=0.1555, simple_loss=0.224, pruned_loss=0.04352, over 4805.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2486, pruned_loss=0.05304, over 952867.43 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:58:31,665 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119154.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:58:34,457 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.653e+02 1.869e+02 2.317e+02 3.451e+02, threshold=3.738e+02, percent-clipped=0.0 +2023-03-27 01:59:05,263 INFO [finetune.py:976] (0/7) Epoch 21, batch 4650, loss[loss=0.2082, simple_loss=0.2624, pruned_loss=0.077, over 4825.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2461, pruned_loss=0.0524, over 953566.81 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:59:38,314 INFO [finetune.py:976] (0/7) Epoch 21, batch 4700, loss[loss=0.1727, simple_loss=0.2309, pruned_loss=0.05725, over 4832.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2436, pruned_loss=0.05188, over 956612.07 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:59:40,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.521e+02 1.791e+02 2.382e+02 6.096e+02, threshold=3.583e+02, percent-clipped=7.0 +2023-03-27 01:59:48,649 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0990, 4.7073, 4.5625, 2.7145, 4.8820, 3.6509, 1.0209, 3.3891], + device='cuda:0'), covar=tensor([0.2282, 0.1879, 0.1429, 0.2850, 0.0707, 0.0864, 0.4254, 0.1351], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0178, 0.0159, 0.0130, 0.0160, 0.0122, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 01:59:48,703 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5828, 2.2995, 2.9839, 1.8167, 2.4647, 2.9161, 2.1084, 3.0158], + device='cuda:0'), covar=tensor([0.1258, 0.1995, 0.1566, 0.2108, 0.1040, 0.1573, 0.2734, 0.0937], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0206, 0.0192, 0.0190, 0.0175, 0.0214, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 01:59:58,286 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6707, 1.5395, 1.5191, 1.6936, 1.0282, 3.4412, 1.3509, 1.6582], + device='cuda:0'), covar=tensor([0.3264, 0.2579, 0.2218, 0.2376, 0.1922, 0.0226, 0.2762, 0.1313], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0114, 0.0097, 0.0095, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 01:59:59,464 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:00:11,568 INFO [finetune.py:976] (0/7) Epoch 21, batch 4750, loss[loss=0.1948, simple_loss=0.2635, pruned_loss=0.06308, over 4741.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2399, pruned_loss=0.05002, over 953889.75 frames. ], batch size: 54, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:00:31,333 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:00:36,048 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1040, 2.1547, 2.1598, 1.5544, 2.0220, 2.2673, 2.2621, 1.7639], + device='cuda:0'), covar=tensor([0.0638, 0.0660, 0.0695, 0.0883, 0.0799, 0.0693, 0.0605, 0.1157], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0120, 0.0126, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:00:37,144 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8635, 1.7591, 1.9180, 1.2438, 1.7958, 1.9731, 1.9700, 1.4828], + device='cuda:0'), covar=tensor([0.0616, 0.0658, 0.0604, 0.0847, 0.0767, 0.0649, 0.0570, 0.1228], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0120, 0.0126, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:00:42,929 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4476, 1.3161, 1.7133, 2.5019, 1.7210, 2.3148, 0.9212, 2.1820], + device='cuda:0'), covar=tensor([0.1618, 0.1451, 0.1125, 0.0770, 0.0901, 0.1034, 0.1516, 0.0567], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0166, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:00:44,656 INFO [finetune.py:976] (0/7) Epoch 21, batch 4800, loss[loss=0.1292, simple_loss=0.2042, pruned_loss=0.02712, over 4722.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2426, pruned_loss=0.05126, over 954884.17 frames. ], batch size: 23, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:00:47,499 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.488e+02 1.781e+02 2.195e+02 3.360e+02, threshold=3.562e+02, percent-clipped=0.0 +2023-03-27 02:01:22,405 INFO [finetune.py:976] (0/7) Epoch 21, batch 4850, loss[loss=0.1662, simple_loss=0.2429, pruned_loss=0.04471, over 4831.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2475, pruned_loss=0.05307, over 955646.57 frames. ], batch size: 47, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:01:47,002 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3231, 2.9398, 2.7862, 1.3165, 3.0379, 2.2200, 0.6513, 1.8430], + device='cuda:0'), covar=tensor([0.2442, 0.2465, 0.1796, 0.3284, 0.1403, 0.1186, 0.4162, 0.1777], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0178, 0.0160, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 02:01:55,527 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119431.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:02:13,628 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 02:02:15,267 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119449.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 02:02:19,182 INFO [finetune.py:976] (0/7) Epoch 21, batch 4900, loss[loss=0.2165, simple_loss=0.2667, pruned_loss=0.08311, over 4165.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.249, pruned_loss=0.05367, over 953599.57 frames. ], batch size: 18, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:02:25,215 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.684e+02 1.937e+02 2.365e+02 4.201e+02, threshold=3.874e+02, percent-clipped=2.0 +2023-03-27 02:02:39,231 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 02:02:47,615 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8276, 1.3327, 1.9603, 1.8693, 1.6773, 1.6148, 1.8399, 1.7746], + device='cuda:0'), covar=tensor([0.4057, 0.3884, 0.3064, 0.3690, 0.4364, 0.3513, 0.4051, 0.2923], + device='cuda:0'), in_proj_covar=tensor([0.0256, 0.0243, 0.0265, 0.0284, 0.0282, 0.0258, 0.0292, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:02:55,580 INFO [finetune.py:976] (0/7) Epoch 21, batch 4950, loss[loss=0.1716, simple_loss=0.2454, pruned_loss=0.0489, over 4892.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2495, pruned_loss=0.05308, over 953678.17 frames. ], batch size: 35, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:02:56,399 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-03-27 02:03:02,769 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119514.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:03:14,838 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0947, 1.6477, 2.3643, 3.8908, 2.7036, 2.7161, 1.0647, 3.2191], + device='cuda:0'), covar=tensor([0.1639, 0.1522, 0.1424, 0.0618, 0.0779, 0.1621, 0.1820, 0.0445], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0102, 0.0140, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:03:29,003 INFO [finetune.py:976] (0/7) Epoch 21, batch 5000, loss[loss=0.1487, simple_loss=0.213, pruned_loss=0.0422, over 4891.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2471, pruned_loss=0.05224, over 953149.96 frames. ], batch size: 32, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:03:32,979 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.554e+02 1.853e+02 2.138e+02 3.358e+02, threshold=3.705e+02, percent-clipped=0.0 +2023-03-27 02:03:43,293 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:03:49,834 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7661, 1.2924, 0.9159, 1.7432, 2.2253, 1.4690, 1.4840, 1.6480], + device='cuda:0'), covar=tensor([0.1347, 0.1995, 0.1904, 0.1114, 0.1703, 0.1826, 0.1484, 0.1941], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 02:04:00,492 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4576, 1.3681, 1.5605, 0.8387, 1.5198, 1.5407, 1.5208, 1.2862], + device='cuda:0'), covar=tensor([0.0634, 0.0860, 0.0676, 0.0954, 0.0875, 0.0712, 0.0662, 0.1346], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0136, 0.0139, 0.0120, 0.0126, 0.0138, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:04:01,117 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0777, 1.8092, 2.3271, 1.5971, 2.1056, 2.2257, 1.6924, 2.4016], + device='cuda:0'), covar=tensor([0.1113, 0.1748, 0.1423, 0.1837, 0.0851, 0.1387, 0.2522, 0.0739], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0204, 0.0191, 0.0189, 0.0173, 0.0213, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:04:02,226 INFO [finetune.py:976] (0/7) Epoch 21, batch 5050, loss[loss=0.1826, simple_loss=0.2504, pruned_loss=0.05745, over 4887.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.245, pruned_loss=0.05181, over 951596.19 frames. ], batch size: 35, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:04:09,249 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-27 02:04:34,251 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-27 02:04:35,255 INFO [finetune.py:976] (0/7) Epoch 21, batch 5100, loss[loss=0.1743, simple_loss=0.2363, pruned_loss=0.05612, over 4764.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2411, pruned_loss=0.05026, over 952955.68 frames. ], batch size: 27, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:04:39,201 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.479e+02 1.750e+02 2.173e+02 3.976e+02, threshold=3.500e+02, percent-clipped=1.0 +2023-03-27 02:04:43,483 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119665.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:08,855 INFO [finetune.py:976] (0/7) Epoch 21, batch 5150, loss[loss=0.1728, simple_loss=0.2454, pruned_loss=0.05015, over 4764.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2405, pruned_loss=0.05027, over 950960.49 frames. ], batch size: 28, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:05:15,891 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3407, 1.2632, 1.2679, 1.2545, 0.7847, 2.2113, 0.7815, 1.1673], + device='cuda:0'), covar=tensor([0.3320, 0.2463, 0.2271, 0.2534, 0.2104, 0.0366, 0.2973, 0.1370], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:05:24,649 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:27,591 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119731.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:30,090 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6713, 1.4733, 1.0262, 0.2628, 1.1868, 1.4954, 1.4612, 1.4084], + device='cuda:0'), covar=tensor([0.1028, 0.0909, 0.1543, 0.2030, 0.1577, 0.2506, 0.2246, 0.0995], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0191, 0.0198, 0.0182, 0.0209, 0.0208, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:05:31,846 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3813, 2.8992, 2.7489, 1.2430, 3.0574, 2.2782, 0.6585, 1.8815], + device='cuda:0'), covar=tensor([0.2417, 0.2849, 0.2020, 0.3764, 0.1616, 0.1111, 0.4410, 0.1845], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0177, 0.0158, 0.0129, 0.0160, 0.0121, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 02:05:38,885 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119749.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:05:42,228 INFO [finetune.py:976] (0/7) Epoch 21, batch 5200, loss[loss=0.1893, simple_loss=0.2675, pruned_loss=0.05556, over 4814.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2442, pruned_loss=0.05111, over 953086.20 frames. ], batch size: 40, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:05:45,722 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.656e+02 1.877e+02 2.270e+02 4.720e+02, threshold=3.754e+02, percent-clipped=1.0 +2023-03-27 02:05:59,329 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:06:10,789 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119797.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:06:15,130 INFO [finetune.py:976] (0/7) Epoch 21, batch 5250, loss[loss=0.1876, simple_loss=0.2634, pruned_loss=0.05591, over 4919.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2458, pruned_loss=0.05141, over 950089.21 frames. ], batch size: 36, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:06:59,457 INFO [finetune.py:976] (0/7) Epoch 21, batch 5300, loss[loss=0.2278, simple_loss=0.3021, pruned_loss=0.07674, over 4891.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2498, pruned_loss=0.05316, over 951086.60 frames. ], batch size: 36, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:07:07,164 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.978e+01 1.512e+02 1.747e+02 2.069e+02 4.039e+02, threshold=3.495e+02, percent-clipped=1.0 +2023-03-27 02:07:18,527 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:07:54,828 INFO [finetune.py:976] (0/7) Epoch 21, batch 5350, loss[loss=0.1809, simple_loss=0.248, pruned_loss=0.05692, over 4792.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2499, pruned_loss=0.05322, over 950906.80 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:08:28,086 INFO [finetune.py:976] (0/7) Epoch 21, batch 5400, loss[loss=0.1581, simple_loss=0.233, pruned_loss=0.04161, over 4872.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2463, pruned_loss=0.0517, over 951520.30 frames. ], batch size: 34, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:08:31,173 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.474e+02 1.662e+02 2.015e+02 3.492e+02, threshold=3.324e+02, percent-clipped=0.0 +2023-03-27 02:08:58,905 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-120000.pt +2023-03-27 02:09:02,904 INFO [finetune.py:976] (0/7) Epoch 21, batch 5450, loss[loss=0.1753, simple_loss=0.2473, pruned_loss=0.0516, over 4859.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2428, pruned_loss=0.05096, over 950627.01 frames. ], batch size: 49, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:09:13,779 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=120021.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:09:36,195 INFO [finetune.py:976] (0/7) Epoch 21, batch 5500, loss[loss=0.2338, simple_loss=0.3011, pruned_loss=0.08323, over 4349.00 frames. ], tot_loss[loss=0.17, simple_loss=0.24, pruned_loss=0.04996, over 948592.92 frames. ], batch size: 65, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:09:37,497 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5733, 2.7099, 2.4547, 1.8053, 2.4876, 2.7870, 2.9012, 2.2976], + device='cuda:0'), covar=tensor([0.0571, 0.0562, 0.0692, 0.0869, 0.0788, 0.0639, 0.0557, 0.1004], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0137, 0.0140, 0.0121, 0.0126, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:09:39,679 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.953e+01 1.422e+02 1.683e+02 2.099e+02 3.794e+02, threshold=3.366e+02, percent-clipped=2.0 +2023-03-27 02:09:48,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5975, 2.5006, 1.9956, 2.7456, 2.5402, 2.1639, 3.0368, 2.5365], + device='cuda:0'), covar=tensor([0.1214, 0.2078, 0.2887, 0.2360, 0.2384, 0.1562, 0.2520, 0.1662], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0187, 0.0233, 0.0251, 0.0245, 0.0202, 0.0213, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:09:49,976 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7922, 3.7529, 3.4998, 1.8757, 3.8623, 2.9467, 0.8136, 2.6293], + device='cuda:0'), covar=tensor([0.2400, 0.2570, 0.1724, 0.3566, 0.1060, 0.0991, 0.4707, 0.1605], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0178, 0.0159, 0.0130, 0.0161, 0.0123, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 02:10:09,946 INFO [finetune.py:976] (0/7) Epoch 21, batch 5550, loss[loss=0.1674, simple_loss=0.2507, pruned_loss=0.04203, over 4863.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2425, pruned_loss=0.05097, over 947365.74 frames. ], batch size: 44, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:10:19,069 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2299, 1.3149, 1.4970, 1.0741, 1.2766, 1.4317, 1.3154, 1.6100], + device='cuda:0'), covar=tensor([0.1241, 0.2139, 0.1233, 0.1448, 0.0976, 0.1311, 0.2936, 0.0882], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0205, 0.0192, 0.0190, 0.0175, 0.0215, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:10:34,171 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3177, 1.2673, 1.1312, 1.3544, 1.5471, 1.4274, 1.2538, 1.1245], + device='cuda:0'), covar=tensor([0.0347, 0.0338, 0.0712, 0.0320, 0.0251, 0.0539, 0.0401, 0.0492], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0108, 0.0145, 0.0112, 0.0100, 0.0112, 0.0101, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.6695e-05, 8.2567e-05, 1.1414e-04, 8.6330e-05, 7.7973e-05, 8.3108e-05, + 7.5475e-05, 8.6595e-05], device='cuda:0') +2023-03-27 02:10:42,282 INFO [finetune.py:976] (0/7) Epoch 21, batch 5600, loss[loss=0.1677, simple_loss=0.2499, pruned_loss=0.04273, over 4926.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2467, pruned_loss=0.05235, over 947976.38 frames. ], batch size: 33, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:10:45,197 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.550e+02 1.831e+02 2.203e+02 3.727e+02, threshold=3.662e+02, percent-clipped=1.0 +2023-03-27 02:10:52,171 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120170.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:11:09,874 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0020, 1.7969, 2.2181, 1.4653, 2.0803, 2.3126, 1.6828, 2.4130], + device='cuda:0'), covar=tensor([0.1441, 0.2166, 0.1631, 0.2122, 0.1032, 0.1451, 0.3023, 0.0968], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0206, 0.0192, 0.0191, 0.0175, 0.0216, 0.0218, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:11:12,065 INFO [finetune.py:976] (0/7) Epoch 21, batch 5650, loss[loss=0.2006, simple_loss=0.2884, pruned_loss=0.05643, over 4931.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2505, pruned_loss=0.05356, over 947483.63 frames. ], batch size: 38, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:11:20,661 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:11:41,964 INFO [finetune.py:976] (0/7) Epoch 21, batch 5700, loss[loss=0.1657, simple_loss=0.2201, pruned_loss=0.05561, over 4183.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2459, pruned_loss=0.05247, over 929473.21 frames. ], batch size: 18, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:11:44,943 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.458e+02 1.705e+02 2.128e+02 3.595e+02, threshold=3.409e+02, percent-clipped=0.0 +2023-03-27 02:11:50,449 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0617, 0.9998, 0.9398, 1.1165, 1.0365, 1.0976, 0.6378, 0.9659], + device='cuda:0'), covar=tensor([0.1146, 0.1009, 0.1006, 0.0869, 0.0983, 0.0654, 0.1489, 0.0973], + device='cuda:0'), in_proj_covar=tensor([0.0243, 0.0209, 0.0212, 0.0194, 0.0243, 0.0188, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:11:58,956 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-21.pt +2023-03-27 02:12:12,191 INFO [finetune.py:976] (0/7) Epoch 22, batch 0, loss[loss=0.1855, simple_loss=0.2584, pruned_loss=0.0563, over 4838.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2584, pruned_loss=0.0563, over 4838.00 frames. ], batch size: 47, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:12:12,193 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 02:12:27,778 INFO [finetune.py:1010] (0/7) Epoch 22, validation: loss=0.1597, simple_loss=0.228, pruned_loss=0.04574, over 2265189.00 frames. +2023-03-27 02:12:27,779 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 02:12:35,250 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2115, 1.8675, 2.1977, 1.5170, 2.1065, 2.3725, 2.4103, 1.3934], + device='cuda:0'), covar=tensor([0.0760, 0.0971, 0.0789, 0.1031, 0.0770, 0.0743, 0.0671, 0.1915], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0137, 0.0139, 0.0121, 0.0125, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:13:15,368 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:13:27,846 INFO [finetune.py:976] (0/7) Epoch 22, batch 50, loss[loss=0.1748, simple_loss=0.2418, pruned_loss=0.05391, over 4922.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2517, pruned_loss=0.05562, over 215914.27 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:13:48,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.631e+02 1.972e+02 2.363e+02 4.295e+02, threshold=3.943e+02, percent-clipped=3.0 +2023-03-27 02:13:55,425 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120369.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:14:04,333 INFO [finetune.py:976] (0/7) Epoch 22, batch 100, loss[loss=0.1922, simple_loss=0.2558, pruned_loss=0.06432, over 4918.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2447, pruned_loss=0.05247, over 380376.37 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:14:11,969 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 02:14:37,019 INFO [finetune.py:976] (0/7) Epoch 22, batch 150, loss[loss=0.1807, simple_loss=0.2266, pruned_loss=0.06736, over 4033.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2405, pruned_loss=0.05197, over 504493.26 frames. ], batch size: 17, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:14:48,867 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 02:14:55,336 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.445e+02 1.737e+02 2.098e+02 4.550e+02, threshold=3.473e+02, percent-clipped=2.0 +2023-03-27 02:15:03,339 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 02:15:10,227 INFO [finetune.py:976] (0/7) Epoch 22, batch 200, loss[loss=0.1746, simple_loss=0.2464, pruned_loss=0.05143, over 4909.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2388, pruned_loss=0.05182, over 605964.77 frames. ], batch size: 37, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:15:14,957 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2092, 3.5761, 3.7941, 4.0869, 3.9643, 3.7159, 4.2939, 1.4647], + device='cuda:0'), covar=tensor([0.0863, 0.0990, 0.0868, 0.1003, 0.1248, 0.1710, 0.0715, 0.5682], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0243, 0.0281, 0.0292, 0.0334, 0.0284, 0.0305, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:15:20,009 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0590, 2.1193, 1.8135, 2.1469, 1.9375, 4.8197, 1.9433, 2.5269], + device='cuda:0'), covar=tensor([0.2984, 0.2373, 0.2026, 0.2194, 0.1348, 0.0169, 0.2225, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:15:42,763 INFO [finetune.py:976] (0/7) Epoch 22, batch 250, loss[loss=0.2038, simple_loss=0.2771, pruned_loss=0.06526, over 4815.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2431, pruned_loss=0.05264, over 685333.73 frames. ], batch size: 38, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:16:01,910 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.519e+02 1.843e+02 2.180e+02 3.548e+02, threshold=3.686e+02, percent-clipped=1.0 +2023-03-27 02:16:06,894 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3081, 3.2409, 2.8266, 1.6127, 2.9690, 2.3941, 2.3925, 2.8107], + device='cuda:0'), covar=tensor([0.0829, 0.0760, 0.1691, 0.2135, 0.1415, 0.2213, 0.1911, 0.1072], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0190, 0.0197, 0.0182, 0.0208, 0.0207, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:16:16,407 INFO [finetune.py:976] (0/7) Epoch 22, batch 300, loss[loss=0.1586, simple_loss=0.2319, pruned_loss=0.04264, over 4923.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2468, pruned_loss=0.05339, over 745538.83 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:16:50,473 INFO [finetune.py:976] (0/7) Epoch 22, batch 350, loss[loss=0.1283, simple_loss=0.2042, pruned_loss=0.02619, over 4771.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2489, pruned_loss=0.05412, over 792218.89 frames. ], batch size: 27, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:17:09,291 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.647e+02 1.881e+02 2.280e+02 4.594e+02, threshold=3.762e+02, percent-clipped=3.0 +2023-03-27 02:17:19,788 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=120676.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:17:23,372 INFO [finetune.py:976] (0/7) Epoch 22, batch 400, loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.05227, over 4773.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.25, pruned_loss=0.05421, over 828449.74 frames. ], batch size: 51, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:18:12,318 INFO [finetune.py:976] (0/7) Epoch 22, batch 450, loss[loss=0.1641, simple_loss=0.2296, pruned_loss=0.0493, over 4876.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2484, pruned_loss=0.05308, over 857669.44 frames. ], batch size: 31, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:18:20,734 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=120737.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:18:43,318 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.581e+02 1.866e+02 2.243e+02 3.725e+02, threshold=3.731e+02, percent-clipped=0.0 +2023-03-27 02:19:07,338 INFO [finetune.py:976] (0/7) Epoch 22, batch 500, loss[loss=0.1562, simple_loss=0.2126, pruned_loss=0.04988, over 4821.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2469, pruned_loss=0.05301, over 879030.66 frames. ], batch size: 41, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:19:40,280 INFO [finetune.py:976] (0/7) Epoch 22, batch 550, loss[loss=0.1899, simple_loss=0.2585, pruned_loss=0.06062, over 4793.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2442, pruned_loss=0.05218, over 895843.66 frames. ], batch size: 29, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:19:42,235 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7658, 1.6636, 1.4963, 1.8603, 2.0563, 1.8636, 1.4163, 1.4830], + device='cuda:0'), covar=tensor([0.2222, 0.2046, 0.2075, 0.1593, 0.1582, 0.1258, 0.2597, 0.2011], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0215, 0.0196, 0.0244, 0.0190, 0.0219, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:19:45,426 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-27 02:19:49,439 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0041, 1.3573, 2.0778, 1.9973, 1.8442, 1.8001, 1.9221, 1.9694], + device='cuda:0'), covar=tensor([0.3527, 0.3779, 0.3237, 0.3529, 0.4839, 0.3549, 0.4159, 0.2908], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0241, 0.0262, 0.0282, 0.0281, 0.0257, 0.0289, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:19:58,120 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.544e+02 1.778e+02 2.172e+02 3.720e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-03-27 02:20:04,754 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0357, 1.8574, 1.6226, 1.6605, 1.7731, 1.8036, 1.8275, 2.4801], + device='cuda:0'), covar=tensor([0.3660, 0.3999, 0.3267, 0.3772, 0.3975, 0.2336, 0.3574, 0.1786], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0233, 0.0278, 0.0255, 0.0224, 0.0254, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:20:13,124 INFO [finetune.py:976] (0/7) Epoch 22, batch 600, loss[loss=0.1965, simple_loss=0.2684, pruned_loss=0.06227, over 4907.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2438, pruned_loss=0.05217, over 907622.70 frames. ], batch size: 46, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:20:46,533 INFO [finetune.py:976] (0/7) Epoch 22, batch 650, loss[loss=0.2238, simple_loss=0.2872, pruned_loss=0.08026, over 4893.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2472, pruned_loss=0.05301, over 918775.40 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:03,251 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-03-27 02:21:04,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.580e+02 1.877e+02 2.237e+02 3.344e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-27 02:21:20,032 INFO [finetune.py:976] (0/7) Epoch 22, batch 700, loss[loss=0.1713, simple_loss=0.2519, pruned_loss=0.04534, over 4825.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2492, pruned_loss=0.0542, over 924594.26 frames. ], batch size: 39, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:32,894 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121003.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:21:53,225 INFO [finetune.py:976] (0/7) Epoch 22, batch 750, loss[loss=0.1771, simple_loss=0.258, pruned_loss=0.04815, over 4780.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2513, pruned_loss=0.05529, over 930971.14 frames. ], batch size: 25, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:53,316 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:22:09,806 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.617e+02 1.906e+02 2.410e+02 4.829e+02, threshold=3.812e+02, percent-clipped=5.0 +2023-03-27 02:22:14,357 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121064.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:22:15,444 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0777, 0.9633, 0.9968, 1.1319, 1.2040, 1.1789, 0.9929, 0.9803], + device='cuda:0'), covar=tensor([0.0378, 0.0317, 0.0648, 0.0303, 0.0284, 0.0440, 0.0370, 0.0420], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0107, 0.0143, 0.0111, 0.0099, 0.0111, 0.0100, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.6103e-05, 8.1905e-05, 1.1247e-04, 8.5177e-05, 7.6899e-05, 8.1930e-05, + 7.4616e-05, 8.5577e-05], device='cuda:0') +2023-03-27 02:22:26,768 INFO [finetune.py:976] (0/7) Epoch 22, batch 800, loss[loss=0.1257, simple_loss=0.1981, pruned_loss=0.02664, over 4812.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2497, pruned_loss=0.05393, over 936028.35 frames. ], batch size: 25, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:22:33,098 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0446, 0.9931, 1.0089, 0.4844, 0.9336, 1.1586, 1.1167, 1.0075], + device='cuda:0'), covar=tensor([0.0858, 0.0565, 0.0544, 0.0546, 0.0559, 0.0606, 0.0422, 0.0715], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0123, 0.0130, 0.0128, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.9036e-05, 1.0794e-04, 8.9631e-05, 8.6375e-05, 9.1455e-05, 9.1485e-05, + 1.0071e-04, 1.0558e-04], device='cuda:0') +2023-03-27 02:22:47,234 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.5965, 4.0155, 4.1877, 4.4787, 4.3907, 4.0065, 4.6953, 1.4070], + device='cuda:0'), covar=tensor([0.0741, 0.0792, 0.0790, 0.0841, 0.1049, 0.1663, 0.0589, 0.5949], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0242, 0.0279, 0.0290, 0.0332, 0.0283, 0.0303, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:23:10,262 INFO [finetune.py:976] (0/7) Epoch 22, batch 850, loss[loss=0.1675, simple_loss=0.2212, pruned_loss=0.05689, over 4007.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2471, pruned_loss=0.05312, over 940111.73 frames. ], batch size: 17, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:23:27,589 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5681, 1.7490, 2.1140, 1.8492, 1.8295, 3.1556, 1.6355, 1.7618], + device='cuda:0'), covar=tensor([0.0848, 0.1439, 0.1110, 0.0792, 0.1160, 0.0274, 0.1175, 0.1422], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:23:28,703 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.269e+01 1.509e+02 1.830e+02 2.168e+02 4.982e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-03-27 02:23:37,192 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:23:58,239 INFO [finetune.py:976] (0/7) Epoch 22, batch 900, loss[loss=0.2019, simple_loss=0.2643, pruned_loss=0.06977, over 4909.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2459, pruned_loss=0.05287, over 943863.35 frames. ], batch size: 46, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:24:09,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9104, 1.8207, 1.7582, 2.0661, 2.3573, 2.1225, 1.9155, 1.6784], + device='cuda:0'), covar=tensor([0.2037, 0.1903, 0.1766, 0.1479, 0.1722, 0.1198, 0.2243, 0.1702], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0209, 0.0213, 0.0195, 0.0242, 0.0189, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:24:38,332 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:24:38,357 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8442, 1.3994, 1.9909, 1.8506, 1.6791, 1.6151, 1.8530, 1.8815], + device='cuda:0'), covar=tensor([0.3137, 0.3017, 0.2415, 0.2792, 0.3758, 0.3076, 0.3151, 0.2208], + device='cuda:0'), in_proj_covar=tensor([0.0255, 0.0243, 0.0263, 0.0284, 0.0282, 0.0258, 0.0291, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:24:42,944 INFO [finetune.py:976] (0/7) Epoch 22, batch 950, loss[loss=0.1752, simple_loss=0.2491, pruned_loss=0.05062, over 4768.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2447, pruned_loss=0.05257, over 946299.64 frames. ], batch size: 28, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:24:44,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3420, 1.2842, 1.2700, 0.7583, 1.2843, 1.4631, 1.5197, 1.2084], + device='cuda:0'), covar=tensor([0.0739, 0.0489, 0.0514, 0.0435, 0.0491, 0.0404, 0.0282, 0.0515], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0122, 0.0130, 0.0129, 0.0141, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.9046e-05, 1.0788e-04, 8.9450e-05, 8.6305e-05, 9.1685e-05, 9.1739e-05, + 1.0081e-04, 1.0533e-04], device='cuda:0') +2023-03-27 02:24:59,302 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.960e+01 1.496e+02 1.804e+02 2.225e+02 4.174e+02, threshold=3.608e+02, percent-clipped=1.0 +2023-03-27 02:25:02,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-27 02:25:16,227 INFO [finetune.py:976] (0/7) Epoch 22, batch 1000, loss[loss=0.1746, simple_loss=0.2396, pruned_loss=0.05481, over 4227.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2467, pruned_loss=0.05334, over 947587.91 frames. ], batch size: 18, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:25:27,111 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6559, 1.2370, 0.9085, 1.6591, 2.0339, 1.5023, 1.3809, 1.6175], + device='cuda:0'), covar=tensor([0.1453, 0.2058, 0.1938, 0.1125, 0.1995, 0.1885, 0.1469, 0.1856], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0120, 0.0093, 0.0099, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 02:25:33,730 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121311.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:25:34,933 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9086, 1.6251, 2.3047, 3.5635, 2.4480, 2.5210, 1.3139, 3.0197], + device='cuda:0'), covar=tensor([0.1607, 0.1412, 0.1265, 0.0515, 0.0755, 0.1422, 0.1663, 0.0412], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0136, 0.0167, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:25:45,964 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0713, 1.9327, 1.4719, 0.6066, 1.5405, 1.7062, 1.5982, 1.7750], + device='cuda:0'), covar=tensor([0.0818, 0.0674, 0.1322, 0.1879, 0.1380, 0.2177, 0.2188, 0.0869], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0191, 0.0198, 0.0181, 0.0209, 0.0207, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:25:49,276 INFO [finetune.py:976] (0/7) Epoch 22, batch 1050, loss[loss=0.1696, simple_loss=0.2494, pruned_loss=0.04489, over 4823.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.25, pruned_loss=0.05446, over 949235.43 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:25:49,373 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:25:49,413 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2049, 2.1555, 1.8235, 2.3016, 2.7964, 2.1988, 2.0630, 1.6241], + device='cuda:0'), covar=tensor([0.2131, 0.1893, 0.1853, 0.1515, 0.1564, 0.1177, 0.2060, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0196, 0.0244, 0.0190, 0.0218, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:26:05,530 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.613e+02 2.056e+02 2.633e+02 6.948e+02, threshold=4.113e+02, percent-clipped=5.0 +2023-03-27 02:26:05,610 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121359.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:13,525 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121372.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:26:19,238 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:20,880 INFO [finetune.py:976] (0/7) Epoch 22, batch 1100, loss[loss=0.2005, simple_loss=0.2814, pruned_loss=0.05981, over 4808.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2489, pruned_loss=0.05372, over 950577.08 frames. ], batch size: 39, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:26:24,526 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.02 vs. limit=5.0 +2023-03-27 02:26:36,775 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121407.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:41,112 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-03-27 02:26:53,110 INFO [finetune.py:976] (0/7) Epoch 22, batch 1150, loss[loss=0.2106, simple_loss=0.2675, pruned_loss=0.07681, over 4812.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2499, pruned_loss=0.05354, over 951111.00 frames. ], batch size: 39, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:27:09,453 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 02:27:10,453 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.055e+01 1.683e+02 1.964e+02 2.424e+02 3.625e+02, threshold=3.928e+02, percent-clipped=0.0 +2023-03-27 02:27:15,985 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:27:25,671 INFO [finetune.py:976] (0/7) Epoch 22, batch 1200, loss[loss=0.1572, simple_loss=0.2347, pruned_loss=0.03981, over 4804.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2479, pruned_loss=0.05258, over 948847.94 frames. ], batch size: 25, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:27:49,694 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121520.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:27:50,948 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6823, 1.5747, 1.9363, 1.3414, 1.7089, 1.9495, 1.4889, 2.0946], + device='cuda:0'), covar=tensor([0.1111, 0.1919, 0.1184, 0.1605, 0.0895, 0.1194, 0.2846, 0.0819], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0206, 0.0192, 0.0190, 0.0175, 0.0214, 0.0217, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:27:57,407 INFO [finetune.py:976] (0/7) Epoch 22, batch 1250, loss[loss=0.1972, simple_loss=0.2647, pruned_loss=0.06487, over 4852.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2453, pruned_loss=0.05195, over 950187.84 frames. ], batch size: 44, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:28:26,677 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.487e+02 1.790e+02 2.203e+02 4.512e+02, threshold=3.581e+02, percent-clipped=1.0 +2023-03-27 02:28:30,941 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7917, 3.7279, 3.6474, 1.7011, 3.9594, 3.0400, 0.9497, 2.6935], + device='cuda:0'), covar=tensor([0.2496, 0.2136, 0.1418, 0.3202, 0.0933, 0.0935, 0.4050, 0.1455], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0130, 0.0160, 0.0123, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 02:28:35,971 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 02:28:40,600 INFO [finetune.py:976] (0/7) Epoch 22, batch 1300, loss[loss=0.1437, simple_loss=0.2092, pruned_loss=0.03908, over 4789.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2439, pruned_loss=0.0519, over 953080.65 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:28:57,902 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7672, 1.7374, 1.5995, 1.9341, 2.3534, 2.0406, 1.7861, 1.4492], + device='cuda:0'), covar=tensor([0.2501, 0.2123, 0.2046, 0.1755, 0.1775, 0.1282, 0.2346, 0.2183], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0196, 0.0244, 0.0190, 0.0218, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:29:34,826 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5372, 1.7682, 1.8392, 1.0903, 1.8335, 2.0822, 2.0784, 1.5981], + device='cuda:0'), covar=tensor([0.0967, 0.0750, 0.0565, 0.0564, 0.0508, 0.0771, 0.0367, 0.0753], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0151, 0.0128, 0.0124, 0.0133, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([9.0634e-05, 1.0943e-04, 9.1367e-05, 8.7745e-05, 9.3894e-05, 9.3403e-05, + 1.0260e-04, 1.0736e-04], device='cuda:0') +2023-03-27 02:29:39,000 INFO [finetune.py:976] (0/7) Epoch 22, batch 1350, loss[loss=0.1935, simple_loss=0.2681, pruned_loss=0.05941, over 4911.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2446, pruned_loss=0.05238, over 954138.07 frames. ], batch size: 37, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:30:02,099 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.546e+02 1.913e+02 2.289e+02 6.231e+02, threshold=3.826e+02, percent-clipped=1.0 +2023-03-27 02:30:02,194 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:30:07,019 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121667.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:30:16,103 INFO [finetune.py:976] (0/7) Epoch 22, batch 1400, loss[loss=0.1979, simple_loss=0.2709, pruned_loss=0.06242, over 4904.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2467, pruned_loss=0.05307, over 955657.33 frames. ], batch size: 36, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:30:34,333 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121707.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:30:49,371 INFO [finetune.py:976] (0/7) Epoch 22, batch 1450, loss[loss=0.1613, simple_loss=0.2426, pruned_loss=0.04002, over 4768.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2487, pruned_loss=0.05347, over 955352.27 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:31:08,590 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.585e+02 1.838e+02 2.176e+02 4.319e+02, threshold=3.676e+02, percent-clipped=1.0 +2023-03-27 02:31:09,542 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 02:31:11,093 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121763.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:31:22,461 INFO [finetune.py:976] (0/7) Epoch 22, batch 1500, loss[loss=0.1937, simple_loss=0.2612, pruned_loss=0.06312, over 4857.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2508, pruned_loss=0.05423, over 955177.47 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:31:49,125 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:31:56,350 INFO [finetune.py:976] (0/7) Epoch 22, batch 1550, loss[loss=0.1606, simple_loss=0.2423, pruned_loss=0.03939, over 4855.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2509, pruned_loss=0.05398, over 953735.77 frames. ], batch size: 34, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:32:00,097 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1721, 3.3607, 3.0506, 2.3886, 3.0045, 3.3268, 3.5053, 2.9400], + device='cuda:0'), covar=tensor([0.0525, 0.0410, 0.0577, 0.0644, 0.0657, 0.0482, 0.0458, 0.0808], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0137, 0.0140, 0.0121, 0.0127, 0.0139, 0.0141, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:32:15,624 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.581e+02 1.856e+02 2.152e+02 3.350e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-27 02:32:21,171 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121868.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:32:29,536 INFO [finetune.py:976] (0/7) Epoch 22, batch 1600, loss[loss=0.1561, simple_loss=0.2133, pruned_loss=0.04942, over 4383.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2481, pruned_loss=0.05299, over 953584.18 frames. ], batch size: 19, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:32:45,489 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7645, 1.0392, 1.8625, 1.7805, 1.6331, 1.6148, 1.7021, 1.8039], + device='cuda:0'), covar=tensor([0.4075, 0.4237, 0.3493, 0.3727, 0.4837, 0.3937, 0.4578, 0.3095], + device='cuda:0'), in_proj_covar=tensor([0.0256, 0.0242, 0.0262, 0.0283, 0.0282, 0.0258, 0.0291, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:32:51,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7818, 1.7123, 1.5443, 1.8265, 2.4745, 1.8955, 1.7116, 1.4635], + device='cuda:0'), covar=tensor([0.2494, 0.2154, 0.2284, 0.1739, 0.1489, 0.1358, 0.2417, 0.2229], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0214, 0.0195, 0.0242, 0.0189, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:33:02,671 INFO [finetune.py:976] (0/7) Epoch 22, batch 1650, loss[loss=0.1747, simple_loss=0.2366, pruned_loss=0.05639, over 4830.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2453, pruned_loss=0.05192, over 954867.32 frames. ], batch size: 25, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:33:08,882 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4758, 1.6164, 1.5834, 0.8859, 1.7612, 1.9479, 1.8740, 1.4442], + device='cuda:0'), covar=tensor([0.0884, 0.0643, 0.0571, 0.0537, 0.0486, 0.0641, 0.0336, 0.0647], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0152, 0.0128, 0.0124, 0.0133, 0.0132, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([9.1020e-05, 1.0963e-04, 9.1508e-05, 8.7699e-05, 9.3708e-05, 9.3936e-05, + 1.0272e-04, 1.0742e-04], device='cuda:0') +2023-03-27 02:33:13,736 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121950.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:33:15,595 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 02:33:22,466 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.553e+02 1.837e+02 2.107e+02 3.976e+02, threshold=3.675e+02, percent-clipped=1.0 +2023-03-27 02:33:33,818 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121967.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:33:46,459 INFO [finetune.py:976] (0/7) Epoch 22, batch 1700, loss[loss=0.1859, simple_loss=0.2513, pruned_loss=0.06029, over 4927.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2429, pruned_loss=0.0512, over 955307.71 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:33:57,550 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-122000.pt +2023-03-27 02:34:13,902 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122011.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:34:16,838 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122015.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:34:36,870 INFO [finetune.py:976] (0/7) Epoch 22, batch 1750, loss[loss=0.1533, simple_loss=0.2236, pruned_loss=0.04144, over 4759.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2445, pruned_loss=0.05149, over 953157.82 frames. ], batch size: 27, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:34:54,082 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.6505, 4.9086, 5.1842, 5.4473, 5.3742, 5.1975, 5.7181, 2.1479], + device='cuda:0'), covar=tensor([0.0630, 0.0814, 0.0803, 0.0777, 0.1055, 0.1231, 0.0540, 0.5494], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0244, 0.0281, 0.0290, 0.0333, 0.0283, 0.0305, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:34:56,622 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-03-27 02:35:06,504 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.629e+02 1.889e+02 2.189e+02 5.095e+02, threshold=3.778e+02, percent-clipped=2.0 +2023-03-27 02:35:10,492 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122063.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:22,845 INFO [finetune.py:976] (0/7) Epoch 22, batch 1800, loss[loss=0.1889, simple_loss=0.26, pruned_loss=0.05889, over 4903.00 frames. ], tot_loss[loss=0.176, simple_loss=0.247, pruned_loss=0.05252, over 951521.16 frames. ], batch size: 36, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:35:41,154 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122111.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:46,352 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:56,273 INFO [finetune.py:976] (0/7) Epoch 22, batch 1850, loss[loss=0.1625, simple_loss=0.2464, pruned_loss=0.03935, over 4815.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.248, pruned_loss=0.05217, over 953505.90 frames. ], batch size: 40, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:36:05,700 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.85 vs. limit=5.0 +2023-03-27 02:36:08,671 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7620, 1.6161, 1.4061, 1.8284, 2.2212, 1.8877, 1.5303, 1.4372], + device='cuda:0'), covar=tensor([0.2071, 0.2040, 0.1992, 0.1530, 0.1665, 0.1187, 0.2409, 0.1898], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0197, 0.0244, 0.0190, 0.0219, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:36:12,755 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.044e+01 1.577e+02 1.909e+02 2.256e+02 5.766e+02, threshold=3.818e+02, percent-clipped=1.0 +2023-03-27 02:36:21,524 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 02:36:25,985 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8804, 4.2152, 3.7930, 2.0290, 4.2869, 3.0794, 0.9017, 2.9262], + device='cuda:0'), covar=tensor([0.2350, 0.1746, 0.1609, 0.3333, 0.0889, 0.1093, 0.4765, 0.1430], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0176, 0.0158, 0.0129, 0.0159, 0.0122, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 02:36:27,283 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122178.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 02:36:29,612 INFO [finetune.py:976] (0/7) Epoch 22, batch 1900, loss[loss=0.1816, simple_loss=0.2532, pruned_loss=0.05502, over 4732.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2495, pruned_loss=0.05248, over 953642.43 frames. ], batch size: 59, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:36:29,732 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0969, 1.8641, 1.9830, 1.3184, 1.9842, 2.0666, 2.1226, 1.6426], + device='cuda:0'), covar=tensor([0.0652, 0.0719, 0.0781, 0.0948, 0.0731, 0.0746, 0.0627, 0.1149], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0121, 0.0127, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:36:31,046 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-03-27 02:36:37,614 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122195.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:03,559 INFO [finetune.py:976] (0/7) Epoch 22, batch 1950, loss[loss=0.2208, simple_loss=0.2771, pruned_loss=0.08227, over 4216.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2489, pruned_loss=0.05259, over 953913.10 frames. ], batch size: 65, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:37:06,785 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-27 02:37:09,770 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-27 02:37:18,270 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122256.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:19,980 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.844e+01 1.567e+02 1.786e+02 2.225e+02 4.203e+02, threshold=3.573e+02, percent-clipped=2.0 +2023-03-27 02:37:36,887 INFO [finetune.py:976] (0/7) Epoch 22, batch 2000, loss[loss=0.1872, simple_loss=0.2372, pruned_loss=0.06865, over 4829.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2461, pruned_loss=0.05197, over 953554.24 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:37:51,521 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122306.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:59,448 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122319.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:38:02,213 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-27 02:38:10,018 INFO [finetune.py:976] (0/7) Epoch 22, batch 2050, loss[loss=0.1554, simple_loss=0.2215, pruned_loss=0.04467, over 4912.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2421, pruned_loss=0.05075, over 953796.15 frames. ], batch size: 37, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:38:26,819 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.990e+01 1.410e+02 1.787e+02 2.088e+02 3.673e+02, threshold=3.574e+02, percent-clipped=1.0 +2023-03-27 02:38:42,521 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:38:44,604 INFO [finetune.py:976] (0/7) Epoch 22, batch 2100, loss[loss=0.1622, simple_loss=0.2412, pruned_loss=0.04159, over 4945.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2418, pruned_loss=0.05058, over 951903.08 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:38:55,963 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7662, 2.5346, 2.1738, 2.9734, 2.7953, 2.3611, 3.1401, 2.7480], + device='cuda:0'), covar=tensor([0.1317, 0.2360, 0.3110, 0.2447, 0.2591, 0.1632, 0.3143, 0.1983], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0254, 0.0248, 0.0204, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:39:28,194 INFO [finetune.py:976] (0/7) Epoch 22, batch 2150, loss[loss=0.193, simple_loss=0.2458, pruned_loss=0.07006, over 4828.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.245, pruned_loss=0.05207, over 952133.28 frames. ], batch size: 30, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:40:03,769 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.591e+02 1.910e+02 2.352e+02 5.051e+02, threshold=3.819e+02, percent-clipped=3.0 +2023-03-27 02:40:04,684 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 02:40:13,689 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:40:16,626 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122473.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:40:25,439 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4516, 2.2489, 1.7126, 1.0375, 1.9446, 2.0335, 1.8250, 1.9982], + device='cuda:0'), covar=tensor([0.0700, 0.0699, 0.1446, 0.1759, 0.1221, 0.1663, 0.1866, 0.0824], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0193, 0.0200, 0.0183, 0.0210, 0.0208, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:40:26,432 INFO [finetune.py:976] (0/7) Epoch 22, batch 2200, loss[loss=0.1503, simple_loss=0.2263, pruned_loss=0.0371, over 3946.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2462, pruned_loss=0.05182, over 951419.85 frames. ], batch size: 17, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:40:56,788 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:40:58,943 INFO [finetune.py:976] (0/7) Epoch 22, batch 2250, loss[loss=0.1592, simple_loss=0.2341, pruned_loss=0.04213, over 4875.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2457, pruned_loss=0.05112, over 949419.74 frames. ], batch size: 34, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:41:12,977 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:41:17,776 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.977e+01 1.468e+02 1.830e+02 2.095e+02 3.153e+02, threshold=3.659e+02, percent-clipped=0.0 +2023-03-27 02:41:31,700 INFO [finetune.py:976] (0/7) Epoch 22, batch 2300, loss[loss=0.143, simple_loss=0.2137, pruned_loss=0.03615, over 3283.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2457, pruned_loss=0.05061, over 950848.51 frames. ], batch size: 14, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:41:49,477 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122606.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:05,213 INFO [finetune.py:976] (0/7) Epoch 22, batch 2350, loss[loss=0.1728, simple_loss=0.2264, pruned_loss=0.05961, over 4823.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.244, pruned_loss=0.05006, over 953437.61 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:42:11,732 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8968, 1.8244, 1.9950, 1.0991, 2.0042, 1.9714, 1.9629, 1.7061], + device='cuda:0'), covar=tensor([0.0579, 0.0692, 0.0654, 0.0938, 0.0791, 0.0689, 0.0680, 0.1029], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0136, 0.0140, 0.0120, 0.0125, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:42:21,478 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:24,441 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.417e+02 1.625e+02 2.018e+02 3.172e+02, threshold=3.250e+02, percent-clipped=0.0 +2023-03-27 02:42:26,479 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-27 02:42:31,438 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 02:42:32,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9172, 1.7892, 1.5949, 1.4132, 1.9066, 1.6728, 1.8437, 1.8865], + device='cuda:0'), covar=tensor([0.1383, 0.1805, 0.2979, 0.2403, 0.2575, 0.1648, 0.2672, 0.1731], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0189, 0.0237, 0.0256, 0.0249, 0.0205, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:42:34,173 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122675.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:38,347 INFO [finetune.py:976] (0/7) Epoch 22, batch 2400, loss[loss=0.1838, simple_loss=0.242, pruned_loss=0.06282, over 4911.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2428, pruned_loss=0.05012, over 953972.73 frames. ], batch size: 43, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:42:41,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2853, 1.8137, 2.2800, 2.1815, 1.9124, 1.9635, 2.1328, 2.0813], + device='cuda:0'), covar=tensor([0.4149, 0.4200, 0.3155, 0.3937, 0.5176, 0.4027, 0.4927, 0.3254], + device='cuda:0'), in_proj_covar=tensor([0.0258, 0.0243, 0.0264, 0.0284, 0.0283, 0.0259, 0.0293, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:43:11,474 INFO [finetune.py:976] (0/7) Epoch 22, batch 2450, loss[loss=0.1585, simple_loss=0.2359, pruned_loss=0.0405, over 4863.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2403, pruned_loss=0.04991, over 955584.77 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:43:28,861 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-27 02:43:31,104 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.588e+02 1.841e+02 2.130e+02 2.968e+02, threshold=3.682e+02, percent-clipped=0.0 +2023-03-27 02:43:39,659 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122773.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:43:44,590 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-27 02:43:45,020 INFO [finetune.py:976] (0/7) Epoch 22, batch 2500, loss[loss=0.2099, simple_loss=0.2764, pruned_loss=0.07173, over 4895.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2409, pruned_loss=0.05063, over 957240.54 frames. ], batch size: 32, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:44:00,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6218, 1.3681, 1.9877, 1.8519, 1.5754, 3.3715, 1.4156, 1.5252], + device='cuda:0'), covar=tensor([0.0873, 0.1750, 0.1074, 0.0844, 0.1538, 0.0235, 0.1415, 0.1719], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0076, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:44:21,539 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122821.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:23,360 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:28,112 INFO [finetune.py:976] (0/7) Epoch 22, batch 2550, loss[loss=0.1595, simple_loss=0.2181, pruned_loss=0.05049, over 4060.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2436, pruned_loss=0.05123, over 956698.95 frames. ], batch size: 17, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:44:42,554 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122851.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:53,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.566e+02 1.889e+02 2.331e+02 3.878e+02, threshold=3.777e+02, percent-clipped=1.0 +2023-03-27 02:45:07,295 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:45:22,082 INFO [finetune.py:976] (0/7) Epoch 22, batch 2600, loss[loss=0.148, simple_loss=0.2385, pruned_loss=0.02881, over 4889.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.245, pruned_loss=0.05154, over 956248.57 frames. ], batch size: 43, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:45:35,522 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6257, 1.4985, 2.0970, 1.9162, 1.6358, 3.5791, 1.4394, 1.6369], + device='cuda:0'), covar=tensor([0.0940, 0.1787, 0.1027, 0.0951, 0.1658, 0.0229, 0.1491, 0.1758], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:45:36,675 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122899.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:02,678 INFO [finetune.py:976] (0/7) Epoch 22, batch 2650, loss[loss=0.185, simple_loss=0.263, pruned_loss=0.05353, over 4907.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.247, pruned_loss=0.05244, over 955539.21 frames. ], batch size: 36, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:46:03,435 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:19,588 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.898e+01 1.493e+02 1.785e+02 2.135e+02 3.458e+02, threshold=3.570e+02, percent-clipped=0.0 +2023-03-27 02:46:31,737 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122975.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:35,902 INFO [finetune.py:976] (0/7) Epoch 22, batch 2700, loss[loss=0.1563, simple_loss=0.227, pruned_loss=0.04282, over 4833.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2465, pruned_loss=0.05195, over 955536.24 frames. ], batch size: 47, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:04,315 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123023.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:47:09,736 INFO [finetune.py:976] (0/7) Epoch 22, batch 2750, loss[loss=0.1534, simple_loss=0.2228, pruned_loss=0.04201, over 4930.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2446, pruned_loss=0.05137, over 956405.46 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:26,634 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.465e+02 1.704e+02 2.045e+02 3.535e+02, threshold=3.409e+02, percent-clipped=0.0 +2023-03-27 02:47:42,976 INFO [finetune.py:976] (0/7) Epoch 22, batch 2800, loss[loss=0.215, simple_loss=0.2771, pruned_loss=0.0764, over 4763.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2422, pruned_loss=0.05103, over 954480.89 frames. ], batch size: 26, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:49,166 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7144, 0.7039, 1.7999, 1.6878, 1.5906, 1.5170, 1.5654, 1.7471], + device='cuda:0'), covar=tensor([0.3161, 0.3295, 0.2510, 0.2861, 0.3567, 0.3033, 0.3435, 0.2529], + device='cuda:0'), in_proj_covar=tensor([0.0258, 0.0243, 0.0264, 0.0284, 0.0283, 0.0258, 0.0293, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:48:10,887 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123124.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:48:16,586 INFO [finetune.py:976] (0/7) Epoch 22, batch 2850, loss[loss=0.2294, simple_loss=0.2871, pruned_loss=0.08584, over 4852.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2415, pruned_loss=0.05127, over 955754.50 frames. ], batch size: 49, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:48:23,799 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5354, 1.3509, 1.6712, 1.6606, 1.6026, 2.9952, 1.3788, 1.5186], + device='cuda:0'), covar=tensor([0.0885, 0.1896, 0.1087, 0.0921, 0.1624, 0.0322, 0.1570, 0.1927], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 02:48:29,329 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4899, 1.4079, 1.8625, 2.8409, 1.9625, 1.9959, 1.0381, 2.3947], + device='cuda:0'), covar=tensor([0.1715, 0.1419, 0.1196, 0.0570, 0.0829, 0.1687, 0.1595, 0.0501], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:48:33,487 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.606e+02 1.919e+02 2.375e+02 6.875e+02, threshold=3.839e+02, percent-clipped=7.0 +2023-03-27 02:48:42,224 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:48:49,677 INFO [finetune.py:976] (0/7) Epoch 22, batch 2900, loss[loss=0.1903, simple_loss=0.2414, pruned_loss=0.06965, over 4075.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2441, pruned_loss=0.05197, over 955487.94 frames. ], batch size: 17, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:04,599 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4488, 1.3566, 1.8590, 2.8661, 1.9692, 1.9792, 0.7878, 2.4291], + device='cuda:0'), covar=tensor([0.1832, 0.1429, 0.1292, 0.0668, 0.0846, 0.1525, 0.1850, 0.0519], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0135, 0.0165, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:49:10,553 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7720, 1.1216, 0.9467, 1.6645, 2.1850, 1.1508, 1.4922, 1.5023], + device='cuda:0'), covar=tensor([0.1358, 0.2195, 0.1795, 0.1132, 0.1724, 0.1863, 0.1417, 0.1992], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 02:49:22,629 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:49:25,026 INFO [finetune.py:976] (0/7) Epoch 22, batch 2950, loss[loss=0.1578, simple_loss=0.2382, pruned_loss=0.03869, over 4861.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.248, pruned_loss=0.05316, over 955084.37 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:28,188 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 02:49:41,287 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5960, 1.4709, 2.1407, 3.3834, 2.3236, 2.2685, 0.9472, 2.7695], + device='cuda:0'), covar=tensor([0.1748, 0.1438, 0.1322, 0.0532, 0.0756, 0.1732, 0.1816, 0.0454], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:49:42,387 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.515e+01 1.556e+02 1.822e+02 2.269e+02 3.192e+02, threshold=3.643e+02, percent-clipped=0.0 +2023-03-27 02:49:42,485 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4673, 1.4596, 2.2012, 3.2444, 2.2067, 2.2148, 1.0713, 2.6883], + device='cuda:0'), covar=tensor([0.1755, 0.1391, 0.1240, 0.0587, 0.0809, 0.1830, 0.1757, 0.0495], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 02:49:59,811 INFO [finetune.py:976] (0/7) Epoch 22, batch 3000, loss[loss=0.1791, simple_loss=0.254, pruned_loss=0.05207, over 4914.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2484, pruned_loss=0.05373, over 955422.89 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:59,812 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 02:50:11,237 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8132, 1.6548, 1.4934, 1.8750, 2.1133, 1.8443, 1.3880, 1.5005], + device='cuda:0'), covar=tensor([0.2306, 0.2155, 0.2077, 0.1811, 0.1528, 0.1277, 0.2520, 0.1983], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0213, 0.0196, 0.0242, 0.0189, 0.0217, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:50:15,181 INFO [finetune.py:1010] (0/7) Epoch 22, validation: loss=0.1575, simple_loss=0.2256, pruned_loss=0.04471, over 2265189.00 frames. +2023-03-27 02:50:15,182 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 02:50:47,926 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:50:48,100 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 02:51:07,177 INFO [finetune.py:976] (0/7) Epoch 22, batch 3050, loss[loss=0.2085, simple_loss=0.2692, pruned_loss=0.07384, over 4829.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2489, pruned_loss=0.05356, over 955794.32 frames. ], batch size: 47, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:51:16,032 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 02:51:26,742 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3159, 1.4573, 1.5049, 0.7689, 1.4780, 1.6882, 1.7209, 1.3935], + device='cuda:0'), covar=tensor([0.0915, 0.0601, 0.0536, 0.0605, 0.0518, 0.0623, 0.0338, 0.0712], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0132, 0.0130, 0.0141, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9726e-05, 1.0823e-04, 9.1111e-05, 8.6610e-05, 9.2816e-05, 9.2981e-05, + 1.0110e-04, 1.0659e-04], device='cuda:0') +2023-03-27 02:51:27,199 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.922e+01 1.647e+02 1.960e+02 2.377e+02 4.726e+02, threshold=3.920e+02, percent-clipped=6.0 +2023-03-27 02:51:34,721 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1874, 2.1097, 1.7240, 2.1096, 2.0798, 1.8809, 2.4777, 2.2377], + device='cuda:0'), covar=tensor([0.1268, 0.1944, 0.2742, 0.2511, 0.2498, 0.1642, 0.2975, 0.1622], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0254, 0.0248, 0.0204, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:51:36,573 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123375.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:51:41,110 INFO [finetune.py:976] (0/7) Epoch 22, batch 3100, loss[loss=0.1579, simple_loss=0.2142, pruned_loss=0.05078, over 4170.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.247, pruned_loss=0.05265, over 955940.99 frames. ], batch size: 18, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:00,805 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 02:52:12,218 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123428.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:52:14,579 INFO [finetune.py:976] (0/7) Epoch 22, batch 3150, loss[loss=0.1355, simple_loss=0.209, pruned_loss=0.03103, over 4786.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2441, pruned_loss=0.05132, over 956268.26 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:34,405 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.485e+02 1.808e+02 2.093e+02 3.688e+02, threshold=3.617e+02, percent-clipped=0.0 +2023-03-27 02:52:47,863 INFO [finetune.py:976] (0/7) Epoch 22, batch 3200, loss[loss=0.1713, simple_loss=0.244, pruned_loss=0.04936, over 4791.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2409, pruned_loss=0.05023, over 953352.15 frames. ], batch size: 51, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:52,366 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123489.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:18,963 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:21,317 INFO [finetune.py:976] (0/7) Epoch 22, batch 3250, loss[loss=0.192, simple_loss=0.2693, pruned_loss=0.05733, over 4845.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2406, pruned_loss=0.05047, over 950429.52 frames. ], batch size: 49, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:53:33,481 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9675, 1.9200, 1.5838, 1.7934, 1.9657, 1.7146, 2.1530, 2.0007], + device='cuda:0'), covar=tensor([0.1330, 0.1909, 0.2744, 0.2261, 0.2402, 0.1557, 0.3070, 0.1642], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0254, 0.0248, 0.0204, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:53:41,211 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.285e+01 1.537e+02 1.740e+02 2.121e+02 3.629e+02, threshold=3.481e+02, percent-clipped=1.0 +2023-03-27 02:53:47,538 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0122, 1.9399, 1.6989, 1.8663, 1.9869, 1.7967, 2.0940, 2.0722], + device='cuda:0'), covar=tensor([0.1235, 0.1925, 0.2510, 0.2191, 0.2106, 0.1453, 0.2848, 0.1489], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0234, 0.0253, 0.0247, 0.0204, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:53:51,141 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:54,732 INFO [finetune.py:976] (0/7) Epoch 22, batch 3300, loss[loss=0.2023, simple_loss=0.2808, pruned_loss=0.06187, over 4925.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2449, pruned_loss=0.05204, over 950681.52 frames. ], batch size: 36, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:54:28,302 INFO [finetune.py:976] (0/7) Epoch 22, batch 3350, loss[loss=0.1624, simple_loss=0.2317, pruned_loss=0.04655, over 4779.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2473, pruned_loss=0.0526, over 950944.54 frames. ], batch size: 25, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:54:47,671 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.542e+02 1.809e+02 2.055e+02 5.285e+02, threshold=3.617e+02, percent-clipped=2.0 +2023-03-27 02:54:54,262 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123670.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:55:01,512 INFO [finetune.py:976] (0/7) Epoch 22, batch 3400, loss[loss=0.168, simple_loss=0.2519, pruned_loss=0.042, over 4812.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2491, pruned_loss=0.05339, over 951465.89 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:55:54,857 INFO [finetune.py:976] (0/7) Epoch 22, batch 3450, loss[loss=0.1842, simple_loss=0.2437, pruned_loss=0.06233, over 4743.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2492, pruned_loss=0.05367, over 950717.74 frames. ], batch size: 54, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:56:26,817 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.545e+02 1.898e+02 2.347e+02 3.548e+02, threshold=3.797e+02, percent-clipped=0.0 +2023-03-27 02:56:44,277 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 02:56:45,230 INFO [finetune.py:976] (0/7) Epoch 22, batch 3500, loss[loss=0.1337, simple_loss=0.2052, pruned_loss=0.03114, over 4842.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2476, pruned_loss=0.05364, over 952725.37 frames. ], batch size: 49, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:56:46,496 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:56:53,203 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7594, 3.3583, 3.4503, 3.6020, 3.5480, 3.2307, 3.8507, 1.3580], + device='cuda:0'), covar=tensor([0.0970, 0.0860, 0.1012, 0.1032, 0.1506, 0.1801, 0.0825, 0.5177], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0280, 0.0290, 0.0334, 0.0284, 0.0305, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:56:57,574 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7993, 1.9909, 1.7119, 1.6635, 2.3671, 2.3654, 2.0331, 1.9785], + device='cuda:0'), covar=tensor([0.0445, 0.0294, 0.0513, 0.0331, 0.0224, 0.0439, 0.0359, 0.0318], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0106, 0.0142, 0.0110, 0.0098, 0.0110, 0.0100, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.5884e-05, 8.1605e-05, 1.1130e-04, 8.4694e-05, 7.5843e-05, 8.1325e-05, + 7.4360e-05, 8.4623e-05], device='cuda:0') +2023-03-27 02:56:59,300 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2509, 2.1372, 1.8535, 2.1498, 2.0626, 2.0895, 2.0717, 2.7531], + device='cuda:0'), covar=tensor([0.3758, 0.3808, 0.3194, 0.3652, 0.3786, 0.2394, 0.3639, 0.1691], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0264, 0.0235, 0.0277, 0.0255, 0.0226, 0.0254, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:57:01,149 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 02:57:18,065 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 02:57:18,476 INFO [finetune.py:976] (0/7) Epoch 22, batch 3550, loss[loss=0.1845, simple_loss=0.246, pruned_loss=0.06152, over 4826.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2448, pruned_loss=0.05246, over 954070.57 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:57:30,909 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-27 02:57:36,087 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.884e+01 1.429e+02 1.766e+02 2.143e+02 3.754e+02, threshold=3.531e+02, percent-clipped=0.0 +2023-03-27 02:57:36,417 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-03-27 02:57:45,445 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:57:52,356 INFO [finetune.py:976] (0/7) Epoch 22, batch 3600, loss[loss=0.1743, simple_loss=0.2429, pruned_loss=0.05283, over 4801.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2429, pruned_loss=0.05182, over 956624.65 frames. ], batch size: 51, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:57:54,902 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6100, 1.4586, 1.8911, 1.2645, 1.5201, 1.8576, 1.4062, 1.9810], + device='cuda:0'), covar=tensor([0.1275, 0.2219, 0.1348, 0.1652, 0.1037, 0.1277, 0.3002, 0.0844], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0206, 0.0191, 0.0189, 0.0174, 0.0213, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:57:58,528 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3054, 2.9799, 3.0719, 3.2097, 3.1039, 2.9182, 3.3417, 0.9584], + device='cuda:0'), covar=tensor([0.1142, 0.0988, 0.1198, 0.1240, 0.1674, 0.1960, 0.1161, 0.5833], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0279, 0.0290, 0.0334, 0.0284, 0.0304, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 02:58:25,719 INFO [finetune.py:976] (0/7) Epoch 22, batch 3650, loss[loss=0.1653, simple_loss=0.2525, pruned_loss=0.0391, over 4841.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2443, pruned_loss=0.05187, over 956651.66 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:58:26,472 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:58:43,171 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.106e+01 1.598e+02 1.925e+02 2.525e+02 5.508e+02, threshold=3.851e+02, percent-clipped=5.0 +2023-03-27 02:58:49,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123970.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:58:57,340 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 02:58:59,577 INFO [finetune.py:976] (0/7) Epoch 22, batch 3700, loss[loss=0.1811, simple_loss=0.2631, pruned_loss=0.0496, over 4913.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2491, pruned_loss=0.05346, over 955856.71 frames. ], batch size: 37, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:59:10,707 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-124000.pt +2023-03-27 02:59:23,314 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124018.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:59:34,573 INFO [finetune.py:976] (0/7) Epoch 22, batch 3750, loss[loss=0.1517, simple_loss=0.2303, pruned_loss=0.03652, over 4855.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2507, pruned_loss=0.05441, over 954267.17 frames. ], batch size: 44, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 02:59:37,113 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124036.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:59:42,186 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 02:59:51,700 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.541e+02 1.786e+02 2.095e+02 2.976e+02, threshold=3.572e+02, percent-clipped=0.0 +2023-03-27 02:59:53,028 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124062.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:06,947 INFO [finetune.py:976] (0/7) Epoch 22, batch 3800, loss[loss=0.1626, simple_loss=0.2441, pruned_loss=0.04051, over 4831.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2507, pruned_loss=0.054, over 954107.69 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:00:08,692 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:17,183 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124097.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:39,613 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:51,384 INFO [finetune.py:976] (0/7) Epoch 22, batch 3850, loss[loss=0.1587, simple_loss=0.2333, pruned_loss=0.04207, over 4755.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2489, pruned_loss=0.05283, over 951708.42 frames. ], batch size: 28, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:00:51,447 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124132.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:01:03,891 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4194, 1.3839, 1.3246, 1.3841, 1.7636, 1.6882, 1.4373, 1.3026], + device='cuda:0'), covar=tensor([0.0403, 0.0301, 0.0597, 0.0306, 0.0187, 0.0430, 0.0322, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0106, 0.0142, 0.0110, 0.0098, 0.0110, 0.0100, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.6032e-05, 8.1328e-05, 1.1134e-04, 8.4687e-05, 7.6051e-05, 8.1368e-05, + 7.4438e-05, 8.4431e-05], device='cuda:0') +2023-03-27 03:01:20,790 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.486e+02 1.746e+02 2.060e+02 3.550e+02, threshold=3.491e+02, percent-clipped=0.0 +2023-03-27 03:01:46,862 INFO [finetune.py:976] (0/7) Epoch 22, batch 3900, loss[loss=0.1494, simple_loss=0.225, pruned_loss=0.03689, over 4822.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2462, pruned_loss=0.05195, over 950728.04 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:01:49,297 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5524, 1.6085, 2.2677, 1.8105, 1.8463, 4.1162, 1.5904, 1.7723], + device='cuda:0'), covar=tensor([0.0982, 0.1660, 0.1295, 0.1016, 0.1497, 0.0217, 0.1404, 0.1665], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:02:11,440 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0935, 1.8570, 2.0411, 0.8843, 2.3568, 2.4701, 2.1793, 1.9397], + device='cuda:0'), covar=tensor([0.0926, 0.1000, 0.0508, 0.0706, 0.0623, 0.0697, 0.0530, 0.0771], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0128, 0.0123, 0.0132, 0.0131, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.0332e-05, 1.0868e-04, 9.1738e-05, 8.6948e-05, 9.2872e-05, 9.3453e-05, + 1.0186e-04, 1.0699e-04], device='cuda:0') +2023-03-27 03:02:17,916 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:02:20,286 INFO [finetune.py:976] (0/7) Epoch 22, batch 3950, loss[loss=0.14, simple_loss=0.2031, pruned_loss=0.03841, over 4394.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2427, pruned_loss=0.0508, over 951506.90 frames. ], batch size: 19, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:02:22,616 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.9532, 4.3046, 4.4725, 4.7352, 4.7070, 4.4333, 5.0702, 1.6184], + device='cuda:0'), covar=tensor([0.0671, 0.0743, 0.0760, 0.0856, 0.1068, 0.1371, 0.0474, 0.5789], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0244, 0.0279, 0.0290, 0.0333, 0.0284, 0.0305, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:02:39,959 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.559e+02 1.897e+02 2.284e+02 3.853e+02, threshold=3.794e+02, percent-clipped=3.0 +2023-03-27 03:02:47,325 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:02:48,662 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 03:02:49,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8503, 4.6042, 4.2740, 2.2489, 4.7048, 3.5992, 0.9024, 3.2088], + device='cuda:0'), covar=tensor([0.2397, 0.2032, 0.1358, 0.3350, 0.0767, 0.0865, 0.4692, 0.1421], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0180, 0.0161, 0.0131, 0.0163, 0.0124, 0.0150, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 03:02:53,668 INFO [finetune.py:976] (0/7) Epoch 22, batch 4000, loss[loss=0.171, simple_loss=0.2437, pruned_loss=0.04918, over 4763.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2425, pruned_loss=0.05169, over 952305.36 frames. ], batch size: 27, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:02:53,843 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-27 03:03:10,246 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6919, 1.5887, 1.9588, 1.2752, 1.6730, 1.9078, 1.5260, 2.0773], + device='cuda:0'), covar=tensor([0.1255, 0.1958, 0.1203, 0.1673, 0.0932, 0.1246, 0.2574, 0.0749], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0206, 0.0190, 0.0189, 0.0174, 0.0215, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:03:18,773 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-03-27 03:03:20,592 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.37 vs. limit=5.0 +2023-03-27 03:03:25,328 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2936, 2.9245, 3.0456, 3.2224, 3.0895, 2.8979, 3.3673, 0.8731], + device='cuda:0'), covar=tensor([0.1127, 0.1055, 0.1071, 0.1127, 0.1622, 0.1847, 0.1124, 0.5762], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0245, 0.0281, 0.0292, 0.0336, 0.0286, 0.0306, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:03:26,973 INFO [finetune.py:976] (0/7) Epoch 22, batch 4050, loss[loss=0.2642, simple_loss=0.3188, pruned_loss=0.1048, over 4812.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2466, pruned_loss=0.05343, over 951986.14 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:03:27,709 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124333.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:03:31,861 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3454, 1.5282, 1.6442, 0.9219, 1.4958, 1.8397, 1.8464, 1.4997], + device='cuda:0'), covar=tensor([0.0839, 0.0614, 0.0514, 0.0494, 0.0524, 0.0539, 0.0306, 0.0620], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0122, 0.0131, 0.0130, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9460e-05, 1.0799e-04, 9.0851e-05, 8.6260e-05, 9.2297e-05, 9.2647e-05, + 1.0101e-04, 1.0631e-04], device='cuda:0') +2023-03-27 03:03:34,306 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1926, 1.2963, 1.3317, 0.7512, 1.2234, 1.5261, 1.5684, 1.2556], + device='cuda:0'), covar=tensor([0.0819, 0.0560, 0.0512, 0.0497, 0.0501, 0.0596, 0.0302, 0.0583], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0122, 0.0131, 0.0130, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9483e-05, 1.0803e-04, 9.0856e-05, 8.6270e-05, 9.2310e-05, 9.2664e-05, + 1.0101e-04, 1.0634e-04], device='cuda:0') +2023-03-27 03:03:39,248 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124348.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:03:39,286 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2465, 2.0626, 1.6377, 2.2595, 2.1014, 1.8635, 2.5181, 2.2881], + device='cuda:0'), covar=tensor([0.1269, 0.2228, 0.2959, 0.2501, 0.2514, 0.1681, 0.3008, 0.1749], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0188, 0.0235, 0.0253, 0.0247, 0.0204, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:03:46,866 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.585e+02 1.909e+02 2.213e+02 3.315e+02, threshold=3.818e+02, percent-clipped=0.0 +2023-03-27 03:04:00,187 INFO [finetune.py:976] (0/7) Epoch 22, batch 4100, loss[loss=0.1938, simple_loss=0.2676, pruned_loss=0.06001, over 4924.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2497, pruned_loss=0.05388, over 953847.41 frames. ], batch size: 33, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:04:01,489 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 03:04:07,292 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124392.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:08,704 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-27 03:04:19,557 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124409.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:04:24,895 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124418.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:33,281 INFO [finetune.py:976] (0/7) Epoch 22, batch 4150, loss[loss=0.1513, simple_loss=0.2217, pruned_loss=0.0405, over 4852.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2509, pruned_loss=0.05413, over 954876.32 frames. ], batch size: 31, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:04:33,385 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:44,092 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3509, 2.2265, 1.7504, 2.3555, 2.2303, 1.9477, 2.6252, 2.3563], + device='cuda:0'), covar=tensor([0.1334, 0.2020, 0.3094, 0.2388, 0.2655, 0.1739, 0.2731, 0.1699], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0189, 0.0236, 0.0254, 0.0248, 0.0204, 0.0216, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:04:53,595 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.819e+01 1.555e+02 1.746e+02 2.178e+02 5.076e+02, threshold=3.492e+02, percent-clipped=2.0 +2023-03-27 03:05:06,935 INFO [finetune.py:976] (0/7) Epoch 22, batch 4200, loss[loss=0.1758, simple_loss=0.2304, pruned_loss=0.06057, over 4764.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2512, pruned_loss=0.05397, over 954950.58 frames. ], batch size: 26, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:05:14,200 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:05:20,638 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 03:05:40,281 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:05:42,631 INFO [finetune.py:976] (0/7) Epoch 22, batch 4250, loss[loss=0.1546, simple_loss=0.228, pruned_loss=0.04057, over 4836.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2478, pruned_loss=0.05221, over 956726.77 frames. ], batch size: 47, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:06:02,064 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.200e+01 1.490e+02 1.704e+02 2.072e+02 3.423e+02, threshold=3.408e+02, percent-clipped=0.0 +2023-03-27 03:06:12,354 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:06:17,860 INFO [finetune.py:976] (0/7) Epoch 22, batch 4300, loss[loss=0.2241, simple_loss=0.2897, pruned_loss=0.07922, over 4423.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2459, pruned_loss=0.05171, over 956842.20 frames. ], batch size: 19, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:06:41,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6800, 3.1781, 3.3952, 3.4775, 3.4833, 3.2146, 3.7256, 1.5038], + device='cuda:0'), covar=tensor([0.0782, 0.0841, 0.0852, 0.1074, 0.1234, 0.1507, 0.0824, 0.4968], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0245, 0.0280, 0.0291, 0.0335, 0.0284, 0.0305, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:07:10,856 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124628.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:07:17,350 INFO [finetune.py:976] (0/7) Epoch 22, batch 4350, loss[loss=0.1806, simple_loss=0.2435, pruned_loss=0.05884, over 4833.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2423, pruned_loss=0.05002, over 957012.85 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:07:21,128 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:07:48,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.554e+01 1.414e+02 1.739e+02 2.148e+02 3.782e+02, threshold=3.479e+02, percent-clipped=2.0 +2023-03-27 03:07:50,744 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124663.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:01,828 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 03:08:06,298 INFO [finetune.py:976] (0/7) Epoch 22, batch 4400, loss[loss=0.2254, simple_loss=0.294, pruned_loss=0.07837, over 4823.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2437, pruned_loss=0.05106, over 956088.66 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:08:11,258 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3815, 3.7754, 3.9845, 4.2040, 4.1301, 3.8305, 4.4860, 1.3833], + device='cuda:0'), covar=tensor([0.0770, 0.0860, 0.0811, 0.0948, 0.1277, 0.1769, 0.0642, 0.5891], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0279, 0.0290, 0.0334, 0.0283, 0.0304, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:08:12,520 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:17,301 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124699.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:20,762 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124704.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:08:31,137 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:31,843 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 03:08:35,322 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124724.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:40,101 INFO [finetune.py:976] (0/7) Epoch 22, batch 4450, loss[loss=0.1455, simple_loss=0.2269, pruned_loss=0.03203, over 4762.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2465, pruned_loss=0.05203, over 955032.42 frames. ], batch size: 27, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:08:45,009 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:58,177 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.229e+02 1.653e+02 1.910e+02 2.206e+02 5.425e+02, threshold=3.820e+02, percent-clipped=3.0 +2023-03-27 03:09:02,352 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:13,853 INFO [finetune.py:976] (0/7) Epoch 22, batch 4500, loss[loss=0.1631, simple_loss=0.2375, pruned_loss=0.04433, over 4773.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.05228, over 954862.69 frames. ], batch size: 26, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:09:17,594 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:38,644 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:47,291 INFO [finetune.py:976] (0/7) Epoch 22, batch 4550, loss[loss=0.1767, simple_loss=0.2535, pruned_loss=0.04997, over 4876.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.25, pruned_loss=0.05307, over 955354.35 frames. ], batch size: 35, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:10:03,813 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-27 03:10:04,841 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.469e+02 1.742e+02 1.998e+02 4.285e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 03:10:19,580 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:10:20,090 INFO [finetune.py:976] (0/7) Epoch 22, batch 4600, loss[loss=0.1449, simple_loss=0.2193, pruned_loss=0.03524, over 4742.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2501, pruned_loss=0.05332, over 955108.14 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:10:26,277 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-27 03:10:47,654 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3823, 2.2362, 1.8736, 2.1102, 2.2661, 2.0176, 2.4936, 2.2829], + device='cuda:0'), covar=tensor([0.1184, 0.1853, 0.2787, 0.2336, 0.2405, 0.1599, 0.2787, 0.1591], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0234, 0.0253, 0.0247, 0.0204, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:10:48,202 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8170, 1.2328, 0.9822, 1.6561, 2.1271, 1.4665, 1.5591, 1.6558], + device='cuda:0'), covar=tensor([0.1345, 0.2067, 0.1838, 0.1065, 0.1846, 0.1850, 0.1370, 0.1945], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:10:50,460 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124928.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:10:53,306 INFO [finetune.py:976] (0/7) Epoch 22, batch 4650, loss[loss=0.1543, simple_loss=0.2256, pruned_loss=0.04153, over 4908.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2468, pruned_loss=0.05238, over 953820.94 frames. ], batch size: 36, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:11:10,703 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.467e+02 1.675e+02 2.008e+02 3.769e+02, threshold=3.350e+02, percent-clipped=1.0 +2023-03-27 03:11:21,849 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:11:26,384 INFO [finetune.py:976] (0/7) Epoch 22, batch 4700, loss[loss=0.1578, simple_loss=0.2288, pruned_loss=0.04342, over 4906.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2432, pruned_loss=0.05087, over 955108.24 frames. ], batch size: 35, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:11:36,919 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:11:43,164 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125004.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:11:49,267 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0759, 1.9880, 1.6378, 1.8006, 2.0092, 1.7604, 2.2754, 2.0614], + device='cuda:0'), covar=tensor([0.1311, 0.1797, 0.2807, 0.2344, 0.2502, 0.1613, 0.3063, 0.1570], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0253, 0.0248, 0.0204, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:11:52,756 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125019.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:12:07,774 INFO [finetune.py:976] (0/7) Epoch 22, batch 4750, loss[loss=0.2406, simple_loss=0.2914, pruned_loss=0.09494, over 4124.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2421, pruned_loss=0.05067, over 955884.98 frames. ], batch size: 65, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:12:30,807 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125052.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:12:39,937 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.467e+02 1.710e+02 2.084e+02 3.277e+02, threshold=3.420e+02, percent-clipped=0.0 +2023-03-27 03:13:08,204 INFO [finetune.py:976] (0/7) Epoch 22, batch 4800, loss[loss=0.1947, simple_loss=0.2695, pruned_loss=0.05998, over 4924.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2441, pruned_loss=0.05144, over 956100.01 frames. ], batch size: 38, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:13:16,561 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:13:42,361 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125128.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:13:45,184 INFO [finetune.py:976] (0/7) Epoch 22, batch 4850, loss[loss=0.1605, simple_loss=0.2376, pruned_loss=0.04175, over 4811.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2465, pruned_loss=0.05187, over 953936.82 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 64.0 +2023-03-27 03:13:45,277 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.1662, 4.4879, 4.7499, 4.9793, 4.8871, 4.5571, 5.2744, 1.5562], + device='cuda:0'), covar=tensor([0.0665, 0.0809, 0.0904, 0.0905, 0.1192, 0.1702, 0.0516, 0.6345], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0243, 0.0279, 0.0291, 0.0334, 0.0284, 0.0305, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:13:47,688 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:04,216 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.665e+02 1.915e+02 2.224e+02 3.844e+02, threshold=3.831e+02, percent-clipped=1.0 +2023-03-27 03:14:14,550 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125176.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:18,598 INFO [finetune.py:976] (0/7) Epoch 22, batch 4900, loss[loss=0.1633, simple_loss=0.2356, pruned_loss=0.04551, over 4772.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2485, pruned_loss=0.05256, over 955159.72 frames. ], batch size: 28, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:14:23,460 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125189.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:52,064 INFO [finetune.py:976] (0/7) Epoch 22, batch 4950, loss[loss=0.1765, simple_loss=0.2555, pruned_loss=0.0487, over 4807.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.25, pruned_loss=0.05278, over 954556.19 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:14:59,821 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 03:15:12,019 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.500e+02 1.887e+02 2.169e+02 5.445e+02, threshold=3.774e+02, percent-clipped=5.0 +2023-03-27 03:15:25,140 INFO [finetune.py:976] (0/7) Epoch 22, batch 5000, loss[loss=0.1608, simple_loss=0.2334, pruned_loss=0.04411, over 4816.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2481, pruned_loss=0.05186, over 954700.89 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:15:33,516 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125294.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:15:33,562 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5080, 1.5528, 1.5522, 0.8030, 1.6338, 1.7992, 1.8722, 1.4158], + device='cuda:0'), covar=tensor([0.0843, 0.0543, 0.0490, 0.0530, 0.0428, 0.0583, 0.0259, 0.0662], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9511e-05, 1.0858e-04, 9.1033e-05, 8.6614e-05, 9.1988e-05, 9.2691e-05, + 1.0159e-04, 1.0708e-04], device='cuda:0') +2023-03-27 03:15:50,208 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125319.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:15:51,456 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4872, 1.3741, 1.3868, 1.3777, 0.8907, 2.3010, 0.7822, 1.2236], + device='cuda:0'), covar=tensor([0.3461, 0.2666, 0.2297, 0.2602, 0.1979, 0.0382, 0.2710, 0.1365], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:15:58,103 INFO [finetune.py:976] (0/7) Epoch 22, batch 5050, loss[loss=0.1919, simple_loss=0.2504, pruned_loss=0.06667, over 4742.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.245, pruned_loss=0.05077, over 952990.54 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:16:05,257 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125342.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:15,317 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:18,273 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.708e+01 1.493e+02 1.877e+02 2.229e+02 3.838e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-27 03:16:22,398 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125367.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:27,959 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:31,149 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-27 03:16:31,543 INFO [finetune.py:976] (0/7) Epoch 22, batch 5100, loss[loss=0.1716, simple_loss=0.2449, pruned_loss=0.0492, over 4784.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2422, pruned_loss=0.04969, over 954206.47 frames. ], batch size: 29, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:16:32,397 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 03:16:55,557 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125417.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 03:16:58,439 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:01,491 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6252, 1.4965, 1.4120, 1.7870, 1.8862, 1.7308, 1.2335, 1.3736], + device='cuda:0'), covar=tensor([0.2331, 0.2116, 0.2033, 0.1626, 0.1709, 0.1268, 0.2657, 0.1942], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0196, 0.0242, 0.0188, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:17:05,015 INFO [finetune.py:976] (0/7) Epoch 22, batch 5150, loss[loss=0.23, simple_loss=0.2889, pruned_loss=0.08549, over 4831.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2436, pruned_loss=0.0511, over 953294.24 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:17:13,012 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:33,158 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.608e+02 1.768e+02 2.290e+02 4.207e+02, threshold=3.536e+02, percent-clipped=1.0 +2023-03-27 03:17:45,304 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125476.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:48,790 INFO [finetune.py:976] (0/7) Epoch 22, batch 5200, loss[loss=0.1499, simple_loss=0.2373, pruned_loss=0.03127, over 4895.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.0523, over 955443.90 frames. ], batch size: 43, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:17:48,912 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125482.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:54,881 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125484.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:18:14,788 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-03-27 03:18:39,396 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125524.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:18:44,720 INFO [finetune.py:976] (0/7) Epoch 22, batch 5250, loss[loss=0.1508, simple_loss=0.2176, pruned_loss=0.04203, over 4773.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2497, pruned_loss=0.0527, over 957302.33 frames. ], batch size: 26, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:19:03,922 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.572e+02 1.864e+02 2.118e+02 3.675e+02, threshold=3.728e+02, percent-clipped=1.0 +2023-03-27 03:19:05,136 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9069, 1.6262, 2.0839, 1.3775, 1.9484, 2.0904, 1.5231, 2.2419], + device='cuda:0'), covar=tensor([0.1265, 0.2211, 0.1476, 0.1926, 0.0968, 0.1336, 0.2796, 0.0812], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0191, 0.0190, 0.0174, 0.0214, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:19:18,602 INFO [finetune.py:976] (0/7) Epoch 22, batch 5300, loss[loss=0.2136, simple_loss=0.2844, pruned_loss=0.0714, over 4139.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.251, pruned_loss=0.05335, over 955547.07 frames. ], batch size: 65, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:19:30,944 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 03:19:32,531 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125603.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:19:52,427 INFO [finetune.py:976] (0/7) Epoch 22, batch 5350, loss[loss=0.2151, simple_loss=0.2829, pruned_loss=0.07359, over 4714.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2513, pruned_loss=0.05326, over 955453.76 frames. ], batch size: 59, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:19:53,862 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 03:20:10,862 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.759e+01 1.403e+02 1.743e+02 2.180e+02 4.274e+02, threshold=3.486e+02, percent-clipped=1.0 +2023-03-27 03:20:13,357 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125664.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:20:25,088 INFO [finetune.py:976] (0/7) Epoch 22, batch 5400, loss[loss=0.1801, simple_loss=0.2285, pruned_loss=0.06584, over 4059.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2487, pruned_loss=0.05327, over 953678.00 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:20:35,918 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5968, 1.6070, 1.6013, 0.8116, 1.7554, 1.9113, 1.9507, 1.4427], + device='cuda:0'), covar=tensor([0.1020, 0.0736, 0.0576, 0.0649, 0.0465, 0.0564, 0.0307, 0.0839], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0131, 0.0131, 0.0141, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9985e-05, 1.0857e-04, 9.1165e-05, 8.6441e-05, 9.2034e-05, 9.3171e-05, + 1.0120e-04, 1.0689e-04], device='cuda:0') +2023-03-27 03:20:44,655 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125712.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 03:20:53,976 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:20:58,061 INFO [finetune.py:976] (0/7) Epoch 22, batch 5450, loss[loss=0.1561, simple_loss=0.2188, pruned_loss=0.04669, over 4768.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2456, pruned_loss=0.05241, over 952909.18 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:20:58,135 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:17,000 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.470e+02 1.717e+02 2.000e+02 3.602e+02, threshold=3.434e+02, percent-clipped=1.0 +2023-03-27 03:21:27,702 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:31,103 INFO [finetune.py:976] (0/7) Epoch 22, batch 5500, loss[loss=0.1742, simple_loss=0.2419, pruned_loss=0.05328, over 4860.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2407, pruned_loss=0.05031, over 952769.41 frames. ], batch size: 44, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:21:32,412 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:33,631 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125786.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:22:04,442 INFO [finetune.py:976] (0/7) Epoch 22, batch 5550, loss[loss=0.1703, simple_loss=0.2499, pruned_loss=0.04541, over 4757.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.243, pruned_loss=0.05093, over 954011.34 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:22:04,491 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125832.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:22:07,128 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 03:22:32,017 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.619e+02 1.949e+02 2.379e+02 4.295e+02, threshold=3.899e+02, percent-clipped=6.0 +2023-03-27 03:22:44,812 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9314, 1.2242, 0.9133, 1.7973, 2.2723, 1.8584, 1.6439, 1.7161], + device='cuda:0'), covar=tensor([0.1443, 0.2169, 0.2071, 0.1209, 0.1886, 0.1859, 0.1411, 0.1878], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:22:48,802 INFO [finetune.py:976] (0/7) Epoch 22, batch 5600, loss[loss=0.1562, simple_loss=0.236, pruned_loss=0.03816, over 4892.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2469, pruned_loss=0.05221, over 954185.09 frames. ], batch size: 32, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:22:58,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5409, 2.4816, 2.3930, 2.7168, 3.2340, 2.5772, 2.5880, 2.0171], + device='cuda:0'), covar=tensor([0.2320, 0.1826, 0.1821, 0.1605, 0.1517, 0.1159, 0.1794, 0.2012], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0209, 0.0212, 0.0196, 0.0242, 0.0188, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:23:20,374 INFO [finetune.py:976] (0/7) Epoch 22, batch 5650, loss[loss=0.179, simple_loss=0.2532, pruned_loss=0.05238, over 4829.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2502, pruned_loss=0.05272, over 953043.60 frames. ], batch size: 39, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:23:47,398 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125959.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:23:48,544 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.919e+01 1.591e+02 1.865e+02 2.221e+02 3.612e+02, threshold=3.730e+02, percent-clipped=0.0 +2023-03-27 03:23:52,405 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 03:23:53,427 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3813, 2.1234, 2.6720, 1.8844, 2.3136, 2.7569, 2.1408, 2.7486], + device='cuda:0'), covar=tensor([0.1047, 0.1426, 0.1221, 0.1557, 0.0813, 0.0976, 0.1979, 0.0704], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0207, 0.0192, 0.0191, 0.0175, 0.0216, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:24:11,010 INFO [finetune.py:976] (0/7) Epoch 22, batch 5700, loss[loss=0.1724, simple_loss=0.2237, pruned_loss=0.06055, over 3978.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2453, pruned_loss=0.05179, over 934120.91 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:24:20,531 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125998.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:24:21,807 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-126000.pt +2023-03-27 03:24:28,445 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-22.pt +2023-03-27 03:24:37,871 INFO [finetune.py:976] (0/7) Epoch 23, batch 0, loss[loss=0.1321, simple_loss=0.2091, pruned_loss=0.02756, over 4328.00 frames. ], tot_loss[loss=0.1321, simple_loss=0.2091, pruned_loss=0.02756, over 4328.00 frames. ], batch size: 19, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:24:37,872 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 03:24:41,482 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1738, 1.3893, 1.4311, 0.7209, 1.3935, 1.5863, 1.7152, 1.3092], + device='cuda:0'), covar=tensor([0.0892, 0.0656, 0.0575, 0.0507, 0.0507, 0.0707, 0.0325, 0.0751], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0126, 0.0121, 0.0129, 0.0129, 0.0140, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.8915e-05, 1.0707e-04, 9.0447e-05, 8.5543e-05, 9.0683e-05, 9.1748e-05, + 9.9864e-05, 1.0572e-04], device='cuda:0') +2023-03-27 03:24:43,264 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8069, 3.4339, 3.5247, 3.6961, 3.5548, 3.3791, 3.8730, 1.3957], + device='cuda:0'), covar=tensor([0.0904, 0.0816, 0.0954, 0.0999, 0.1563, 0.1744, 0.0866, 0.5423], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0241, 0.0276, 0.0288, 0.0331, 0.0281, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:24:45,150 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8658, 3.4747, 3.5700, 3.7340, 3.6397, 3.4127, 3.9231, 1.3095], + device='cuda:0'), covar=tensor([0.0936, 0.0796, 0.0938, 0.1024, 0.1485, 0.1840, 0.0920, 0.5390], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0241, 0.0276, 0.0288, 0.0331, 0.0281, 0.0301, 0.0295], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:24:53,039 INFO [finetune.py:1010] (0/7) Epoch 23, validation: loss=0.1587, simple_loss=0.2268, pruned_loss=0.04533, over 2265189.00 frames. +2023-03-27 03:24:53,040 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 03:24:55,410 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126012.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:25:01,296 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9713, 1.8262, 1.5313, 1.6461, 1.7032, 1.7315, 1.7287, 2.4235], + device='cuda:0'), covar=tensor([0.3677, 0.4067, 0.3337, 0.3670, 0.4165, 0.2303, 0.3606, 0.1604], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0234, 0.0276, 0.0255, 0.0224, 0.0253, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:25:11,568 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-27 03:25:12,068 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:13,320 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1084, 2.0619, 2.1074, 1.5292, 2.1156, 2.1967, 2.2879, 1.7313], + device='cuda:0'), covar=tensor([0.0592, 0.0636, 0.0709, 0.0852, 0.0699, 0.0730, 0.0587, 0.1141], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0134, 0.0138, 0.0119, 0.0124, 0.0137, 0.0138, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:25:23,367 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6165, 1.0905, 0.8995, 1.4871, 2.0535, 1.1681, 1.4153, 1.4489], + device='cuda:0'), covar=tensor([0.1501, 0.2261, 0.1997, 0.1268, 0.2063, 0.1914, 0.1581, 0.1997], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:25:30,342 INFO [finetune.py:976] (0/7) Epoch 23, batch 50, loss[loss=0.1518, simple_loss=0.2316, pruned_loss=0.036, over 4895.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2488, pruned_loss=0.05208, over 216368.98 frames. ], batch size: 43, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:25:30,441 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126059.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:31,491 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:32,453 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.266e+01 1.545e+02 1.923e+02 2.325e+02 3.929e+02, threshold=3.846e+02, percent-clipped=1.0 +2023-03-27 03:25:37,669 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-27 03:25:42,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:44,664 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:44,744 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7565, 1.6253, 1.4994, 1.8282, 2.1201, 1.8252, 1.3821, 1.4216], + device='cuda:0'), covar=tensor([0.2237, 0.2092, 0.1983, 0.1720, 0.1566, 0.1224, 0.2463, 0.1962], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0212, 0.0196, 0.0243, 0.0189, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:25:45,298 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126081.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:26:03,753 INFO [finetune.py:976] (0/7) Epoch 23, batch 100, loss[loss=0.1684, simple_loss=0.2314, pruned_loss=0.05266, over 4811.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2427, pruned_loss=0.05107, over 381410.10 frames. ], batch size: 25, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:26:14,894 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:26:37,115 INFO [finetune.py:976] (0/7) Epoch 23, batch 150, loss[loss=0.1527, simple_loss=0.2249, pruned_loss=0.04029, over 4874.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2396, pruned_loss=0.05033, over 511207.77 frames. ], batch size: 34, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:26:38,298 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.556e+02 1.791e+02 2.255e+02 5.687e+02, threshold=3.583e+02, percent-clipped=3.0 +2023-03-27 03:26:52,069 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1999, 2.0716, 2.1689, 1.3604, 2.1513, 2.2239, 2.2406, 1.8445], + device='cuda:0'), covar=tensor([0.0547, 0.0604, 0.0625, 0.0892, 0.0617, 0.0676, 0.0512, 0.0970], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0134, 0.0138, 0.0120, 0.0124, 0.0137, 0.0137, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:26:53,944 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2651, 2.1482, 1.9375, 2.3361, 2.8284, 2.2976, 2.1847, 1.7004], + device='cuda:0'), covar=tensor([0.2210, 0.1906, 0.1922, 0.1599, 0.1658, 0.1110, 0.1993, 0.2010], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0213, 0.0197, 0.0244, 0.0189, 0.0216, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:27:10,689 INFO [finetune.py:976] (0/7) Epoch 23, batch 200, loss[loss=0.1672, simple_loss=0.2368, pruned_loss=0.04878, over 4156.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2379, pruned_loss=0.04993, over 606821.93 frames. ], batch size: 65, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:27:45,406 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 03:28:05,167 INFO [finetune.py:976] (0/7) Epoch 23, batch 250, loss[loss=0.1556, simple_loss=0.2383, pruned_loss=0.03646, over 4816.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2436, pruned_loss=0.05141, over 684756.09 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:28:05,278 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126259.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:28:06,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.625e+02 1.849e+02 2.180e+02 4.181e+02, threshold=3.697e+02, percent-clipped=1.0 +2023-03-27 03:28:37,018 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:28:40,579 INFO [finetune.py:976] (0/7) Epoch 23, batch 300, loss[loss=0.1692, simple_loss=0.2403, pruned_loss=0.04902, over 4878.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2467, pruned_loss=0.0518, over 744781.85 frames. ], batch size: 32, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:29:00,155 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.57 vs. limit=5.0 +2023-03-27 03:29:20,734 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126354.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:24,668 INFO [finetune.py:976] (0/7) Epoch 23, batch 350, loss[loss=0.1559, simple_loss=0.2371, pruned_loss=0.03741, over 4138.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2486, pruned_loss=0.05185, over 792086.07 frames. ], batch size: 65, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:29:25,831 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.521e+01 1.514e+02 1.805e+02 2.248e+02 3.946e+02, threshold=3.610e+02, percent-clipped=1.0 +2023-03-27 03:29:39,016 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:39,578 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126381.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:44,529 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 03:29:46,327 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-03-27 03:29:59,416 INFO [finetune.py:976] (0/7) Epoch 23, batch 400, loss[loss=0.1815, simple_loss=0.234, pruned_loss=0.06445, over 4399.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2479, pruned_loss=0.05155, over 827753.31 frames. ], batch size: 19, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:30:21,837 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126429.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:27,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4106, 2.2969, 1.8391, 2.5150, 2.3330, 1.9926, 2.8859, 2.3905], + device='cuda:0'), covar=tensor([0.1295, 0.2379, 0.3122, 0.2715, 0.2494, 0.1633, 0.3567, 0.1769], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0190, 0.0238, 0.0257, 0.0252, 0.0207, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:30:29,739 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126441.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:35,170 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:40,547 INFO [finetune.py:976] (0/7) Epoch 23, batch 450, loss[loss=0.1631, simple_loss=0.2286, pruned_loss=0.04881, over 4713.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2469, pruned_loss=0.05124, over 856582.03 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:30:42,257 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.347e+01 1.414e+02 1.639e+02 2.017e+02 3.767e+02, threshold=3.277e+02, percent-clipped=3.0 +2023-03-27 03:31:09,179 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 03:31:13,829 INFO [finetune.py:976] (0/7) Epoch 23, batch 500, loss[loss=0.1378, simple_loss=0.2138, pruned_loss=0.03091, over 4742.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2447, pruned_loss=0.05078, over 878495.64 frames. ], batch size: 27, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:31:15,653 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:31:42,155 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8035, 1.4066, 1.0831, 1.7384, 2.0429, 1.5969, 1.5963, 1.7729], + device='cuda:0'), covar=tensor([0.1162, 0.1688, 0.1740, 0.0962, 0.1878, 0.1824, 0.1207, 0.1455], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:31:42,805 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7721, 2.5325, 2.0020, 1.2046, 2.2724, 2.2882, 2.0471, 2.2654], + device='cuda:0'), covar=tensor([0.0737, 0.0767, 0.1521, 0.1843, 0.1309, 0.1627, 0.1836, 0.0889], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0191, 0.0199, 0.0182, 0.0209, 0.0208, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:31:47,405 INFO [finetune.py:976] (0/7) Epoch 23, batch 550, loss[loss=0.204, simple_loss=0.2535, pruned_loss=0.07726, over 4828.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2428, pruned_loss=0.0508, over 896250.69 frames. ], batch size: 30, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:31:48,616 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.588e+02 1.845e+02 2.407e+02 4.330e+02, threshold=3.691e+02, percent-clipped=4.0 +2023-03-27 03:32:21,218 INFO [finetune.py:976] (0/7) Epoch 23, batch 600, loss[loss=0.1529, simple_loss=0.2223, pruned_loss=0.04177, over 4821.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2448, pruned_loss=0.05233, over 908878.97 frames. ], batch size: 25, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:32:54,312 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8260, 1.6682, 1.4608, 1.2451, 1.6316, 1.5836, 1.5987, 2.1641], + device='cuda:0'), covar=tensor([0.3569, 0.3695, 0.3019, 0.3425, 0.3310, 0.2294, 0.3205, 0.1667], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0235, 0.0279, 0.0257, 0.0227, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:33:02,502 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:33:05,530 INFO [finetune.py:976] (0/7) Epoch 23, batch 650, loss[loss=0.1462, simple_loss=0.2087, pruned_loss=0.04184, over 4718.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.248, pruned_loss=0.05314, over 920132.66 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:33:06,762 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.579e+02 1.902e+02 2.270e+02 1.001e+03, threshold=3.804e+02, percent-clipped=1.0 +2023-03-27 03:33:06,996 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 03:33:13,314 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 03:33:42,518 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126702.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:33:47,146 INFO [finetune.py:976] (0/7) Epoch 23, batch 700, loss[loss=0.1502, simple_loss=0.2148, pruned_loss=0.0428, over 4719.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2488, pruned_loss=0.05262, over 927799.36 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:34:11,540 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126736.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:34:29,141 INFO [finetune.py:976] (0/7) Epoch 23, batch 750, loss[loss=0.1877, simple_loss=0.2678, pruned_loss=0.05382, over 4898.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2497, pruned_loss=0.05264, over 933928.03 frames. ], batch size: 36, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:34:30,819 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.063e+01 1.442e+02 1.710e+02 2.057e+02 3.398e+02, threshold=3.419e+02, percent-clipped=0.0 +2023-03-27 03:34:43,537 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 03:34:44,558 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6823, 1.5774, 2.0589, 3.2631, 2.1595, 2.3569, 0.9569, 2.6991], + device='cuda:0'), covar=tensor([0.1710, 0.1446, 0.1281, 0.0520, 0.0776, 0.1262, 0.1850, 0.0488], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 03:34:56,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7648, 1.5391, 2.1069, 3.4347, 2.2280, 2.5135, 1.3463, 2.8072], + device='cuda:0'), covar=tensor([0.1744, 0.1460, 0.1404, 0.0598, 0.0849, 0.1281, 0.1675, 0.0569], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 03:35:00,639 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:35:02,394 INFO [finetune.py:976] (0/7) Epoch 23, batch 800, loss[loss=0.1275, simple_loss=0.2031, pruned_loss=0.02601, over 4787.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2473, pruned_loss=0.05131, over 936349.91 frames. ], batch size: 28, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:35:07,273 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126816.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:35:44,514 INFO [finetune.py:976] (0/7) Epoch 23, batch 850, loss[loss=0.1599, simple_loss=0.2268, pruned_loss=0.04644, over 4862.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2455, pruned_loss=0.05085, over 940319.16 frames. ], batch size: 34, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:35:45,682 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.329e+01 1.478e+02 1.768e+02 2.024e+02 3.574e+02, threshold=3.536e+02, percent-clipped=1.0 +2023-03-27 03:35:56,141 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126877.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:36:18,424 INFO [finetune.py:976] (0/7) Epoch 23, batch 900, loss[loss=0.1824, simple_loss=0.2377, pruned_loss=0.06359, over 4706.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2427, pruned_loss=0.05036, over 943049.23 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:36:52,033 INFO [finetune.py:976] (0/7) Epoch 23, batch 950, loss[loss=0.1241, simple_loss=0.1984, pruned_loss=0.02488, over 4722.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2407, pruned_loss=0.04996, over 946420.92 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:36:53,225 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.486e+02 1.775e+02 2.132e+02 3.351e+02, threshold=3.551e+02, percent-clipped=0.0 +2023-03-27 03:37:26,096 INFO [finetune.py:976] (0/7) Epoch 23, batch 1000, loss[loss=0.183, simple_loss=0.2522, pruned_loss=0.05693, over 4930.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2431, pruned_loss=0.05073, over 947606.23 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:37:43,503 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127036.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:37:51,209 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8302, 1.8920, 1.6112, 2.0533, 2.4167, 1.9630, 1.8214, 1.4658], + device='cuda:0'), covar=tensor([0.2086, 0.1852, 0.1744, 0.1515, 0.1617, 0.1155, 0.2090, 0.1889], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0198, 0.0245, 0.0190, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:38:01,448 INFO [finetune.py:976] (0/7) Epoch 23, batch 1050, loss[loss=0.174, simple_loss=0.2465, pruned_loss=0.05077, over 4818.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2466, pruned_loss=0.05183, over 947148.49 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:38:02,652 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.597e+02 1.830e+02 2.248e+02 5.450e+02, threshold=3.660e+02, percent-clipped=4.0 +2023-03-27 03:38:27,252 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:38:42,637 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127106.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:38:44,898 INFO [finetune.py:976] (0/7) Epoch 23, batch 1100, loss[loss=0.207, simple_loss=0.2778, pruned_loss=0.06812, over 4799.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2471, pruned_loss=0.05139, over 948588.26 frames. ], batch size: 45, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:39:08,663 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 03:39:14,542 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127154.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:39:21,916 INFO [finetune.py:976] (0/7) Epoch 23, batch 1150, loss[loss=0.1935, simple_loss=0.2605, pruned_loss=0.06326, over 4919.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2472, pruned_loss=0.05096, over 949617.67 frames. ], batch size: 42, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:39:23,617 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.547e+02 1.781e+02 2.032e+02 3.739e+02, threshold=3.562e+02, percent-clipped=1.0 +2023-03-27 03:39:33,498 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.30 vs. limit=5.0 +2023-03-27 03:39:35,834 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:40:04,152 INFO [finetune.py:976] (0/7) Epoch 23, batch 1200, loss[loss=0.178, simple_loss=0.2489, pruned_loss=0.05356, over 4887.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2461, pruned_loss=0.05099, over 950817.68 frames. ], batch size: 32, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:40:10,137 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7967, 1.5616, 1.4743, 0.8342, 1.6456, 1.8520, 1.7918, 1.3923], + device='cuda:0'), covar=tensor([0.0768, 0.0592, 0.0545, 0.0509, 0.0490, 0.0479, 0.0345, 0.0618], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0148, 0.0127, 0.0122, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9516e-05, 1.0709e-04, 9.0898e-05, 8.5929e-05, 9.1560e-05, 9.1630e-05, + 1.0076e-04, 1.0597e-04], device='cuda:0') +2023-03-27 03:40:33,000 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127249.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:40:42,285 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8052, 1.3043, 0.8800, 1.6140, 2.1879, 1.3806, 1.5355, 1.5814], + device='cuda:0'), covar=tensor([0.1452, 0.2170, 0.2089, 0.1260, 0.1821, 0.1865, 0.1534, 0.2086], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:40:44,043 INFO [finetune.py:976] (0/7) Epoch 23, batch 1250, loss[loss=0.1643, simple_loss=0.2355, pruned_loss=0.04654, over 4813.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2445, pruned_loss=0.05097, over 951402.29 frames. ], batch size: 39, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:40:45,212 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.489e+02 1.742e+02 2.248e+02 3.707e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 03:41:21,264 INFO [finetune.py:976] (0/7) Epoch 23, batch 1300, loss[loss=0.1653, simple_loss=0.2368, pruned_loss=0.04692, over 4823.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2421, pruned_loss=0.05039, over 953546.41 frames. ], batch size: 39, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:41:22,019 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127310.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:41:23,216 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5373, 1.4570, 1.4565, 1.4815, 1.0519, 2.8147, 1.1334, 1.5145], + device='cuda:0'), covar=tensor([0.2958, 0.2317, 0.1909, 0.2164, 0.1667, 0.0327, 0.2630, 0.1160], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0116, 0.0120, 0.0123, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:41:54,394 INFO [finetune.py:976] (0/7) Epoch 23, batch 1350, loss[loss=0.1691, simple_loss=0.2461, pruned_loss=0.04608, over 4907.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2418, pruned_loss=0.05, over 954450.74 frames. ], batch size: 43, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:41:55,605 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.500e+02 1.797e+02 2.250e+02 4.549e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-03-27 03:41:56,847 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7728, 1.7321, 1.7091, 1.7751, 1.3743, 3.5961, 1.4935, 2.0462], + device='cuda:0'), covar=tensor([0.3122, 0.2466, 0.2039, 0.2298, 0.1688, 0.0227, 0.2446, 0.1148], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:42:27,768 INFO [finetune.py:976] (0/7) Epoch 23, batch 1400, loss[loss=0.1645, simple_loss=0.2485, pruned_loss=0.04028, over 4805.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2429, pruned_loss=0.05018, over 952059.14 frames. ], batch size: 51, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:42:28,496 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5651, 1.6584, 1.6560, 1.0173, 1.8632, 1.9850, 2.0186, 1.5071], + device='cuda:0'), covar=tensor([0.0915, 0.0590, 0.0537, 0.0476, 0.0389, 0.0658, 0.0319, 0.0685], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0149, 0.0128, 0.0122, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.0077e-05, 1.0739e-04, 9.1468e-05, 8.6158e-05, 9.1887e-05, 9.2330e-05, + 1.0113e-04, 1.0634e-04], device='cuda:0') +2023-03-27 03:42:40,734 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127427.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:43:01,001 INFO [finetune.py:976] (0/7) Epoch 23, batch 1450, loss[loss=0.1535, simple_loss=0.2314, pruned_loss=0.03774, over 4770.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2451, pruned_loss=0.05072, over 952944.88 frames. ], batch size: 28, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:43:03,312 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.817e+01 1.596e+02 1.913e+02 2.194e+02 3.811e+02, threshold=3.827e+02, percent-clipped=1.0 +2023-03-27 03:43:09,867 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:43:25,060 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127488.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:43:26,247 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5499, 3.7561, 3.5302, 1.3860, 3.7531, 2.8699, 0.8570, 2.6466], + device='cuda:0'), covar=tensor([0.2297, 0.2247, 0.1495, 0.3528, 0.1103, 0.0946, 0.4381, 0.1322], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0160, 0.0129, 0.0161, 0.0124, 0.0150, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 03:43:46,709 INFO [finetune.py:976] (0/7) Epoch 23, batch 1500, loss[loss=0.1594, simple_loss=0.2409, pruned_loss=0.03896, over 4902.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2466, pruned_loss=0.05132, over 954340.23 frames. ], batch size: 37, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:43:53,894 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127520.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:43:59,933 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:44:20,425 INFO [finetune.py:976] (0/7) Epoch 23, batch 1550, loss[loss=0.1444, simple_loss=0.2329, pruned_loss=0.02793, over 4756.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2464, pruned_loss=0.05106, over 953801.03 frames. ], batch size: 28, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:44:22,241 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.462e+02 1.755e+02 2.123e+02 3.197e+02, threshold=3.511e+02, percent-clipped=0.0 +2023-03-27 03:44:48,018 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127590.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:45:04,796 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:45:07,683 INFO [finetune.py:976] (0/7) Epoch 23, batch 1600, loss[loss=0.1607, simple_loss=0.2356, pruned_loss=0.04293, over 4890.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2448, pruned_loss=0.05037, over 954845.81 frames. ], batch size: 32, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:45:34,515 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-27 03:45:40,959 INFO [finetune.py:976] (0/7) Epoch 23, batch 1650, loss[loss=0.1456, simple_loss=0.2188, pruned_loss=0.03623, over 4850.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2426, pruned_loss=0.05012, over 955376.08 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:45:43,334 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.567e+02 1.812e+02 2.212e+02 4.212e+02, threshold=3.624e+02, percent-clipped=4.0 +2023-03-27 03:45:43,477 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127662.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:46:26,409 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.5034, 3.1482, 3.3127, 3.1927, 3.1407, 3.0452, 3.5908, 1.5509], + device='cuda:0'), covar=tensor([0.1492, 0.1913, 0.1822, 0.2079, 0.2387, 0.2644, 0.1867, 0.6713], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0247, 0.0281, 0.0293, 0.0336, 0.0287, 0.0306, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:46:26,934 INFO [finetune.py:976] (0/7) Epoch 23, batch 1700, loss[loss=0.2768, simple_loss=0.3083, pruned_loss=0.1226, over 4071.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2418, pruned_loss=0.05056, over 954961.29 frames. ], batch size: 65, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:46:36,935 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127723.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:47:00,092 INFO [finetune.py:976] (0/7) Epoch 23, batch 1750, loss[loss=0.1804, simple_loss=0.2697, pruned_loss=0.0455, over 4815.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2437, pruned_loss=0.05105, over 953504.39 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:47:01,899 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.511e+02 1.783e+02 2.157e+02 3.427e+02, threshold=3.565e+02, percent-clipped=0.0 +2023-03-27 03:47:06,844 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9393, 1.9273, 1.6369, 2.0811, 2.5879, 2.0247, 1.8511, 1.4953], + device='cuda:0'), covar=tensor([0.2175, 0.1926, 0.2010, 0.1646, 0.1724, 0.1259, 0.2239, 0.1925], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0213, 0.0196, 0.0244, 0.0189, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:47:12,610 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8406, 4.1542, 4.4568, 4.6438, 4.5746, 4.3086, 4.9368, 1.5669], + device='cuda:0'), covar=tensor([0.0736, 0.0862, 0.0745, 0.0855, 0.1159, 0.1696, 0.0538, 0.6150], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0247, 0.0281, 0.0294, 0.0337, 0.0288, 0.0306, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:47:13,902 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:47:16,795 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127783.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:47:22,115 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2325, 1.4070, 1.1041, 1.2718, 1.6211, 1.5481, 1.3413, 1.2770], + device='cuda:0'), covar=tensor([0.0425, 0.0344, 0.0605, 0.0342, 0.0203, 0.0505, 0.0352, 0.0444], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0105, 0.0143, 0.0111, 0.0098, 0.0111, 0.0100, 0.0111], + device='cuda:0'), out_proj_covar=tensor([7.6405e-05, 8.0848e-05, 1.1179e-04, 8.4827e-05, 7.6365e-05, 8.1909e-05, + 7.4560e-05, 8.4357e-05], device='cuda:0') +2023-03-27 03:47:33,985 INFO [finetune.py:976] (0/7) Epoch 23, batch 1800, loss[loss=0.1457, simple_loss=0.2092, pruned_loss=0.04116, over 4719.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2454, pruned_loss=0.05142, over 954388.94 frames. ], batch size: 23, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:47:54,748 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127840.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:01,839 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:07,744 INFO [finetune.py:976] (0/7) Epoch 23, batch 1850, loss[loss=0.1712, simple_loss=0.2501, pruned_loss=0.04614, over 4816.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2474, pruned_loss=0.05177, over 953744.83 frames. ], batch size: 39, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:48:09,575 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.256e+01 1.480e+02 1.674e+02 2.095e+02 4.093e+02, threshold=3.347e+02, percent-clipped=1.0 +2023-03-27 03:48:12,658 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.7919, 3.3501, 3.5242, 3.6575, 3.5650, 3.3426, 3.8734, 1.2432], + device='cuda:0'), covar=tensor([0.0939, 0.0971, 0.1024, 0.1132, 0.1373, 0.1658, 0.0881, 0.5423], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0248, 0.0283, 0.0296, 0.0339, 0.0290, 0.0307, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:48:16,150 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:24,924 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127885.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:40,218 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127905.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:42,623 INFO [finetune.py:976] (0/7) Epoch 23, batch 1900, loss[loss=0.1926, simple_loss=0.273, pruned_loss=0.05613, over 4888.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2492, pruned_loss=0.05199, over 955017.92 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:48:43,941 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:58,804 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:49:11,956 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127953.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:49:16,014 INFO [finetune.py:976] (0/7) Epoch 23, batch 1950, loss[loss=0.1744, simple_loss=0.2409, pruned_loss=0.05391, over 4843.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2479, pruned_loss=0.05168, over 953181.38 frames. ], batch size: 44, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:49:17,847 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.391e+02 1.764e+02 2.084e+02 3.552e+02, threshold=3.528e+02, percent-clipped=1.0 +2023-03-27 03:49:43,486 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-128000.pt +2023-03-27 03:49:51,363 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9238, 4.6638, 4.3918, 2.4923, 4.7801, 3.6052, 0.8877, 3.3700], + device='cuda:0'), covar=tensor([0.2368, 0.1704, 0.1334, 0.2962, 0.0882, 0.0860, 0.4725, 0.1374], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0128, 0.0161, 0.0124, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 03:49:53,065 INFO [finetune.py:976] (0/7) Epoch 23, batch 2000, loss[loss=0.1591, simple_loss=0.2176, pruned_loss=0.05027, over 4805.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2451, pruned_loss=0.05071, over 955282.28 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:50:03,203 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128018.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:50:19,609 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5447, 1.6567, 0.7260, 2.2996, 2.6439, 1.9193, 2.1255, 1.9079], + device='cuda:0'), covar=tensor([0.1266, 0.2061, 0.2181, 0.1041, 0.1746, 0.1706, 0.1411, 0.2069], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:50:38,969 INFO [finetune.py:976] (0/7) Epoch 23, batch 2050, loss[loss=0.1619, simple_loss=0.2298, pruned_loss=0.04695, over 4823.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2419, pruned_loss=0.05003, over 955631.96 frames. ], batch size: 33, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:50:41,271 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.483e+02 1.749e+02 2.101e+02 3.191e+02, threshold=3.498e+02, percent-clipped=0.0 +2023-03-27 03:50:54,685 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128083.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:51:13,972 INFO [finetune.py:976] (0/7) Epoch 23, batch 2100, loss[loss=0.131, simple_loss=0.2156, pruned_loss=0.02319, over 4818.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2405, pruned_loss=0.04924, over 956171.19 frames. ], batch size: 51, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:51:40,712 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128131.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:51:43,601 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128135.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:00,526 INFO [finetune.py:976] (0/7) Epoch 23, batch 2150, loss[loss=0.2026, simple_loss=0.2541, pruned_loss=0.0755, over 4895.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2448, pruned_loss=0.05088, over 954439.53 frames. ], batch size: 32, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:52:02,382 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.558e+02 1.811e+02 2.178e+02 3.611e+02, threshold=3.622e+02, percent-clipped=2.0 +2023-03-27 03:52:17,013 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128185.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:31,847 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128206.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:33,602 INFO [finetune.py:976] (0/7) Epoch 23, batch 2200, loss[loss=0.1905, simple_loss=0.2633, pruned_loss=0.05887, over 4819.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2484, pruned_loss=0.05234, over 951938.72 frames. ], batch size: 30, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:52:46,630 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:47,949 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9180, 1.9299, 1.5693, 1.7472, 1.7504, 1.7326, 1.8332, 2.4797], + device='cuda:0'), covar=tensor([0.3955, 0.3853, 0.3236, 0.3599, 0.4034, 0.2282, 0.3581, 0.1576], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0235, 0.0275, 0.0256, 0.0227, 0.0254, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:52:49,634 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:53:07,264 INFO [finetune.py:976] (0/7) Epoch 23, batch 2250, loss[loss=0.149, simple_loss=0.2131, pruned_loss=0.0425, over 4027.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2484, pruned_loss=0.05224, over 951695.85 frames. ], batch size: 16, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:53:09,087 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.465e+01 1.491e+02 1.772e+02 2.217e+02 3.841e+02, threshold=3.544e+02, percent-clipped=1.0 +2023-03-27 03:53:28,171 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8248, 1.7678, 1.7350, 1.7763, 1.5058, 4.5150, 1.7008, 2.0806], + device='cuda:0'), covar=tensor([0.3190, 0.2474, 0.2012, 0.2195, 0.1523, 0.0107, 0.2451, 0.1219], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0114, 0.0096, 0.0095, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:53:40,845 INFO [finetune.py:976] (0/7) Epoch 23, batch 2300, loss[loss=0.1774, simple_loss=0.2523, pruned_loss=0.05124, over 4882.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.249, pruned_loss=0.0524, over 954601.16 frames. ], batch size: 43, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:53:47,297 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128318.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:53:47,918 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7884, 1.6817, 1.4604, 1.6924, 2.0878, 2.0248, 1.6721, 1.5166], + device='cuda:0'), covar=tensor([0.0331, 0.0373, 0.0614, 0.0357, 0.0233, 0.0503, 0.0403, 0.0482], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0106, 0.0144, 0.0111, 0.0100, 0.0111, 0.0101, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7315e-05, 8.1618e-05, 1.1282e-04, 8.5458e-05, 7.7366e-05, 8.1967e-05, + 7.5318e-05, 8.5303e-05], device='cuda:0') +2023-03-27 03:53:48,502 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128320.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:13,558 INFO [finetune.py:976] (0/7) Epoch 23, batch 2350, loss[loss=0.1396, simple_loss=0.2119, pruned_loss=0.03364, over 4258.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2471, pruned_loss=0.05223, over 952836.16 frames. ], batch size: 65, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:54:15,920 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.510e+02 1.720e+02 2.103e+02 3.385e+02, threshold=3.440e+02, percent-clipped=0.0 +2023-03-27 03:54:18,448 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128366.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:28,982 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128381.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:46,896 INFO [finetune.py:976] (0/7) Epoch 23, batch 2400, loss[loss=0.2058, simple_loss=0.2691, pruned_loss=0.07125, over 4929.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.245, pruned_loss=0.05194, over 953265.45 frames. ], batch size: 38, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:54:49,471 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128413.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:03,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4750, 1.2258, 1.9602, 3.0784, 1.9181, 2.3966, 1.0096, 2.6914], + device='cuda:0'), covar=tensor([0.2275, 0.2290, 0.1608, 0.0980, 0.1135, 0.1359, 0.2140, 0.0718], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0164, 0.0100, 0.0137, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 03:55:06,798 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128435.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:17,496 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2647, 1.3941, 1.5518, 1.4948, 1.5606, 2.9568, 1.2658, 1.4831], + device='cuda:0'), covar=tensor([0.1073, 0.1750, 0.1158, 0.1032, 0.1666, 0.0321, 0.1542, 0.1869], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0076, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:55:35,546 INFO [finetune.py:976] (0/7) Epoch 23, batch 2450, loss[loss=0.1394, simple_loss=0.2136, pruned_loss=0.0326, over 4923.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.242, pruned_loss=0.05091, over 956704.69 frames. ], batch size: 37, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:55:35,669 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7157, 1.5797, 1.9954, 1.2138, 1.6854, 1.9665, 1.5260, 2.0789], + device='cuda:0'), covar=tensor([0.1146, 0.2235, 0.1325, 0.1793, 0.1031, 0.1301, 0.2749, 0.0822], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0191, 0.0187, 0.0172, 0.0212, 0.0214, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:55:41,395 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.054e+01 1.492e+02 1.779e+02 2.209e+02 4.084e+02, threshold=3.557e+02, percent-clipped=2.0 +2023-03-27 03:55:49,826 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:55,742 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:10,248 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:12,479 INFO [finetune.py:976] (0/7) Epoch 23, batch 2500, loss[loss=0.1542, simple_loss=0.2343, pruned_loss=0.03705, over 4792.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2433, pruned_loss=0.05118, over 957472.46 frames. ], batch size: 29, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:56:24,785 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4973, 1.5117, 1.2453, 1.4980, 1.9226, 1.7338, 1.5227, 1.3390], + device='cuda:0'), covar=tensor([0.0341, 0.0329, 0.0664, 0.0344, 0.0222, 0.0515, 0.0322, 0.0424], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0101, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7326e-05, 8.1845e-05, 1.1291e-04, 8.5483e-05, 7.7395e-05, 8.1959e-05, + 7.5416e-05, 8.5430e-05], device='cuda:0') +2023-03-27 03:56:25,580 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-27 03:56:25,957 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:28,300 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9611, 1.3040, 0.9455, 1.9082, 2.2318, 1.7319, 1.6228, 1.5867], + device='cuda:0'), covar=tensor([0.1525, 0.2143, 0.1963, 0.1177, 0.1954, 0.2038, 0.1467, 0.2174], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0120, 0.0095, 0.0100, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:56:33,675 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5516, 1.5020, 1.4983, 1.5670, 1.2075, 3.5445, 1.4826, 1.8652], + device='cuda:0'), covar=tensor([0.3322, 0.2683, 0.2187, 0.2300, 0.1769, 0.0184, 0.2551, 0.1255], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0095, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 03:56:48,370 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128554.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:55,781 INFO [finetune.py:976] (0/7) Epoch 23, batch 2550, loss[loss=0.2214, simple_loss=0.298, pruned_loss=0.0724, over 4873.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2474, pruned_loss=0.05235, over 957417.89 frames. ], batch size: 34, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:56:58,589 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.563e+02 1.837e+02 2.202e+02 4.665e+02, threshold=3.674e+02, percent-clipped=3.0 +2023-03-27 03:57:11,627 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:57:19,462 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0386, 1.8552, 1.6978, 1.7166, 1.8232, 1.7882, 1.8331, 2.5401], + device='cuda:0'), covar=tensor([0.3842, 0.4448, 0.3342, 0.3505, 0.3936, 0.2497, 0.3661, 0.1642], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0262, 0.0235, 0.0276, 0.0255, 0.0227, 0.0254, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:57:33,477 INFO [finetune.py:976] (0/7) Epoch 23, batch 2600, loss[loss=0.2423, simple_loss=0.2914, pruned_loss=0.09658, over 4912.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.248, pruned_loss=0.05252, over 956563.58 frames. ], batch size: 36, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:57:48,176 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1163, 1.3678, 0.6624, 2.0765, 2.4315, 1.6844, 1.6901, 1.7793], + device='cuda:0'), covar=tensor([0.1403, 0.2125, 0.2202, 0.1099, 0.1794, 0.1989, 0.1462, 0.2023], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0120, 0.0095, 0.0100, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 03:57:49,258 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.0444, 4.4202, 4.5835, 4.9196, 4.7532, 4.5627, 5.1509, 1.5829], + device='cuda:0'), covar=tensor([0.0759, 0.0787, 0.0720, 0.0896, 0.1299, 0.1640, 0.0585, 0.6074], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0245, 0.0279, 0.0290, 0.0336, 0.0284, 0.0301, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:57:49,890 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4885, 1.3640, 1.5277, 0.7976, 1.5766, 1.5717, 1.4589, 1.3200], + device='cuda:0'), covar=tensor([0.0609, 0.0779, 0.0662, 0.0954, 0.0829, 0.0668, 0.0607, 0.1192], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0134, 0.0137, 0.0118, 0.0123, 0.0135, 0.0136, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 03:57:55,266 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-27 03:58:07,053 INFO [finetune.py:976] (0/7) Epoch 23, batch 2650, loss[loss=0.1713, simple_loss=0.2439, pruned_loss=0.04929, over 4822.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2488, pruned_loss=0.05221, over 956586.74 frames. ], batch size: 39, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:58:08,890 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.579e+02 1.820e+02 2.183e+02 3.562e+02, threshold=3.640e+02, percent-clipped=0.0 +2023-03-27 03:58:18,908 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128676.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:58:40,903 INFO [finetune.py:976] (0/7) Epoch 23, batch 2700, loss[loss=0.1753, simple_loss=0.245, pruned_loss=0.05276, over 4822.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2461, pruned_loss=0.05085, over 954387.73 frames. ], batch size: 41, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:59:14,674 INFO [finetune.py:976] (0/7) Epoch 23, batch 2750, loss[loss=0.1861, simple_loss=0.2508, pruned_loss=0.06076, over 4707.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2436, pruned_loss=0.04995, over 954538.94 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:59:16,466 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.314e+01 1.516e+02 1.813e+02 2.146e+02 3.615e+02, threshold=3.627e+02, percent-clipped=0.0 +2023-03-27 03:59:21,318 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128769.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:59:29,516 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:59:48,512 INFO [finetune.py:976] (0/7) Epoch 23, batch 2800, loss[loss=0.1717, simple_loss=0.2388, pruned_loss=0.05228, over 4701.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2419, pruned_loss=0.05016, over 955722.41 frames. ], batch size: 23, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 04:00:00,997 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7921, 1.6924, 1.6916, 1.6508, 1.1565, 3.0372, 1.3111, 1.8012], + device='cuda:0'), covar=tensor([0.2966, 0.2397, 0.1966, 0.2345, 0.1774, 0.0290, 0.2359, 0.1145], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0095, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:00:10,898 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128842.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:00:22,070 INFO [finetune.py:976] (0/7) Epoch 23, batch 2850, loss[loss=0.1602, simple_loss=0.2252, pruned_loss=0.04758, over 4762.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2413, pruned_loss=0.05045, over 955669.31 frames. ], batch size: 26, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 04:00:23,887 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.021e+01 1.432e+02 1.754e+02 2.169e+02 4.729e+02, threshold=3.508e+02, percent-clipped=1.0 +2023-03-27 04:00:41,214 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8778, 4.5114, 4.2463, 2.3069, 4.6176, 3.5397, 0.8964, 3.1186], + device='cuda:0'), covar=tensor([0.2406, 0.1912, 0.1437, 0.3130, 0.0829, 0.0789, 0.4448, 0.1391], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0160, 0.0129, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:00:53,960 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8374, 1.9997, 1.5982, 1.6993, 2.4161, 2.3057, 2.0026, 1.8842], + device='cuda:0'), covar=tensor([0.0383, 0.0343, 0.0592, 0.0335, 0.0235, 0.0638, 0.0383, 0.0423], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0101, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7279e-05, 8.1839e-05, 1.1321e-04, 8.5216e-05, 7.7363e-05, 8.2099e-05, + 7.5528e-05, 8.5189e-05], device='cuda:0') +2023-03-27 04:00:55,200 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128890.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:01:08,141 INFO [finetune.py:976] (0/7) Epoch 23, batch 2900, loss[loss=0.1903, simple_loss=0.2735, pruned_loss=0.0535, over 4923.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2438, pruned_loss=0.05126, over 955582.86 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:01:17,949 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8599, 1.5692, 1.4995, 1.2836, 1.6439, 1.6489, 1.6028, 2.1996], + device='cuda:0'), covar=tensor([0.3945, 0.4084, 0.3244, 0.3698, 0.3889, 0.2491, 0.3563, 0.1909], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0234, 0.0276, 0.0255, 0.0227, 0.0254, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:01:36,747 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128951.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:01:41,988 INFO [finetune.py:976] (0/7) Epoch 23, batch 2950, loss[loss=0.1498, simple_loss=0.2127, pruned_loss=0.04348, over 4705.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2459, pruned_loss=0.05144, over 955009.08 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:01:43,785 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.806e+01 1.599e+02 2.012e+02 2.314e+02 4.261e+02, threshold=4.024e+02, percent-clipped=1.0 +2023-03-27 04:01:52,354 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:02:31,486 INFO [finetune.py:976] (0/7) Epoch 23, batch 3000, loss[loss=0.1806, simple_loss=0.2579, pruned_loss=0.05165, over 4834.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2487, pruned_loss=0.05257, over 955559.82 frames. ], batch size: 49, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:02:31,488 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 04:02:42,305 INFO [finetune.py:1010] (0/7) Epoch 23, validation: loss=0.1567, simple_loss=0.225, pruned_loss=0.04424, over 2265189.00 frames. +2023-03-27 04:02:42,305 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 04:02:51,935 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129024.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:03:12,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9109, 1.6448, 2.2499, 1.4492, 2.0258, 2.1290, 1.5777, 2.2988], + device='cuda:0'), covar=tensor([0.1179, 0.1931, 0.1294, 0.1877, 0.0814, 0.1300, 0.2702, 0.0725], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0206, 0.0192, 0.0189, 0.0173, 0.0214, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:03:14,563 INFO [finetune.py:976] (0/7) Epoch 23, batch 3050, loss[loss=0.159, simple_loss=0.2291, pruned_loss=0.04439, over 4801.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2499, pruned_loss=0.05317, over 954986.04 frames. ], batch size: 25, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:03:16,828 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.606e+02 1.899e+02 2.236e+02 5.313e+02, threshold=3.798e+02, percent-clipped=3.0 +2023-03-27 04:03:22,096 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:03:46,318 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 04:03:47,837 INFO [finetune.py:976] (0/7) Epoch 23, batch 3100, loss[loss=0.2163, simple_loss=0.2725, pruned_loss=0.08006, over 4907.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2467, pruned_loss=0.0517, over 954976.01 frames. ], batch size: 32, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:03:53,277 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:04:06,393 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129137.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:04:16,570 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7261, 1.5575, 2.0238, 3.2003, 2.1992, 2.4375, 1.0825, 2.6341], + device='cuda:0'), covar=tensor([0.1651, 0.1360, 0.1281, 0.0579, 0.0731, 0.1186, 0.1735, 0.0498], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0165, 0.0101, 0.0138, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 04:04:20,611 INFO [finetune.py:976] (0/7) Epoch 23, batch 3150, loss[loss=0.1448, simple_loss=0.2222, pruned_loss=0.03371, over 4911.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2445, pruned_loss=0.05138, over 954747.21 frames. ], batch size: 35, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:04:22,455 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.503e+02 1.751e+02 2.296e+02 3.694e+02, threshold=3.502e+02, percent-clipped=0.0 +2023-03-27 04:05:01,880 INFO [finetune.py:976] (0/7) Epoch 23, batch 3200, loss[loss=0.157, simple_loss=0.2228, pruned_loss=0.04558, over 4872.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2409, pruned_loss=0.04982, over 957461.36 frames. ], batch size: 31, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:05:06,632 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.8339, 4.2615, 4.4432, 4.6451, 4.5755, 4.3961, 4.9581, 1.6627], + device='cuda:0'), covar=tensor([0.0849, 0.0829, 0.0736, 0.1151, 0.1317, 0.1474, 0.0610, 0.5635], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0245, 0.0278, 0.0290, 0.0335, 0.0284, 0.0303, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:05:26,810 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129246.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:05:37,440 INFO [finetune.py:976] (0/7) Epoch 23, batch 3250, loss[loss=0.253, simple_loss=0.3046, pruned_loss=0.1007, over 4815.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2416, pruned_loss=0.05039, over 955879.24 frames. ], batch size: 41, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:05:39,765 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.902e+01 1.487e+02 1.735e+02 2.015e+02 4.622e+02, threshold=3.470e+02, percent-clipped=1.0 +2023-03-27 04:06:22,328 INFO [finetune.py:976] (0/7) Epoch 23, batch 3300, loss[loss=0.1828, simple_loss=0.2675, pruned_loss=0.04902, over 4868.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2467, pruned_loss=0.0524, over 954162.39 frames. ], batch size: 34, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:06:52,072 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6632, 1.5385, 1.0887, 0.2827, 1.2850, 1.4733, 1.4623, 1.4694], + device='cuda:0'), covar=tensor([0.0982, 0.0817, 0.1486, 0.2039, 0.1456, 0.2546, 0.2323, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0183, 0.0212, 0.0210, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:06:56,062 INFO [finetune.py:976] (0/7) Epoch 23, batch 3350, loss[loss=0.1834, simple_loss=0.2592, pruned_loss=0.05377, over 4895.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2484, pruned_loss=0.05267, over 954558.63 frames. ], batch size: 37, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:06:57,831 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.264e+01 1.616e+02 1.807e+02 2.144e+02 4.365e+02, threshold=3.613e+02, percent-clipped=1.0 +2023-03-27 04:06:59,781 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4933, 1.4425, 1.7183, 1.6474, 1.6112, 3.2884, 1.3351, 1.4463], + device='cuda:0'), covar=tensor([0.1118, 0.2141, 0.1282, 0.1177, 0.1911, 0.0334, 0.1799, 0.2257], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0082, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:07:23,441 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7720, 3.9288, 3.6451, 1.9396, 4.0406, 3.1787, 1.0314, 2.6303], + device='cuda:0'), covar=tensor([0.2211, 0.1689, 0.1472, 0.3244, 0.0920, 0.0925, 0.4382, 0.1465], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0159, 0.0128, 0.0160, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:07:47,796 INFO [finetune.py:976] (0/7) Epoch 23, batch 3400, loss[loss=0.2373, simple_loss=0.304, pruned_loss=0.08531, over 4254.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2496, pruned_loss=0.05333, over 955712.92 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:07:50,373 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129413.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:07:52,240 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-27 04:08:07,354 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:08:21,221 INFO [finetune.py:976] (0/7) Epoch 23, batch 3450, loss[loss=0.1474, simple_loss=0.2143, pruned_loss=0.04024, over 4401.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2497, pruned_loss=0.05285, over 956585.58 frames. ], batch size: 19, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:08:23,471 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.549e+02 1.981e+02 2.372e+02 4.494e+02, threshold=3.962e+02, percent-clipped=6.0 +2023-03-27 04:08:31,809 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129474.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:08:39,421 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129485.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:08:53,377 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 04:08:54,911 INFO [finetune.py:976] (0/7) Epoch 23, batch 3500, loss[loss=0.1435, simple_loss=0.2119, pruned_loss=0.03753, over 4735.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2474, pruned_loss=0.05229, over 956502.13 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:09:15,742 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0109, 1.8588, 1.6523, 1.6832, 1.7780, 1.7576, 1.8326, 2.4889], + device='cuda:0'), covar=tensor([0.3204, 0.3469, 0.2751, 0.3237, 0.3487, 0.2183, 0.3038, 0.1485], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0263, 0.0235, 0.0277, 0.0256, 0.0228, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:09:20,949 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129546.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:09:21,696 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-27 04:09:28,784 INFO [finetune.py:976] (0/7) Epoch 23, batch 3550, loss[loss=0.1517, simple_loss=0.224, pruned_loss=0.0397, over 4912.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2447, pruned_loss=0.05143, over 957364.29 frames. ], batch size: 43, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:09:30,139 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-03-27 04:09:30,571 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.427e+01 1.468e+02 1.701e+02 2.000e+02 4.470e+02, threshold=3.402e+02, percent-clipped=1.0 +2023-03-27 04:09:45,524 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 04:09:52,428 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129594.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:10:11,047 INFO [finetune.py:976] (0/7) Epoch 23, batch 3600, loss[loss=0.1729, simple_loss=0.2478, pruned_loss=0.04902, over 4870.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2427, pruned_loss=0.05053, over 957765.46 frames. ], batch size: 34, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:10:12,954 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129612.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:10:44,782 INFO [finetune.py:976] (0/7) Epoch 23, batch 3650, loss[loss=0.1416, simple_loss=0.2145, pruned_loss=0.03436, over 4794.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2437, pruned_loss=0.05075, over 956816.98 frames. ], batch size: 25, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:10:46,575 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.535e+02 1.802e+02 2.202e+02 3.404e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-27 04:10:49,075 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6087, 3.6363, 3.4670, 1.5837, 3.7653, 2.9463, 1.1438, 2.4200], + device='cuda:0'), covar=tensor([0.2448, 0.2324, 0.1454, 0.3507, 0.1061, 0.0900, 0.4058, 0.1580], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0160, 0.0129, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:10:52,238 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0892, 1.9629, 1.7049, 1.7785, 1.8438, 1.8549, 1.8936, 2.6346], + device='cuda:0'), covar=tensor([0.3600, 0.3862, 0.3230, 0.3801, 0.3880, 0.2316, 0.3629, 0.1531], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0261, 0.0234, 0.0274, 0.0254, 0.0226, 0.0252, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:10:53,481 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129673.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:11:23,356 INFO [finetune.py:976] (0/7) Epoch 23, batch 3700, loss[loss=0.2025, simple_loss=0.2814, pruned_loss=0.06178, over 4809.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2476, pruned_loss=0.05227, over 955130.60 frames. ], batch size: 51, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:12:00,260 INFO [finetune.py:976] (0/7) Epoch 23, batch 3750, loss[loss=0.1849, simple_loss=0.2535, pruned_loss=0.05819, over 4785.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2487, pruned_loss=0.05281, over 953096.27 frames. ], batch size: 29, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:12:02,069 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.237e+01 1.564e+02 1.874e+02 2.284e+02 3.839e+02, threshold=3.748e+02, percent-clipped=2.0 +2023-03-27 04:12:06,381 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129769.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:12:09,459 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4657, 1.3676, 1.9067, 1.7734, 1.5787, 3.3891, 1.2849, 1.4619], + device='cuda:0'), covar=tensor([0.1091, 0.2072, 0.1356, 0.1091, 0.1783, 0.0277, 0.1788, 0.2197], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:12:20,300 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129791.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:12:35,162 INFO [finetune.py:976] (0/7) Epoch 23, batch 3800, loss[loss=0.1595, simple_loss=0.2393, pruned_loss=0.03983, over 4748.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2501, pruned_loss=0.05296, over 954356.54 frames. ], batch size: 28, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:13:04,420 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5564, 1.1703, 0.9153, 1.4994, 1.9442, 1.1901, 1.3447, 1.5400], + device='cuda:0'), covar=tensor([0.1510, 0.2043, 0.1812, 0.1160, 0.1979, 0.1967, 0.1464, 0.1779], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 04:13:16,852 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129852.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:13:21,830 INFO [finetune.py:976] (0/7) Epoch 23, batch 3850, loss[loss=0.1634, simple_loss=0.2382, pruned_loss=0.04431, over 4831.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.248, pruned_loss=0.05153, over 955043.06 frames. ], batch size: 47, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:13:24,150 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.288e+01 1.597e+02 1.881e+02 2.160e+02 3.613e+02, threshold=3.763e+02, percent-clipped=0.0 +2023-03-27 04:13:50,579 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9665, 1.4281, 0.8055, 2.0691, 2.3907, 1.8208, 1.6270, 1.9436], + device='cuda:0'), covar=tensor([0.1385, 0.1921, 0.1999, 0.1022, 0.1828, 0.1769, 0.1439, 0.1781], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0091, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 04:13:55,057 INFO [finetune.py:976] (0/7) Epoch 23, batch 3900, loss[loss=0.1657, simple_loss=0.2421, pruned_loss=0.04471, over 4842.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2443, pruned_loss=0.05034, over 955286.58 frames. ], batch size: 49, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:14:27,720 INFO [finetune.py:976] (0/7) Epoch 23, batch 3950, loss[loss=0.1774, simple_loss=0.2338, pruned_loss=0.06051, over 4829.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2412, pruned_loss=0.04973, over 952993.90 frames. ], batch size: 30, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:14:29,947 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.481e+02 1.820e+02 2.101e+02 4.779e+02, threshold=3.640e+02, percent-clipped=1.0 +2023-03-27 04:14:34,548 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129968.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:14:34,599 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4908, 1.5094, 1.3144, 1.5154, 1.7326, 1.6842, 1.4739, 1.2885], + device='cuda:0'), covar=tensor([0.0350, 0.0278, 0.0678, 0.0274, 0.0265, 0.0513, 0.0358, 0.0436], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0145, 0.0112, 0.0100, 0.0112, 0.0102, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7685e-05, 8.1909e-05, 1.1321e-04, 8.5495e-05, 7.7529e-05, 8.2758e-05, + 7.5952e-05, 8.5254e-05], device='cuda:0') +2023-03-27 04:14:52,745 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-27 04:14:55,077 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-130000.pt +2023-03-27 04:14:56,353 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130000.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:14:58,093 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:15:02,811 INFO [finetune.py:976] (0/7) Epoch 23, batch 4000, loss[loss=0.1651, simple_loss=0.2521, pruned_loss=0.03903, over 4820.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2416, pruned_loss=0.04979, over 954104.59 frames. ], batch size: 39, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:15:45,422 INFO [finetune.py:976] (0/7) Epoch 23, batch 4050, loss[loss=0.2195, simple_loss=0.2902, pruned_loss=0.07441, over 4930.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2449, pruned_loss=0.05103, over 955039.18 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:15:47,207 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130061.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:15:47,689 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.665e+02 1.960e+02 2.479e+02 5.275e+02, threshold=3.921e+02, percent-clipped=4.0 +2023-03-27 04:15:48,433 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130063.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:15:53,625 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130069.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:16:04,328 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3165, 3.7858, 3.9688, 4.1460, 4.0792, 3.8100, 4.3933, 1.4183], + device='cuda:0'), covar=tensor([0.0766, 0.0823, 0.0905, 0.0966, 0.1175, 0.1614, 0.0667, 0.5450], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0247, 0.0280, 0.0294, 0.0339, 0.0287, 0.0306, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:16:19,201 INFO [finetune.py:976] (0/7) Epoch 23, batch 4100, loss[loss=0.1951, simple_loss=0.2698, pruned_loss=0.06017, over 4928.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2475, pruned_loss=0.05184, over 952944.18 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:16:26,575 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130117.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:16:54,980 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:17:02,631 INFO [finetune.py:976] (0/7) Epoch 23, batch 4150, loss[loss=0.1774, simple_loss=0.2521, pruned_loss=0.05138, over 4817.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2488, pruned_loss=0.05237, over 953477.87 frames. ], batch size: 30, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:17:04,904 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.505e+02 1.863e+02 2.291e+02 4.324e+02, threshold=3.726e+02, percent-clipped=3.0 +2023-03-27 04:17:34,884 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7970, 3.9382, 3.7739, 1.7993, 4.0379, 3.0173, 0.8790, 2.7154], + device='cuda:0'), covar=tensor([0.2283, 0.2108, 0.1581, 0.3336, 0.1063, 0.0932, 0.4433, 0.1512], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0160, 0.0129, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:17:36,688 INFO [finetune.py:976] (0/7) Epoch 23, batch 4200, loss[loss=0.1272, simple_loss=0.2065, pruned_loss=0.02395, over 4780.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2486, pruned_loss=0.05164, over 952433.49 frames. ], batch size: 26, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:17:39,240 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 04:17:50,647 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8228, 1.0343, 1.8468, 1.8142, 1.6338, 1.5863, 1.7018, 1.7732], + device='cuda:0'), covar=tensor([0.3587, 0.3992, 0.3298, 0.3445, 0.4607, 0.3611, 0.4162, 0.2996], + device='cuda:0'), in_proj_covar=tensor([0.0256, 0.0241, 0.0262, 0.0285, 0.0284, 0.0260, 0.0292, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:17:53,662 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5583, 1.4241, 1.4885, 0.8463, 1.5772, 1.5081, 1.5085, 1.3671], + device='cuda:0'), covar=tensor([0.0587, 0.0837, 0.0718, 0.0900, 0.0831, 0.0763, 0.0649, 0.1220], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0135, 0.0138, 0.0119, 0.0125, 0.0138, 0.0137, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:18:24,042 INFO [finetune.py:976] (0/7) Epoch 23, batch 4250, loss[loss=0.181, simple_loss=0.2459, pruned_loss=0.05804, over 4813.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.246, pruned_loss=0.05081, over 954636.75 frames. ], batch size: 39, lr: 3.08e-03, grad_scale: 64.0 +2023-03-27 04:18:25,853 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.516e+02 1.759e+02 2.094e+02 3.793e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-27 04:18:30,091 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130268.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:18:55,859 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 04:18:57,489 INFO [finetune.py:976] (0/7) Epoch 23, batch 4300, loss[loss=0.2095, simple_loss=0.2697, pruned_loss=0.07464, over 4862.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2421, pruned_loss=0.04974, over 954512.20 frames. ], batch size: 34, lr: 3.08e-03, grad_scale: 64.0 +2023-03-27 04:19:02,810 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130316.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:19:20,886 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8294, 1.7288, 1.5734, 1.9506, 2.3238, 1.9996, 1.6719, 1.5237], + device='cuda:0'), covar=tensor([0.2087, 0.1951, 0.1774, 0.1492, 0.1696, 0.1165, 0.2242, 0.1807], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0210, 0.0213, 0.0197, 0.0245, 0.0190, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:19:29,228 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:19:30,470 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130358.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:19:31,026 INFO [finetune.py:976] (0/7) Epoch 23, batch 4350, loss[loss=0.1296, simple_loss=0.1932, pruned_loss=0.03302, over 4768.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2395, pruned_loss=0.04911, over 955306.74 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:19:33,423 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.417e+02 1.746e+02 2.112e+02 4.412e+02, threshold=3.492e+02, percent-clipped=1.0 +2023-03-27 04:19:34,170 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5964, 1.5589, 1.5097, 1.5734, 1.2032, 3.5071, 1.4233, 1.6616], + device='cuda:0'), covar=tensor([0.3282, 0.2427, 0.2173, 0.2410, 0.1784, 0.0195, 0.2486, 0.1324], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:19:36,536 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 04:19:47,746 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:04,340 INFO [finetune.py:976] (0/7) Epoch 23, batch 4400, loss[loss=0.1905, simple_loss=0.2505, pruned_loss=0.06527, over 4809.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2411, pruned_loss=0.05002, over 952413.04 frames. ], batch size: 29, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:20:19,477 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:24,171 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130438.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:30,584 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:31,765 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130447.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:46,865 INFO [finetune.py:976] (0/7) Epoch 23, batch 4450, loss[loss=0.1658, simple_loss=0.24, pruned_loss=0.04581, over 4825.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2444, pruned_loss=0.05113, over 954059.20 frames. ], batch size: 40, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:20:49,237 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.491e+02 1.813e+02 2.246e+02 3.707e+02, threshold=3.626e+02, percent-clipped=3.0 +2023-03-27 04:20:54,211 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8633, 1.7289, 2.2287, 3.5206, 2.3895, 2.4216, 1.1906, 2.8800], + device='cuda:0'), covar=tensor([0.1595, 0.1351, 0.1228, 0.0526, 0.0689, 0.1287, 0.1791, 0.0462], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 04:20:56,036 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6912, 2.7935, 2.3942, 1.8350, 2.5975, 2.6336, 2.7636, 2.2919], + device='cuda:0'), covar=tensor([0.0562, 0.0534, 0.0765, 0.0867, 0.0834, 0.0784, 0.0608, 0.1060], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0134, 0.0137, 0.0118, 0.0124, 0.0136, 0.0136, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:21:10,572 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:11,728 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130495.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:14,693 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130499.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:20,676 INFO [finetune.py:976] (0/7) Epoch 23, batch 4500, loss[loss=0.1959, simple_loss=0.2717, pruned_loss=0.06, over 4897.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2479, pruned_loss=0.05278, over 953983.54 frames. ], batch size: 37, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:21:50,890 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8158, 2.0617, 2.2967, 2.1499, 2.0361, 4.6487, 2.1223, 2.0874], + device='cuda:0'), covar=tensor([0.0874, 0.1552, 0.1069, 0.0889, 0.1505, 0.0156, 0.1204, 0.1585], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:21:53,309 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:22:04,071 INFO [finetune.py:976] (0/7) Epoch 23, batch 4550, loss[loss=0.1557, simple_loss=0.225, pruned_loss=0.0432, over 4713.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2482, pruned_loss=0.05243, over 952996.79 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:22:04,148 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1151, 2.9257, 2.7888, 1.2068, 2.9966, 2.2437, 0.7379, 1.9124], + device='cuda:0'), covar=tensor([0.2361, 0.2112, 0.1841, 0.3435, 0.1354, 0.1134, 0.3831, 0.1596], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0176, 0.0159, 0.0128, 0.0159, 0.0122, 0.0146, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:22:06,499 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.768e+01 1.546e+02 1.777e+02 2.233e+02 3.779e+02, threshold=3.553e+02, percent-clipped=2.0 +2023-03-27 04:22:07,813 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130565.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:22:37,456 INFO [finetune.py:976] (0/7) Epoch 23, batch 4600, loss[loss=0.1532, simple_loss=0.2253, pruned_loss=0.04057, over 4759.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2475, pruned_loss=0.05203, over 951114.59 frames. ], batch size: 54, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:22:37,588 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130609.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:22:47,773 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:10,854 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130656.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:17,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130658.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:23:17,589 INFO [finetune.py:976] (0/7) Epoch 23, batch 4650, loss[loss=0.1961, simple_loss=0.2547, pruned_loss=0.0687, over 4757.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2459, pruned_loss=0.05179, over 953977.18 frames. ], batch size: 27, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:23:19,987 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.906e+01 1.515e+02 1.768e+02 2.232e+02 6.495e+02, threshold=3.536e+02, percent-clipped=3.0 +2023-03-27 04:23:20,717 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6824, 1.6844, 2.3218, 2.0494, 1.8899, 4.3641, 1.6772, 1.7769], + device='cuda:0'), covar=tensor([0.1015, 0.1817, 0.1058, 0.0955, 0.1615, 0.0191, 0.1431, 0.1838], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:23:54,289 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 04:23:54,689 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:55,883 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130706.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:58,153 INFO [finetune.py:976] (0/7) Epoch 23, batch 4700, loss[loss=0.1636, simple_loss=0.2366, pruned_loss=0.04531, over 4688.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2427, pruned_loss=0.05098, over 952324.62 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:24:18,051 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:24:31,365 INFO [finetune.py:976] (0/7) Epoch 23, batch 4750, loss[loss=0.1949, simple_loss=0.2589, pruned_loss=0.06542, over 4940.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2398, pruned_loss=0.04986, over 952025.15 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:24:34,231 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.528e+02 1.803e+02 2.150e+02 3.686e+02, threshold=3.606e+02, percent-clipped=2.0 +2023-03-27 04:24:47,134 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-03-27 04:24:49,897 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:24:54,018 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:25:04,663 INFO [finetune.py:976] (0/7) Epoch 23, batch 4800, loss[loss=0.2037, simple_loss=0.2802, pruned_loss=0.06362, over 4832.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2428, pruned_loss=0.0509, over 952611.51 frames. ], batch size: 47, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:25:37,286 INFO [finetune.py:976] (0/7) Epoch 23, batch 4850, loss[loss=0.17, simple_loss=0.2524, pruned_loss=0.04384, over 4870.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2466, pruned_loss=0.0523, over 953067.23 frames. ], batch size: 34, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:25:40,092 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.613e+02 1.947e+02 2.336e+02 6.046e+02, threshold=3.894e+02, percent-clipped=4.0 +2023-03-27 04:25:57,025 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7406, 1.5493, 1.4349, 0.7521, 1.6387, 1.7833, 1.6790, 1.3548], + device='cuda:0'), covar=tensor([0.0865, 0.0667, 0.0622, 0.0661, 0.0495, 0.0568, 0.0442, 0.0805], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0132, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9653e-05, 1.0748e-04, 9.0968e-05, 8.6536e-05, 9.2717e-05, 9.2594e-05, + 1.0168e-04, 1.0669e-04], device='cuda:0') +2023-03-27 04:26:15,652 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130904.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:26:19,115 INFO [finetune.py:976] (0/7) Epoch 23, batch 4900, loss[loss=0.1826, simple_loss=0.2375, pruned_loss=0.0638, over 4742.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2478, pruned_loss=0.05227, over 954890.79 frames. ], batch size: 23, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:26:28,435 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130921.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:26:46,419 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8199, 1.6946, 1.5189, 1.4372, 1.8492, 1.5807, 1.8377, 1.7927], + device='cuda:0'), covar=tensor([0.1418, 0.1816, 0.2806, 0.2487, 0.2458, 0.1629, 0.2639, 0.1707], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0188, 0.0234, 0.0253, 0.0247, 0.0204, 0.0212, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:26:52,302 INFO [finetune.py:976] (0/7) Epoch 23, batch 4950, loss[loss=0.1412, simple_loss=0.2175, pruned_loss=0.0325, over 4783.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.249, pruned_loss=0.05232, over 955033.00 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:26:57,594 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.290e+01 1.586e+02 1.789e+02 2.374e+02 3.586e+02, threshold=3.578e+02, percent-clipped=0.0 +2023-03-27 04:26:57,820 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 04:27:21,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3368, 1.3044, 1.6252, 1.1976, 1.3176, 1.4932, 1.2901, 1.6546], + device='cuda:0'), covar=tensor([0.1158, 0.2220, 0.1299, 0.1479, 0.1030, 0.1305, 0.3042, 0.0955], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0192, 0.0191, 0.0174, 0.0214, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:27:25,602 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7364, 2.4669, 2.4022, 1.4910, 2.5620, 2.0279, 1.8940, 2.3867], + device='cuda:0'), covar=tensor([0.1170, 0.0723, 0.1728, 0.2060, 0.1498, 0.2071, 0.2005, 0.1077], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0210, 0.0210, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:27:36,348 INFO [finetune.py:976] (0/7) Epoch 23, batch 5000, loss[loss=0.1227, simple_loss=0.1896, pruned_loss=0.02786, over 4826.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2473, pruned_loss=0.05148, over 955318.41 frames. ], batch size: 49, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:27:36,657 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-27 04:27:57,565 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:09,927 INFO [finetune.py:976] (0/7) Epoch 23, batch 5050, loss[loss=0.1468, simple_loss=0.2239, pruned_loss=0.03484, over 4823.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2454, pruned_loss=0.05175, over 955243.63 frames. ], batch size: 25, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:28:12,371 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.381e+02 1.770e+02 2.059e+02 4.416e+02, threshold=3.539e+02, percent-clipped=4.0 +2023-03-27 04:28:41,482 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:41,508 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:45,126 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131094.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:45,168 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6548, 3.1995, 3.0025, 1.5394, 3.1443, 2.5992, 2.5510, 2.8680], + device='cuda:0'), covar=tensor([0.0889, 0.0717, 0.1495, 0.1961, 0.1358, 0.1740, 0.1703, 0.0959], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0192, 0.0201, 0.0181, 0.0209, 0.0209, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:28:57,921 INFO [finetune.py:976] (0/7) Epoch 23, batch 5100, loss[loss=0.1273, simple_loss=0.2084, pruned_loss=0.02312, over 4783.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2411, pruned_loss=0.04989, over 956109.91 frames. ], batch size: 29, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:28:59,147 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6277, 1.6245, 1.6294, 0.8774, 1.8891, 2.0151, 1.9333, 1.5087], + device='cuda:0'), covar=tensor([0.1280, 0.0937, 0.0578, 0.0695, 0.0436, 0.0689, 0.0411, 0.0871], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0127, 0.0123, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9193e-05, 1.0689e-04, 9.0535e-05, 8.6332e-05, 9.1806e-05, 9.2198e-05, + 1.0076e-04, 1.0600e-04], device='cuda:0') +2023-03-27 04:29:04,674 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 04:29:17,257 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:20,885 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:31,083 INFO [finetune.py:976] (0/7) Epoch 23, batch 5150, loss[loss=0.2211, simple_loss=0.2695, pruned_loss=0.0863, over 4817.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2411, pruned_loss=0.05009, over 952675.57 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:29:34,465 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.214e+01 1.572e+02 1.903e+02 2.241e+02 4.010e+02, threshold=3.805e+02, percent-clipped=1.0 +2023-03-27 04:29:41,263 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131174.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:46,550 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5096, 1.4758, 1.7705, 1.7205, 1.5674, 3.3095, 1.3644, 1.5553], + device='cuda:0'), covar=tensor([0.1028, 0.1779, 0.1101, 0.0966, 0.1687, 0.0271, 0.1501, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:29:57,493 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2745, 1.4467, 0.7471, 2.0530, 2.5066, 1.9359, 1.9091, 1.9203], + device='cuda:0'), covar=tensor([0.1317, 0.2121, 0.2073, 0.1189, 0.1764, 0.1782, 0.1387, 0.2000], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 04:30:01,315 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131204.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:30:04,269 INFO [finetune.py:976] (0/7) Epoch 23, batch 5200, loss[loss=0.1278, simple_loss=0.1928, pruned_loss=0.03145, over 4174.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2454, pruned_loss=0.05149, over 954469.00 frames. ], batch size: 18, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:30:10,898 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 04:30:12,619 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:30:23,011 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131235.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:30:33,241 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131252.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:30:37,393 INFO [finetune.py:976] (0/7) Epoch 23, batch 5250, loss[loss=0.1595, simple_loss=0.234, pruned_loss=0.04253, over 4873.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2476, pruned_loss=0.05185, over 956574.34 frames. ], batch size: 31, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:30:40,888 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.531e+02 1.792e+02 2.239e+02 3.281e+02, threshold=3.585e+02, percent-clipped=0.0 +2023-03-27 04:30:44,379 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:31:21,562 INFO [finetune.py:976] (0/7) Epoch 23, batch 5300, loss[loss=0.1851, simple_loss=0.2548, pruned_loss=0.05773, over 4817.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2484, pruned_loss=0.05224, over 955461.31 frames. ], batch size: 40, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:31:49,605 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1485, 2.0521, 1.5684, 0.6000, 1.7007, 1.8332, 1.7015, 1.9058], + device='cuda:0'), covar=tensor([0.0965, 0.0759, 0.1549, 0.2073, 0.1328, 0.2313, 0.2222, 0.0850], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0210, 0.0210, 0.0224, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:31:54,360 INFO [finetune.py:976] (0/7) Epoch 23, batch 5350, loss[loss=0.1194, simple_loss=0.192, pruned_loss=0.02337, over 4015.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2494, pruned_loss=0.05241, over 957228.98 frames. ], batch size: 17, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:31:57,391 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.529e+02 1.830e+02 2.196e+02 3.219e+02, threshold=3.659e+02, percent-clipped=0.0 +2023-03-27 04:31:57,519 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2410, 1.1570, 1.4531, 1.0406, 1.1639, 1.3584, 1.1220, 1.4956], + device='cuda:0'), covar=tensor([0.1329, 0.2522, 0.1495, 0.1688, 0.1208, 0.1522, 0.3394, 0.1220], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0191, 0.0190, 0.0172, 0.0213, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:32:32,842 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-27 04:32:38,097 INFO [finetune.py:976] (0/7) Epoch 23, batch 5400, loss[loss=0.1643, simple_loss=0.2302, pruned_loss=0.04917, over 4762.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2461, pruned_loss=0.05144, over 957191.21 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:32:38,215 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131409.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:32:42,428 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131416.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:11,755 INFO [finetune.py:976] (0/7) Epoch 23, batch 5450, loss[loss=0.2182, simple_loss=0.2772, pruned_loss=0.07965, over 4821.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2439, pruned_loss=0.05123, over 954876.86 frames. ], batch size: 40, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:33:14,788 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.514e+02 1.875e+02 2.409e+02 5.439e+02, threshold=3.749e+02, percent-clipped=4.0 +2023-03-27 04:33:18,557 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131470.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:23,291 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131477.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:33,344 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131492.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:51,830 INFO [finetune.py:976] (0/7) Epoch 23, batch 5500, loss[loss=0.1604, simple_loss=0.2376, pruned_loss=0.04163, over 4813.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2408, pruned_loss=0.05004, over 955144.69 frames. ], batch size: 41, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:33:58,481 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3649, 1.4454, 1.9692, 1.6219, 1.4840, 3.2855, 1.3184, 1.5269], + device='cuda:0'), covar=tensor([0.0922, 0.1618, 0.1057, 0.0930, 0.1559, 0.0260, 0.1430, 0.1712], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0081, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:34:12,685 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131530.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:34:29,163 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131553.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:34:33,176 INFO [finetune.py:976] (0/7) Epoch 23, batch 5550, loss[loss=0.1844, simple_loss=0.2609, pruned_loss=0.05397, over 4919.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2425, pruned_loss=0.05048, over 953917.71 frames. ], batch size: 33, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:34:36,710 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.823e+01 1.422e+02 1.728e+02 2.186e+02 5.215e+02, threshold=3.457e+02, percent-clipped=2.0 +2023-03-27 04:34:39,877 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5739, 0.7101, 1.6648, 1.5595, 1.4620, 1.4145, 1.5110, 1.6155], + device='cuda:0'), covar=tensor([0.3191, 0.3416, 0.2626, 0.2976, 0.3870, 0.3194, 0.3422, 0.2476], + device='cuda:0'), in_proj_covar=tensor([0.0259, 0.0245, 0.0265, 0.0288, 0.0288, 0.0264, 0.0296, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:35:04,721 INFO [finetune.py:976] (0/7) Epoch 23, batch 5600, loss[loss=0.1714, simple_loss=0.2478, pruned_loss=0.04748, over 4820.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2457, pruned_loss=0.0514, over 950524.69 frames. ], batch size: 25, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:35:30,064 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7218, 1.7082, 1.9302, 1.3285, 1.6338, 1.9501, 1.6159, 2.1121], + device='cuda:0'), covar=tensor([0.1096, 0.1918, 0.1284, 0.1669, 0.1013, 0.1213, 0.2604, 0.0781], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0191, 0.0189, 0.0172, 0.0213, 0.0215, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:35:31,205 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4061, 1.4763, 1.7030, 1.5393, 1.6748, 3.0354, 1.4255, 1.5754], + device='cuda:0'), covar=tensor([0.1008, 0.1818, 0.1077, 0.1029, 0.1614, 0.0303, 0.1484, 0.1817], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0081, 0.0072, 0.0076, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:35:34,623 INFO [finetune.py:976] (0/7) Epoch 23, batch 5650, loss[loss=0.2129, simple_loss=0.2828, pruned_loss=0.07153, over 4844.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2478, pruned_loss=0.05188, over 951096.88 frames. ], batch size: 47, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:35:37,860 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.955e+01 1.494e+02 1.801e+02 2.339e+02 4.576e+02, threshold=3.601e+02, percent-clipped=4.0 +2023-03-27 04:35:57,404 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.7565, 4.2443, 4.4289, 4.5527, 4.5365, 4.2734, 4.8397, 2.0838], + device='cuda:0'), covar=tensor([0.0654, 0.0744, 0.0737, 0.0730, 0.1003, 0.1285, 0.0616, 0.4342], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0246, 0.0279, 0.0292, 0.0337, 0.0286, 0.0304, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:36:04,502 INFO [finetune.py:976] (0/7) Epoch 23, batch 5700, loss[loss=0.1574, simple_loss=0.2133, pruned_loss=0.05081, over 4316.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2452, pruned_loss=0.05136, over 935524.22 frames. ], batch size: 19, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:36:07,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6018, 1.5351, 1.4870, 1.5633, 1.1841, 2.7865, 1.2855, 1.5739], + device='cuda:0'), covar=tensor([0.2959, 0.2371, 0.2074, 0.2248, 0.1675, 0.0269, 0.2519, 0.1189], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:36:16,339 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5016, 2.3317, 2.4311, 1.7732, 2.3327, 2.4955, 2.5127, 2.1026], + device='cuda:0'), covar=tensor([0.0608, 0.0681, 0.0687, 0.0861, 0.0872, 0.0764, 0.0606, 0.1026], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0135, 0.0139, 0.0119, 0.0125, 0.0138, 0.0137, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:36:27,940 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-23.pt +2023-03-27 04:36:40,049 INFO [finetune.py:976] (0/7) Epoch 24, batch 0, loss[loss=0.1246, simple_loss=0.1899, pruned_loss=0.02962, over 4524.00 frames. ], tot_loss[loss=0.1246, simple_loss=0.1899, pruned_loss=0.02962, over 4524.00 frames. ], batch size: 19, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:36:40,050 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 04:36:49,303 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6426, 1.5630, 1.5184, 1.5647, 1.0651, 2.9379, 1.2114, 1.5671], + device='cuda:0'), covar=tensor([0.3294, 0.2414, 0.2126, 0.2288, 0.1798, 0.0261, 0.2492, 0.1267], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:36:50,762 INFO [finetune.py:1010] (0/7) Epoch 24, validation: loss=0.1594, simple_loss=0.227, pruned_loss=0.04592, over 2265189.00 frames. +2023-03-27 04:36:50,763 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 04:37:07,461 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.206e+01 1.398e+02 1.674e+02 2.004e+02 3.219e+02, threshold=3.348e+02, percent-clipped=0.0 +2023-03-27 04:37:08,156 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131765.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:37:12,933 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131772.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:37:18,151 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3021, 3.7818, 3.9503, 4.1718, 4.0430, 3.7974, 4.4070, 1.2691], + device='cuda:0'), covar=tensor([0.0771, 0.0870, 0.0868, 0.0885, 0.1140, 0.1634, 0.0663, 0.6023], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0244, 0.0276, 0.0290, 0.0335, 0.0284, 0.0302, 0.0297], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:37:25,288 INFO [finetune.py:976] (0/7) Epoch 24, batch 50, loss[loss=0.1563, simple_loss=0.2259, pruned_loss=0.04329, over 4753.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2483, pruned_loss=0.05112, over 215222.07 frames. ], batch size: 28, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:38:02,583 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:07,275 INFO [finetune.py:976] (0/7) Epoch 24, batch 100, loss[loss=0.1104, simple_loss=0.1683, pruned_loss=0.0262, over 4300.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2396, pruned_loss=0.04823, over 379386.97 frames. ], batch size: 18, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:38:15,489 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131848.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:20,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131857.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:25,097 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.662e+01 1.465e+02 1.761e+02 2.142e+02 3.724e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-03-27 04:38:34,589 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131878.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:40,507 INFO [finetune.py:976] (0/7) Epoch 24, batch 150, loss[loss=0.1562, simple_loss=0.223, pruned_loss=0.04468, over 4813.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2368, pruned_loss=0.04778, over 505649.53 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:38:44,583 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3299, 1.4506, 1.8244, 1.6321, 1.5930, 3.2968, 1.4599, 1.5561], + device='cuda:0'), covar=tensor([0.1037, 0.1764, 0.1162, 0.1026, 0.1511, 0.0247, 0.1402, 0.1803], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:39:10,817 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:39:29,543 INFO [finetune.py:976] (0/7) Epoch 24, batch 200, loss[loss=0.1925, simple_loss=0.2609, pruned_loss=0.06202, over 4103.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2371, pruned_loss=0.04918, over 604019.03 frames. ], batch size: 65, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:39:51,183 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.517e+02 1.799e+02 2.123e+02 6.232e+02, threshold=3.598e+02, percent-clipped=3.0 +2023-03-27 04:39:52,062 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-03-27 04:40:06,650 INFO [finetune.py:976] (0/7) Epoch 24, batch 250, loss[loss=0.1731, simple_loss=0.2406, pruned_loss=0.05283, over 4739.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2406, pruned_loss=0.05016, over 681897.80 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:40:10,322 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 04:40:16,532 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-132000.pt +2023-03-27 04:40:36,969 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132031.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:40:41,436 INFO [finetune.py:976] (0/7) Epoch 24, batch 300, loss[loss=0.1688, simple_loss=0.2194, pruned_loss=0.05913, over 4723.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2441, pruned_loss=0.05126, over 740317.23 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:40:51,685 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-27 04:40:53,241 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132054.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:40:59,162 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.635e+02 1.887e+02 2.261e+02 6.512e+02, threshold=3.774e+02, percent-clipped=2.0 +2023-03-27 04:40:59,870 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132065.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:04,179 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:08,894 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5210, 1.4186, 1.4098, 1.4823, 0.8441, 2.8959, 0.9943, 1.4137], + device='cuda:0'), covar=tensor([0.3274, 0.2548, 0.2167, 0.2500, 0.2001, 0.0246, 0.2724, 0.1337], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:41:14,122 INFO [finetune.py:976] (0/7) Epoch 24, batch 350, loss[loss=0.1475, simple_loss=0.2196, pruned_loss=0.03772, over 4095.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.247, pruned_loss=0.05224, over 788221.17 frames. ], batch size: 17, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:41:17,750 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132092.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:33,451 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132113.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:34,747 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132115.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:41:42,163 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132120.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:42,201 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9399, 1.4139, 0.7509, 1.8671, 2.2964, 1.8036, 1.6868, 1.9064], + device='cuda:0'), covar=tensor([0.1359, 0.1837, 0.2008, 0.1103, 0.1786, 0.1872, 0.1311, 0.1715], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 04:41:56,059 INFO [finetune.py:976] (0/7) Epoch 24, batch 400, loss[loss=0.1859, simple_loss=0.2495, pruned_loss=0.06113, over 4835.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2479, pruned_loss=0.05253, over 825356.29 frames. ], batch size: 30, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:42:03,880 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132148.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:42:15,409 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.594e+02 1.883e+02 2.349e+02 4.739e+02, threshold=3.766e+02, percent-clipped=1.0 +2023-03-27 04:42:29,849 INFO [finetune.py:976] (0/7) Epoch 24, batch 450, loss[loss=0.1842, simple_loss=0.2451, pruned_loss=0.06163, over 4812.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2468, pruned_loss=0.05201, over 853389.70 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:42:33,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7926, 2.4645, 2.3947, 1.3617, 2.5909, 2.1559, 1.9737, 2.3817], + device='cuda:0'), covar=tensor([0.1197, 0.0871, 0.1974, 0.1984, 0.1524, 0.1863, 0.2080, 0.1168], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0193, 0.0201, 0.0182, 0.0211, 0.0211, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:42:36,324 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132196.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:42:55,015 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:43:02,541 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6376, 1.6039, 1.4659, 1.6761, 1.9562, 1.8399, 1.6990, 1.4381], + device='cuda:0'), covar=tensor([0.0335, 0.0311, 0.0589, 0.0309, 0.0221, 0.0467, 0.0278, 0.0415], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0105, 0.0142, 0.0110, 0.0098, 0.0110, 0.0100, 0.0110], + device='cuda:0'), out_proj_covar=tensor([7.6346e-05, 8.0442e-05, 1.1112e-04, 8.3899e-05, 7.6392e-05, 8.1165e-05, + 7.4058e-05, 8.3925e-05], device='cuda:0') +2023-03-27 04:43:13,243 INFO [finetune.py:976] (0/7) Epoch 24, batch 500, loss[loss=0.1536, simple_loss=0.2247, pruned_loss=0.04126, over 4821.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2433, pruned_loss=0.05044, over 876357.63 frames. ], batch size: 51, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:43:32,460 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.902e+01 1.519e+02 1.809e+02 2.135e+02 3.897e+02, threshold=3.617e+02, percent-clipped=1.0 +2023-03-27 04:43:36,957 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9267, 1.7557, 1.5857, 1.3764, 1.7638, 1.7606, 1.7421, 2.2551], + device='cuda:0'), covar=tensor([0.3558, 0.3601, 0.3111, 0.3427, 0.3304, 0.2201, 0.3334, 0.1731], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0232, 0.0274, 0.0256, 0.0225, 0.0252, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:43:46,923 INFO [finetune.py:976] (0/7) Epoch 24, batch 550, loss[loss=0.1603, simple_loss=0.2342, pruned_loss=0.04316, over 4829.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2396, pruned_loss=0.04926, over 895997.55 frames. ], batch size: 33, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:44:26,892 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 04:44:30,202 INFO [finetune.py:976] (0/7) Epoch 24, batch 600, loss[loss=0.1808, simple_loss=0.245, pruned_loss=0.05828, over 4719.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2403, pruned_loss=0.04974, over 910757.53 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:44:58,205 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.737e+01 1.576e+02 1.859e+02 2.330e+02 3.343e+02, threshold=3.718e+02, percent-clipped=0.0 +2023-03-27 04:45:12,585 INFO [finetune.py:976] (0/7) Epoch 24, batch 650, loss[loss=0.1825, simple_loss=0.2528, pruned_loss=0.05611, over 4824.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2444, pruned_loss=0.05138, over 920860.54 frames. ], batch size: 45, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:45:12,659 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132387.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:45:28,113 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132410.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:45:43,335 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 04:45:46,139 INFO [finetune.py:976] (0/7) Epoch 24, batch 700, loss[loss=0.1719, simple_loss=0.2292, pruned_loss=0.0573, over 4751.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2452, pruned_loss=0.05201, over 927370.10 frames. ], batch size: 26, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:46:03,856 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.592e+02 1.911e+02 2.262e+02 4.191e+02, threshold=3.822e+02, percent-clipped=2.0 +2023-03-27 04:46:19,328 INFO [finetune.py:976] (0/7) Epoch 24, batch 750, loss[loss=0.1693, simple_loss=0.248, pruned_loss=0.04526, over 4811.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2467, pruned_loss=0.05227, over 930908.13 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:46:36,490 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:46:47,371 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6845, 1.6094, 1.6590, 1.6343, 1.0702, 2.8310, 1.2008, 1.5568], + device='cuda:0'), covar=tensor([0.2730, 0.2068, 0.1806, 0.1992, 0.1634, 0.0278, 0.2180, 0.1079], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:46:58,683 INFO [finetune.py:976] (0/7) Epoch 24, batch 800, loss[loss=0.1788, simple_loss=0.2481, pruned_loss=0.0547, over 4856.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2463, pruned_loss=0.05173, over 933921.35 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:46:58,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5240, 1.4162, 1.4661, 1.4832, 0.8983, 2.8867, 1.1018, 1.4566], + device='cuda:0'), covar=tensor([0.3275, 0.2531, 0.2138, 0.2331, 0.1915, 0.0236, 0.2699, 0.1312], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:47:16,685 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 04:47:17,150 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132561.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:47:19,958 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.775e+01 1.480e+02 1.729e+02 2.075e+02 4.531e+02, threshold=3.459e+02, percent-clipped=1.0 +2023-03-27 04:47:35,955 INFO [finetune.py:976] (0/7) Epoch 24, batch 850, loss[loss=0.1864, simple_loss=0.255, pruned_loss=0.05895, over 4902.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.245, pruned_loss=0.05161, over 938518.22 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:47:42,777 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6073, 1.6158, 1.9830, 1.3463, 1.7551, 1.9875, 1.4728, 2.1166], + device='cuda:0'), covar=tensor([0.1187, 0.1880, 0.1164, 0.1644, 0.0807, 0.1202, 0.2745, 0.0710], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0206, 0.0191, 0.0190, 0.0172, 0.0213, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:48:18,617 INFO [finetune.py:976] (0/7) Epoch 24, batch 900, loss[loss=0.1481, simple_loss=0.2149, pruned_loss=0.04065, over 4836.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2419, pruned_loss=0.05094, over 939398.65 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:48:33,138 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132660.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:48:35,407 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.417e+02 1.718e+02 2.002e+02 3.598e+02, threshold=3.436e+02, percent-clipped=1.0 +2023-03-27 04:48:52,530 INFO [finetune.py:976] (0/7) Epoch 24, batch 950, loss[loss=0.2296, simple_loss=0.2926, pruned_loss=0.08324, over 4734.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2399, pruned_loss=0.0498, over 943561.32 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:48:52,623 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132687.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:01,240 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-03-27 04:49:03,546 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7404, 1.1909, 0.9056, 1.5551, 2.1638, 1.1204, 1.4353, 1.4253], + device='cuda:0'), covar=tensor([0.1439, 0.2140, 0.1859, 0.1262, 0.1797, 0.1828, 0.1495, 0.2216], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 04:49:03,670 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 04:49:07,062 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132710.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:49:14,308 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132721.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:26,441 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:28,062 INFO [finetune.py:976] (0/7) Epoch 24, batch 1000, loss[loss=0.1958, simple_loss=0.2674, pruned_loss=0.06209, over 4902.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2423, pruned_loss=0.05118, over 944355.38 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:49:38,677 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:50,801 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132758.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:54,881 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.613e+02 1.803e+02 2.355e+02 4.590e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-27 04:50:17,564 INFO [finetune.py:976] (0/7) Epoch 24, batch 1050, loss[loss=0.2148, simple_loss=0.2676, pruned_loss=0.08099, over 4919.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2448, pruned_loss=0.0514, over 947846.59 frames. ], batch size: 33, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:50:31,463 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:50:45,669 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1298, 3.6113, 3.7738, 3.9142, 3.9104, 3.6507, 4.2094, 1.4836], + device='cuda:0'), covar=tensor([0.0675, 0.0903, 0.0928, 0.0981, 0.1069, 0.1473, 0.0649, 0.5404], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0248, 0.0282, 0.0294, 0.0340, 0.0287, 0.0307, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:50:51,431 INFO [finetune.py:976] (0/7) Epoch 24, batch 1100, loss[loss=0.148, simple_loss=0.2209, pruned_loss=0.03752, over 4771.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2472, pruned_loss=0.05213, over 951801.39 frames. ], batch size: 28, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:51:08,747 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.623e+02 1.902e+02 2.304e+02 4.937e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-27 04:51:24,188 INFO [finetune.py:976] (0/7) Epoch 24, batch 1150, loss[loss=0.1797, simple_loss=0.2567, pruned_loss=0.05131, over 4835.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2475, pruned_loss=0.05212, over 951892.78 frames. ], batch size: 49, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:51:28,500 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 04:51:57,324 INFO [finetune.py:976] (0/7) Epoch 24, batch 1200, loss[loss=0.1697, simple_loss=0.232, pruned_loss=0.05375, over 4749.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2466, pruned_loss=0.05169, over 953351.77 frames. ], batch size: 27, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:52:24,716 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.055e+01 1.470e+02 1.716e+02 2.148e+02 3.548e+02, threshold=3.432e+02, percent-clipped=0.0 +2023-03-27 04:52:36,777 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4941, 3.4143, 3.2966, 1.5827, 3.4401, 2.6559, 0.8218, 2.1950], + device='cuda:0'), covar=tensor([0.2150, 0.1654, 0.1552, 0.3068, 0.1159, 0.0977, 0.4120, 0.1553], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0129, 0.0161, 0.0124, 0.0149, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:52:40,277 INFO [finetune.py:976] (0/7) Epoch 24, batch 1250, loss[loss=0.172, simple_loss=0.2482, pruned_loss=0.04787, over 4867.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2446, pruned_loss=0.0511, over 954140.09 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:52:46,408 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.57 vs. limit=5.0 +2023-03-27 04:52:59,659 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133016.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:53:15,462 INFO [finetune.py:976] (0/7) Epoch 24, batch 1300, loss[loss=0.1308, simple_loss=0.1985, pruned_loss=0.03154, over 4745.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2412, pruned_loss=0.04997, over 951471.62 frames. ], batch size: 27, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:53:42,213 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.345e+01 1.496e+02 1.852e+02 2.141e+02 4.041e+02, threshold=3.705e+02, percent-clipped=2.0 +2023-03-27 04:53:57,207 INFO [finetune.py:976] (0/7) Epoch 24, batch 1350, loss[loss=0.1664, simple_loss=0.2338, pruned_loss=0.04951, over 4807.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2419, pruned_loss=0.05049, over 951573.81 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:54:07,446 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:54:10,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6564, 1.5227, 1.3858, 1.3052, 1.4773, 1.4430, 1.4837, 1.9977], + device='cuda:0'), covar=tensor([0.3099, 0.3312, 0.2485, 0.2807, 0.3128, 0.1997, 0.2925, 0.1565], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0262, 0.0233, 0.0276, 0.0257, 0.0225, 0.0253, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:54:31,058 INFO [finetune.py:976] (0/7) Epoch 24, batch 1400, loss[loss=0.1797, simple_loss=0.2586, pruned_loss=0.05042, over 4819.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2454, pruned_loss=0.05104, over 951428.42 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:54:59,477 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.355e+01 1.624e+02 1.914e+02 2.315e+02 3.947e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-03-27 04:55:19,623 INFO [finetune.py:976] (0/7) Epoch 24, batch 1450, loss[loss=0.1543, simple_loss=0.2301, pruned_loss=0.0392, over 4755.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.247, pruned_loss=0.05174, over 952449.47 frames. ], batch size: 28, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:55:20,983 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3929, 1.3008, 1.7704, 1.6375, 1.5136, 3.2061, 1.3404, 1.4218], + device='cuda:0'), covar=tensor([0.1020, 0.1868, 0.1168, 0.1025, 0.1663, 0.0269, 0.1553, 0.1894], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 04:55:56,690 INFO [finetune.py:976] (0/7) Epoch 24, batch 1500, loss[loss=0.2076, simple_loss=0.2824, pruned_loss=0.06643, over 4808.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2488, pruned_loss=0.05192, over 954769.74 frames. ], batch size: 39, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:56:15,020 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.579e+02 1.864e+02 2.355e+02 5.095e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-27 04:56:30,467 INFO [finetune.py:976] (0/7) Epoch 24, batch 1550, loss[loss=0.1791, simple_loss=0.2578, pruned_loss=0.05019, over 4907.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2496, pruned_loss=0.05219, over 952926.90 frames. ], batch size: 37, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:56:50,661 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133316.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:04,264 INFO [finetune.py:976] (0/7) Epoch 24, batch 1600, loss[loss=0.1522, simple_loss=0.221, pruned_loss=0.04173, over 4811.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2461, pruned_loss=0.05085, over 953503.44 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:57:08,021 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4318, 1.3632, 1.4449, 0.8000, 1.4915, 1.4792, 1.4496, 1.3074], + device='cuda:0'), covar=tensor([0.0610, 0.0849, 0.0713, 0.0984, 0.1055, 0.0710, 0.0675, 0.1320], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0136, 0.0139, 0.0119, 0.0127, 0.0138, 0.0138, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:57:12,168 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5557, 3.3465, 3.2589, 1.5894, 3.4789, 2.6165, 0.7405, 2.2889], + device='cuda:0'), covar=tensor([0.2245, 0.2172, 0.1599, 0.3291, 0.1164, 0.1085, 0.4215, 0.1604], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0130, 0.0162, 0.0124, 0.0149, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 04:57:28,471 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.451e+02 1.796e+02 2.043e+02 3.402e+02, threshold=3.593e+02, percent-clipped=0.0 +2023-03-27 04:57:28,564 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133364.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:38,826 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133374.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:57:40,055 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:46,648 INFO [finetune.py:976] (0/7) Epoch 24, batch 1650, loss[loss=0.1857, simple_loss=0.2568, pruned_loss=0.05733, over 4918.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2422, pruned_loss=0.04951, over 953383.13 frames. ], batch size: 46, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:57:55,364 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1127, 1.9923, 1.7317, 1.9150, 1.8743, 1.9054, 1.9086, 2.6447], + device='cuda:0'), covar=tensor([0.3824, 0.4247, 0.3228, 0.3764, 0.4087, 0.2333, 0.3780, 0.1637], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0264, 0.0235, 0.0278, 0.0258, 0.0228, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:57:56,869 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133402.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:05,764 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6850, 1.4659, 2.2211, 1.3774, 1.9178, 1.9476, 1.3610, 2.1774], + device='cuda:0'), covar=tensor([0.1438, 0.2281, 0.1141, 0.1811, 0.0938, 0.1535, 0.3010, 0.0997], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0208, 0.0192, 0.0190, 0.0173, 0.0214, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:58:19,419 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133435.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:58:20,501 INFO [finetune.py:976] (0/7) Epoch 24, batch 1700, loss[loss=0.1798, simple_loss=0.2512, pruned_loss=0.05423, over 4822.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2407, pruned_loss=0.04902, over 953855.05 frames. ], batch size: 30, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:58:20,618 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:30,779 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 04:58:31,233 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:41,291 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3448, 2.2647, 1.9425, 0.8876, 2.0117, 1.7822, 1.6659, 2.1463], + device='cuda:0'), covar=tensor([0.0830, 0.0868, 0.1565, 0.1969, 0.1360, 0.2172, 0.2127, 0.0846], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0200, 0.0181, 0.0210, 0.0208, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 04:58:48,615 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.499e+01 1.457e+02 1.770e+02 2.219e+02 3.253e+02, threshold=3.541e+02, percent-clipped=0.0 +2023-03-27 04:58:55,849 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:59:04,125 INFO [finetune.py:976] (0/7) Epoch 24, batch 1750, loss[loss=0.1772, simple_loss=0.2468, pruned_loss=0.05383, over 4217.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2427, pruned_loss=0.05014, over 953138.23 frames. ], batch size: 65, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:59:35,045 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-03-27 04:59:36,819 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133535.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:59:37,911 INFO [finetune.py:976] (0/7) Epoch 24, batch 1800, loss[loss=0.1732, simple_loss=0.2478, pruned_loss=0.04931, over 4922.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2455, pruned_loss=0.05024, over 954225.02 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:59:46,448 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 04:59:57,744 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.575e+02 1.839e+02 2.282e+02 3.463e+02, threshold=3.677e+02, percent-clipped=0.0 +2023-03-27 04:59:57,884 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9112, 1.9123, 1.6300, 2.1019, 2.5394, 2.0748, 1.8325, 1.5107], + device='cuda:0'), covar=tensor([0.2122, 0.1876, 0.1807, 0.1485, 0.1512, 0.1087, 0.2086, 0.1981], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0213, 0.0196, 0.0244, 0.0190, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:00:23,456 INFO [finetune.py:976] (0/7) Epoch 24, batch 1850, loss[loss=0.1534, simple_loss=0.227, pruned_loss=0.03989, over 4797.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2482, pruned_loss=0.0519, over 953699.19 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:00:47,732 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5021, 1.0765, 0.7309, 1.3141, 2.0139, 0.7442, 1.2455, 1.3690], + device='cuda:0'), covar=tensor([0.1634, 0.2199, 0.1798, 0.1340, 0.1982, 0.2085, 0.1576, 0.2142], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 05:01:04,053 INFO [finetune.py:976] (0/7) Epoch 24, batch 1900, loss[loss=0.1817, simple_loss=0.2467, pruned_loss=0.05833, over 4830.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2498, pruned_loss=0.05252, over 955519.33 frames. ], batch size: 47, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:01:14,206 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133652.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:01:21,802 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.540e+02 1.881e+02 2.277e+02 3.366e+02, threshold=3.762e+02, percent-clipped=0.0 +2023-03-27 05:01:28,980 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2618, 2.2066, 1.8629, 2.0980, 2.0284, 2.0511, 2.0788, 2.7878], + device='cuda:0'), covar=tensor([0.3939, 0.4291, 0.3214, 0.3932, 0.3790, 0.2479, 0.3780, 0.1683], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0263, 0.0235, 0.0277, 0.0258, 0.0228, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:01:36,060 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 05:01:37,658 INFO [finetune.py:976] (0/7) Epoch 24, batch 1950, loss[loss=0.1185, simple_loss=0.1893, pruned_loss=0.02382, over 4714.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2484, pruned_loss=0.0517, over 956385.32 frames. ], batch size: 23, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:01:43,155 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3361, 2.3071, 1.9948, 2.5116, 2.8721, 2.2819, 2.1086, 1.7623], + device='cuda:0'), covar=tensor([0.2010, 0.1722, 0.1741, 0.1431, 0.1536, 0.1092, 0.2007, 0.1885], + device='cuda:0'), in_proj_covar=tensor([0.0243, 0.0209, 0.0213, 0.0195, 0.0243, 0.0190, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:01:55,006 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133713.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:02:02,673 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4832, 2.2042, 2.6972, 4.4077, 3.2374, 3.0437, 1.1135, 3.6981], + device='cuda:0'), covar=tensor([0.1402, 0.1204, 0.1232, 0.0455, 0.0606, 0.1238, 0.1865, 0.0382], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:02:06,306 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133730.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 05:02:07,530 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:02:11,409 INFO [finetune.py:976] (0/7) Epoch 24, batch 2000, loss[loss=0.1439, simple_loss=0.2324, pruned_loss=0.02774, over 4930.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2439, pruned_loss=0.04983, over 955866.22 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:02:28,712 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.381e+01 1.373e+02 1.735e+02 2.243e+02 3.912e+02, threshold=3.469e+02, percent-clipped=2.0 +2023-03-27 05:02:54,158 INFO [finetune.py:976] (0/7) Epoch 24, batch 2050, loss[loss=0.2009, simple_loss=0.2586, pruned_loss=0.07162, over 4814.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2416, pruned_loss=0.04968, over 955460.41 frames. ], batch size: 41, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:03:23,229 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:03:25,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5794, 3.7079, 3.5249, 1.7997, 3.8142, 2.8376, 0.7788, 2.5123], + device='cuda:0'), covar=tensor([0.2408, 0.2701, 0.1699, 0.3604, 0.1132, 0.1048, 0.4800, 0.1657], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0178, 0.0161, 0.0129, 0.0160, 0.0123, 0.0148, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 05:03:26,946 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-27 05:03:27,929 INFO [finetune.py:976] (0/7) Epoch 24, batch 2100, loss[loss=0.1751, simple_loss=0.237, pruned_loss=0.05658, over 4886.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2409, pruned_loss=0.04984, over 956226.52 frames. ], batch size: 32, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:03:40,563 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-03-27 05:03:47,588 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.827e+01 1.541e+02 1.860e+02 2.293e+02 6.118e+02, threshold=3.720e+02, percent-clipped=3.0 +2023-03-27 05:04:11,228 INFO [finetune.py:976] (0/7) Epoch 24, batch 2150, loss[loss=0.1245, simple_loss=0.1869, pruned_loss=0.03107, over 3807.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2442, pruned_loss=0.05111, over 955518.98 frames. ], batch size: 16, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:04:29,015 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3543, 1.2660, 1.1874, 1.3871, 1.5746, 1.5112, 1.3467, 1.1792], + device='cuda:0'), covar=tensor([0.0330, 0.0315, 0.0627, 0.0288, 0.0261, 0.0447, 0.0354, 0.0398], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0145, 0.0112, 0.0101, 0.0113, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7931e-05, 8.1910e-05, 1.1356e-04, 8.5809e-05, 7.8318e-05, 8.3536e-05, + 7.6234e-05, 8.6019e-05], device='cuda:0') +2023-03-27 05:04:32,691 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8069, 0.7157, 1.8618, 1.7138, 1.6650, 1.6213, 1.6119, 1.8172], + device='cuda:0'), covar=tensor([0.3479, 0.3698, 0.3109, 0.3488, 0.4484, 0.3363, 0.4078, 0.2840], + device='cuda:0'), in_proj_covar=tensor([0.0259, 0.0244, 0.0264, 0.0288, 0.0287, 0.0264, 0.0295, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:04:36,804 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2402, 1.8823, 2.6804, 1.6310, 2.2233, 2.4837, 1.7969, 2.6133], + device='cuda:0'), covar=tensor([0.1306, 0.2047, 0.1394, 0.2165, 0.0919, 0.1506, 0.2697, 0.0814], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0209, 0.0192, 0.0192, 0.0174, 0.0215, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:04:44,943 INFO [finetune.py:976] (0/7) Epoch 24, batch 2200, loss[loss=0.2318, simple_loss=0.2953, pruned_loss=0.08409, over 4907.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2465, pruned_loss=0.05205, over 954531.85 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:04:47,499 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1639, 3.6007, 3.7972, 3.9081, 3.9458, 3.6866, 4.2142, 1.5702], + device='cuda:0'), covar=tensor([0.0683, 0.0835, 0.0802, 0.0905, 0.1059, 0.1455, 0.0636, 0.5303], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0247, 0.0281, 0.0293, 0.0337, 0.0287, 0.0306, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:05:02,713 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.464e+02 1.824e+02 2.239e+02 3.694e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-27 05:05:05,870 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 05:05:06,976 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7910, 1.7086, 1.4756, 1.9125, 2.2434, 1.9237, 1.4385, 1.4597], + device='cuda:0'), covar=tensor([0.2229, 0.2002, 0.1927, 0.1617, 0.1562, 0.1174, 0.2534, 0.1972], + device='cuda:0'), in_proj_covar=tensor([0.0243, 0.0209, 0.0213, 0.0195, 0.0243, 0.0189, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:05:24,728 INFO [finetune.py:976] (0/7) Epoch 24, batch 2250, loss[loss=0.1716, simple_loss=0.247, pruned_loss=0.04811, over 4892.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2483, pruned_loss=0.0532, over 952745.57 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:05:26,028 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.64 vs. limit=2.0 +2023-03-27 05:05:34,977 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133997.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:05:39,392 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-134000.pt +2023-03-27 05:05:49,822 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:04,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8078, 1.6497, 2.1153, 1.4170, 1.9338, 2.0785, 1.6081, 2.2797], + device='cuda:0'), covar=tensor([0.1240, 0.2096, 0.1223, 0.1753, 0.0919, 0.1400, 0.2745, 0.0809], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0209, 0.0193, 0.0191, 0.0174, 0.0215, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:06:07,534 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134030.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 05:06:09,749 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:12,682 INFO [finetune.py:976] (0/7) Epoch 24, batch 2300, loss[loss=0.1362, simple_loss=0.2132, pruned_loss=0.02954, over 4691.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.248, pruned_loss=0.05209, over 954570.85 frames. ], batch size: 23, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:06:27,069 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134058.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:31,056 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.533e+02 1.723e+02 2.089e+02 4.293e+02, threshold=3.445e+02, percent-clipped=2.0 +2023-03-27 05:06:32,439 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0239, 2.0417, 1.4163, 2.0579, 1.9994, 1.6343, 2.7489, 2.0459], + device='cuda:0'), covar=tensor([0.1480, 0.2021, 0.3303, 0.2893, 0.2622, 0.1831, 0.2266, 0.1773], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0191, 0.0237, 0.0256, 0.0251, 0.0208, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:06:40,133 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134078.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 05:06:41,359 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:41,507 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-27 05:06:46,524 INFO [finetune.py:976] (0/7) Epoch 24, batch 2350, loss[loss=0.1632, simple_loss=0.2371, pruned_loss=0.04467, over 4228.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2459, pruned_loss=0.05179, over 954053.52 frames. ], batch size: 66, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:06:57,200 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7897, 1.8235, 1.6719, 1.7518, 1.3985, 4.2671, 1.7747, 2.0414], + device='cuda:0'), covar=tensor([0.3057, 0.2310, 0.2081, 0.2355, 0.1662, 0.0131, 0.2381, 0.1162], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0123, 0.0113, 0.0097, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:07:02,516 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6474, 1.4332, 1.0898, 0.2938, 1.2888, 1.4478, 1.4084, 1.4339], + device='cuda:0'), covar=tensor([0.0832, 0.0820, 0.1269, 0.1839, 0.1234, 0.2185, 0.2316, 0.0843], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0191, 0.0199, 0.0180, 0.0209, 0.0208, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:07:04,249 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0459, 0.8796, 0.8702, 0.3376, 0.8693, 1.0230, 1.0612, 0.8049], + device='cuda:0'), covar=tensor([0.0775, 0.0596, 0.0564, 0.0571, 0.0581, 0.0652, 0.0482, 0.0666], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0147, 0.0126, 0.0121, 0.0130, 0.0129, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.8556e-05, 1.0588e-04, 9.0481e-05, 8.5399e-05, 9.1567e-05, 9.1797e-05, + 1.0019e-04, 1.0523e-04], device='cuda:0') +2023-03-27 05:07:09,037 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1551, 1.3987, 0.7904, 1.9645, 2.5046, 1.9969, 1.7578, 2.1057], + device='cuda:0'), covar=tensor([0.1317, 0.1987, 0.1990, 0.1114, 0.1758, 0.1698, 0.1368, 0.1830], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0093, 0.0109, 0.0091, 0.0118, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:07:14,978 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:07:19,176 INFO [finetune.py:976] (0/7) Epoch 24, batch 2400, loss[loss=0.1422, simple_loss=0.2129, pruned_loss=0.03573, over 4943.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2435, pruned_loss=0.05109, over 955829.18 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:07:38,332 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.553e+01 1.434e+02 1.789e+02 2.166e+02 3.942e+02, threshold=3.577e+02, percent-clipped=1.0 +2023-03-27 05:07:47,957 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:07:55,473 INFO [finetune.py:976] (0/7) Epoch 24, batch 2450, loss[loss=0.1802, simple_loss=0.2423, pruned_loss=0.05902, over 4902.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2402, pruned_loss=0.04987, over 956081.91 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:08:36,910 INFO [finetune.py:976] (0/7) Epoch 24, batch 2500, loss[loss=0.1844, simple_loss=0.2671, pruned_loss=0.0509, over 4943.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2411, pruned_loss=0.05006, over 957121.76 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:08:55,715 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.396e+01 1.499e+02 1.865e+02 2.171e+02 5.575e+02, threshold=3.730e+02, percent-clipped=1.0 +2023-03-27 05:09:03,579 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5225, 2.4336, 2.2267, 1.0705, 2.3678, 1.8149, 1.7357, 2.2994], + device='cuda:0'), covar=tensor([0.0900, 0.0769, 0.1484, 0.2029, 0.1304, 0.2625, 0.2309, 0.0928], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0190, 0.0199, 0.0179, 0.0209, 0.0208, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:09:20,384 INFO [finetune.py:976] (0/7) Epoch 24, batch 2550, loss[loss=0.1371, simple_loss=0.212, pruned_loss=0.03111, over 4785.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2447, pruned_loss=0.05074, over 955146.83 frames. ], batch size: 26, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:09:32,220 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134304.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:34,653 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134308.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:35,292 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134309.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:54,095 INFO [finetune.py:976] (0/7) Epoch 24, batch 2600, loss[loss=0.2008, simple_loss=0.28, pruned_loss=0.06083, over 4732.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2467, pruned_loss=0.05119, over 956478.50 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:10:04,825 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134353.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:07,201 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:12,059 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.471e+02 1.806e+02 2.184e+02 4.519e+02, threshold=3.612e+02, percent-clipped=2.0 +2023-03-27 05:10:12,825 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134365.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:16,847 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134370.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:19,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5295, 1.4250, 2.0752, 3.1009, 2.0553, 2.2178, 1.1479, 2.6013], + device='cuda:0'), covar=tensor([0.1886, 0.1502, 0.1219, 0.0564, 0.0877, 0.1489, 0.1676, 0.0509], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0114, 0.0132, 0.0162, 0.0100, 0.0136, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:10:29,856 INFO [finetune.py:976] (0/7) Epoch 24, batch 2650, loss[loss=0.1955, simple_loss=0.2492, pruned_loss=0.0709, over 4771.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2479, pruned_loss=0.05145, over 957348.95 frames. ], batch size: 27, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:10:54,485 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-03-27 05:11:21,024 INFO [finetune.py:976] (0/7) Epoch 24, batch 2700, loss[loss=0.1493, simple_loss=0.2185, pruned_loss=0.04004, over 4884.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2463, pruned_loss=0.05025, over 956602.28 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:11:39,163 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.416e+02 1.758e+02 2.159e+02 3.599e+02, threshold=3.516e+02, percent-clipped=0.0 +2023-03-27 05:11:54,602 INFO [finetune.py:976] (0/7) Epoch 24, batch 2750, loss[loss=0.1815, simple_loss=0.2546, pruned_loss=0.05425, over 4907.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2437, pruned_loss=0.04983, over 955728.74 frames. ], batch size: 37, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:12:27,873 INFO [finetune.py:976] (0/7) Epoch 24, batch 2800, loss[loss=0.1609, simple_loss=0.2348, pruned_loss=0.0435, over 4749.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2404, pruned_loss=0.0489, over 955597.29 frames. ], batch size: 27, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:12:46,115 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.456e+02 1.823e+02 2.196e+02 4.309e+02, threshold=3.645e+02, percent-clipped=3.0 +2023-03-27 05:12:59,885 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5830, 1.4032, 1.5643, 0.8308, 1.5193, 1.5378, 1.5800, 1.3583], + device='cuda:0'), covar=tensor([0.0575, 0.0848, 0.0637, 0.0958, 0.1014, 0.0731, 0.0634, 0.1331], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0136, 0.0139, 0.0119, 0.0126, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:13:01,598 INFO [finetune.py:976] (0/7) Epoch 24, batch 2850, loss[loss=0.1668, simple_loss=0.2458, pruned_loss=0.04392, over 4920.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2397, pruned_loss=0.04872, over 957377.79 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:13:40,091 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2248, 2.2548, 1.8519, 2.2562, 2.1191, 2.0920, 2.1342, 3.0135], + device='cuda:0'), covar=tensor([0.3592, 0.4526, 0.3342, 0.4098, 0.4714, 0.2602, 0.3989, 0.1637], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0264, 0.0235, 0.0277, 0.0257, 0.0228, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:13:45,376 INFO [finetune.py:976] (0/7) Epoch 24, batch 2900, loss[loss=0.1772, simple_loss=0.2575, pruned_loss=0.0485, over 4906.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2428, pruned_loss=0.05016, over 956604.35 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:13:50,393 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8679, 1.6549, 2.2826, 1.5175, 2.1598, 2.3115, 1.5720, 2.3821], + device='cuda:0'), covar=tensor([0.1395, 0.2062, 0.1475, 0.1923, 0.0878, 0.1307, 0.2635, 0.0846], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0207, 0.0191, 0.0190, 0.0172, 0.0213, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:13:56,288 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134653.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:00,528 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134660.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:03,944 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.559e+02 1.783e+02 2.063e+02 3.902e+02, threshold=3.566e+02, percent-clipped=1.0 +2023-03-27 05:14:04,023 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134665.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:13,729 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 05:14:20,897 INFO [finetune.py:976] (0/7) Epoch 24, batch 2950, loss[loss=0.2178, simple_loss=0.2914, pruned_loss=0.07212, over 4752.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2453, pruned_loss=0.05087, over 955407.72 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:14:28,680 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2524, 2.0902, 1.8425, 2.0045, 1.9528, 1.9729, 2.0082, 2.7945], + device='cuda:0'), covar=tensor([0.3803, 0.4256, 0.3245, 0.3805, 0.3767, 0.2499, 0.3680, 0.1707], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0263, 0.0234, 0.0276, 0.0256, 0.0227, 0.0254, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:14:37,438 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134701.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:48,755 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3695, 2.3602, 2.0082, 2.4591, 2.2921, 2.2053, 2.2595, 3.2057], + device='cuda:0'), covar=tensor([0.4015, 0.4945, 0.3376, 0.4300, 0.4112, 0.2525, 0.4044, 0.1772], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0264, 0.0235, 0.0277, 0.0257, 0.0227, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:14:59,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6191, 0.7678, 1.6877, 1.6380, 1.5343, 1.4567, 1.5526, 1.6390], + device='cuda:0'), covar=tensor([0.3564, 0.3622, 0.3132, 0.3136, 0.4279, 0.3309, 0.4022, 0.2859], + device='cuda:0'), in_proj_covar=tensor([0.0259, 0.0244, 0.0264, 0.0289, 0.0289, 0.0265, 0.0297, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:15:02,624 INFO [finetune.py:976] (0/7) Epoch 24, batch 3000, loss[loss=0.1641, simple_loss=0.2379, pruned_loss=0.04512, over 4765.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2482, pruned_loss=0.05227, over 955316.22 frames. ], batch size: 26, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:15:02,625 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 05:15:13,334 INFO [finetune.py:1010] (0/7) Epoch 24, validation: loss=0.1561, simple_loss=0.2251, pruned_loss=0.0436, over 2265189.00 frames. +2023-03-27 05:15:13,334 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6393MB +2023-03-27 05:15:31,258 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.527e+02 1.858e+02 2.241e+02 5.364e+02, threshold=3.716e+02, percent-clipped=3.0 +2023-03-27 05:15:48,108 INFO [finetune.py:976] (0/7) Epoch 24, batch 3050, loss[loss=0.1533, simple_loss=0.2154, pruned_loss=0.04561, over 3897.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2487, pruned_loss=0.05177, over 954731.49 frames. ], batch size: 17, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:16:14,467 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7739, 1.8597, 1.6785, 1.9091, 1.8704, 4.6985, 1.9035, 2.3887], + device='cuda:0'), covar=tensor([0.3625, 0.2636, 0.2193, 0.2566, 0.1532, 0.0130, 0.2306, 0.1200], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:16:14,476 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7349, 2.7797, 2.8039, 1.7612, 2.6557, 2.9328, 2.9340, 2.3402], + device='cuda:0'), covar=tensor([0.0628, 0.0617, 0.0663, 0.0914, 0.0656, 0.0723, 0.0626, 0.1207], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0120, 0.0127, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:16:37,801 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8563, 2.5659, 2.1252, 1.0489, 2.3443, 2.2153, 1.9597, 2.4380], + device='cuda:0'), covar=tensor([0.0844, 0.0775, 0.1580, 0.2122, 0.1297, 0.2160, 0.2028, 0.0931], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0190, 0.0198, 0.0180, 0.0209, 0.0208, 0.0223, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:16:39,366 INFO [finetune.py:976] (0/7) Epoch 24, batch 3100, loss[loss=0.1751, simple_loss=0.2416, pruned_loss=0.05426, over 4891.00 frames. ], tot_loss[loss=0.174, simple_loss=0.246, pruned_loss=0.05101, over 954159.21 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:16:42,486 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.21 vs. limit=5.0 +2023-03-27 05:16:53,068 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1722, 1.1800, 1.4119, 0.6366, 1.3714, 1.5660, 1.5403, 1.3466], + device='cuda:0'), covar=tensor([0.1121, 0.1055, 0.0678, 0.0638, 0.0670, 0.0803, 0.0528, 0.0866], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0125, 0.0120, 0.0129, 0.0128, 0.0139, 0.0146], + device='cuda:0'), out_proj_covar=tensor([8.7861e-05, 1.0541e-04, 8.9298e-05, 8.4687e-05, 9.0863e-05, 9.1283e-05, + 9.9402e-05, 1.0468e-04], device='cuda:0') +2023-03-27 05:16:56,163 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-27 05:16:56,677 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134862.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:16:58,376 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.835e+01 1.385e+02 1.668e+02 2.115e+02 5.080e+02, threshold=3.336e+02, percent-clipped=1.0 +2023-03-27 05:17:06,134 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6494, 1.4943, 2.3509, 3.2457, 2.1924, 2.4472, 1.4107, 2.6790], + device='cuda:0'), covar=tensor([0.1683, 0.1451, 0.1141, 0.0589, 0.0825, 0.1610, 0.1489, 0.0484], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:17:12,755 INFO [finetune.py:976] (0/7) Epoch 24, batch 3150, loss[loss=0.1795, simple_loss=0.2392, pruned_loss=0.05991, over 4839.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2421, pruned_loss=0.04944, over 954650.57 frames. ], batch size: 30, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:17:27,484 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2180, 1.9145, 2.4472, 1.6849, 2.2858, 2.5170, 1.9288, 2.5865], + device='cuda:0'), covar=tensor([0.1088, 0.1936, 0.1249, 0.1693, 0.0786, 0.0975, 0.2352, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0192, 0.0190, 0.0173, 0.0214, 0.0216, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:17:37,234 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134923.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:17:39,606 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4036, 1.9753, 2.3899, 2.4563, 2.1151, 2.1163, 2.3617, 2.3306], + device='cuda:0'), covar=tensor([0.3950, 0.4145, 0.3357, 0.3759, 0.4968, 0.3797, 0.4689, 0.2854], + device='cuda:0'), in_proj_covar=tensor([0.0260, 0.0245, 0.0266, 0.0291, 0.0290, 0.0266, 0.0297, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:17:46,592 INFO [finetune.py:976] (0/7) Epoch 24, batch 3200, loss[loss=0.1856, simple_loss=0.2453, pruned_loss=0.063, over 4928.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.239, pruned_loss=0.04875, over 953744.23 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:18:03,010 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134960.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:05,933 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.665e+01 1.513e+02 1.793e+02 2.252e+02 3.579e+02, threshold=3.586e+02, percent-clipped=1.0 +2023-03-27 05:18:06,029 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134965.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:22,479 INFO [finetune.py:976] (0/7) Epoch 24, batch 3250, loss[loss=0.1952, simple_loss=0.251, pruned_loss=0.06971, over 4147.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2395, pruned_loss=0.04918, over 955161.28 frames. ], batch size: 65, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:18:45,538 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:48,594 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135013.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:59,443 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4365, 1.3688, 1.3453, 1.4082, 0.8964, 2.8994, 1.0541, 1.4035], + device='cuda:0'), covar=tensor([0.3416, 0.2600, 0.2166, 0.2580, 0.2029, 0.0273, 0.2973, 0.1356], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:19:04,067 INFO [finetune.py:976] (0/7) Epoch 24, batch 3300, loss[loss=0.1791, simple_loss=0.2553, pruned_loss=0.05148, over 4867.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2446, pruned_loss=0.05077, over 954430.76 frames. ], batch size: 31, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:19:13,236 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-03-27 05:19:23,522 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.563e+02 1.899e+02 2.275e+02 5.700e+02, threshold=3.799e+02, percent-clipped=2.0 +2023-03-27 05:19:44,198 INFO [finetune.py:976] (0/7) Epoch 24, batch 3350, loss[loss=0.188, simple_loss=0.2612, pruned_loss=0.05743, over 4755.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2453, pruned_loss=0.05077, over 952399.58 frames. ], batch size: 27, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:20:21,440 INFO [finetune.py:976] (0/7) Epoch 24, batch 3400, loss[loss=0.2173, simple_loss=0.2813, pruned_loss=0.07662, over 4813.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2461, pruned_loss=0.05138, over 952136.53 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:20:23,967 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1374, 1.8346, 2.4737, 3.9137, 2.6736, 2.8426, 0.9721, 3.3080], + device='cuda:0'), covar=tensor([0.1700, 0.1547, 0.1536, 0.0745, 0.0810, 0.1637, 0.2226, 0.0449], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0114, 0.0132, 0.0162, 0.0100, 0.0135, 0.0123, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:20:40,369 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.584e+02 1.828e+02 2.150e+02 3.792e+02, threshold=3.656e+02, percent-clipped=0.0 +2023-03-27 05:20:44,615 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:20:53,769 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3287, 1.3178, 1.5801, 1.0590, 1.3898, 1.5083, 1.2915, 1.6026], + device='cuda:0'), covar=tensor([0.1072, 0.2201, 0.1226, 0.1429, 0.0877, 0.1079, 0.2825, 0.0803], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0190, 0.0189, 0.0172, 0.0212, 0.0214, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:20:54,276 INFO [finetune.py:976] (0/7) Epoch 24, batch 3450, loss[loss=0.1902, simple_loss=0.257, pruned_loss=0.06169, over 4915.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2461, pruned_loss=0.05128, over 951768.50 frames. ], batch size: 43, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:21:03,895 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.26 vs. limit=5.0 +2023-03-27 05:21:27,764 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:21:40,733 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135232.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:21:47,100 INFO [finetune.py:976] (0/7) Epoch 24, batch 3500, loss[loss=0.1534, simple_loss=0.2158, pruned_loss=0.04543, over 4866.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2423, pruned_loss=0.04975, over 953150.01 frames. ], batch size: 31, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:22:06,078 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.678e+01 1.506e+02 1.714e+02 2.011e+02 3.544e+02, threshold=3.428e+02, percent-clipped=0.0 +2023-03-27 05:22:08,098 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6682, 1.5243, 1.3947, 1.5467, 1.8568, 1.7478, 1.5536, 1.3605], + device='cuda:0'), covar=tensor([0.0362, 0.0337, 0.0661, 0.0344, 0.0240, 0.0466, 0.0332, 0.0447], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0145, 0.0113, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7599e-05, 8.1580e-05, 1.1369e-04, 8.6203e-05, 7.8249e-05, 8.4485e-05, + 7.6719e-05, 8.5871e-05], device='cuda:0') +2023-03-27 05:22:12,076 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7566, 3.6946, 3.4954, 1.7269, 3.7779, 2.8264, 0.8259, 2.5425], + device='cuda:0'), covar=tensor([0.2188, 0.1777, 0.1494, 0.3325, 0.1027, 0.0956, 0.4445, 0.1405], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0180, 0.0161, 0.0129, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 05:22:20,453 INFO [finetune.py:976] (0/7) Epoch 24, batch 3550, loss[loss=0.1716, simple_loss=0.2369, pruned_loss=0.05316, over 4805.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2408, pruned_loss=0.04959, over 954183.70 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:22:54,386 INFO [finetune.py:976] (0/7) Epoch 24, batch 3600, loss[loss=0.1762, simple_loss=0.2345, pruned_loss=0.05893, over 4868.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2386, pruned_loss=0.04877, over 953787.40 frames. ], batch size: 31, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:23:09,354 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2803, 2.2248, 2.3661, 1.5715, 2.1669, 2.3704, 2.4512, 1.9109], + device='cuda:0'), covar=tensor([0.0617, 0.0682, 0.0681, 0.0895, 0.0763, 0.0704, 0.0584, 0.1101], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0136, 0.0140, 0.0119, 0.0127, 0.0138, 0.0138, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:23:11,687 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4696, 1.2087, 2.0497, 3.1014, 1.9691, 2.3173, 0.8323, 2.7304], + device='cuda:0'), covar=tensor([0.2053, 0.2084, 0.1541, 0.0934, 0.1082, 0.1569, 0.2317, 0.0559], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:23:12,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.699e+01 1.474e+02 1.759e+02 2.084e+02 3.295e+02, threshold=3.517e+02, percent-clipped=0.0 +2023-03-27 05:23:28,231 INFO [finetune.py:976] (0/7) Epoch 24, batch 3650, loss[loss=0.1944, simple_loss=0.2681, pruned_loss=0.06035, over 4820.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2418, pruned_loss=0.05013, over 953718.16 frames. ], batch size: 39, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:24:11,249 INFO [finetune.py:976] (0/7) Epoch 24, batch 3700, loss[loss=0.1533, simple_loss=0.2353, pruned_loss=0.03562, over 4766.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2436, pruned_loss=0.05022, over 951811.65 frames. ], batch size: 28, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:24:28,521 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.118e+01 1.614e+02 1.999e+02 2.429e+02 5.138e+02, threshold=3.998e+02, percent-clipped=6.0 +2023-03-27 05:24:32,172 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0657, 1.9823, 1.6556, 1.9083, 1.8397, 1.8488, 1.8950, 2.6120], + device='cuda:0'), covar=tensor([0.3671, 0.4075, 0.3256, 0.3677, 0.3788, 0.2386, 0.3601, 0.1678], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0276, 0.0257, 0.0228, 0.0254, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:24:43,345 INFO [finetune.py:976] (0/7) Epoch 24, batch 3750, loss[loss=0.1711, simple_loss=0.2374, pruned_loss=0.05233, over 4844.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2453, pruned_loss=0.05086, over 953076.29 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:25:12,887 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135518.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:14,154 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0892, 1.8071, 2.3524, 1.5635, 2.2149, 2.3818, 1.7071, 2.4728], + device='cuda:0'), covar=tensor([0.1243, 0.2024, 0.1447, 0.1959, 0.0900, 0.1349, 0.2650, 0.0904], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0208, 0.0193, 0.0191, 0.0175, 0.0215, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:25:18,836 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135527.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:26,760 INFO [finetune.py:976] (0/7) Epoch 24, batch 3800, loss[loss=0.1601, simple_loss=0.2437, pruned_loss=0.03823, over 4851.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2469, pruned_loss=0.05098, over 952491.48 frames. ], batch size: 44, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:25:44,706 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.524e+02 1.815e+02 2.221e+02 4.659e+02, threshold=3.630e+02, percent-clipped=3.0 +2023-03-27 05:25:45,382 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:52,195 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 05:26:00,453 INFO [finetune.py:976] (0/7) Epoch 24, batch 3850, loss[loss=0.1522, simple_loss=0.2221, pruned_loss=0.04116, over 4700.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2458, pruned_loss=0.05025, over 952247.90 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:26:20,880 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6708, 1.5144, 1.0876, 0.3494, 1.3390, 1.4798, 1.3806, 1.4510], + device='cuda:0'), covar=tensor([0.0855, 0.0756, 0.1249, 0.1791, 0.1194, 0.2139, 0.2248, 0.0807], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0189, 0.0197, 0.0178, 0.0207, 0.0206, 0.0220, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:26:38,615 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135630.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:26:45,798 INFO [finetune.py:976] (0/7) Epoch 24, batch 3900, loss[loss=0.1789, simple_loss=0.253, pruned_loss=0.05238, over 4820.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2428, pruned_loss=0.04962, over 954802.22 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:27:10,711 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.672e+01 1.400e+02 1.667e+02 1.961e+02 4.314e+02, threshold=3.334e+02, percent-clipped=1.0 +2023-03-27 05:27:26,034 INFO [finetune.py:976] (0/7) Epoch 24, batch 3950, loss[loss=0.1195, simple_loss=0.1975, pruned_loss=0.0207, over 4774.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2403, pruned_loss=0.04911, over 955950.13 frames. ], batch size: 28, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:27:28,546 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:27:43,664 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2534, 1.3656, 1.5158, 1.0927, 1.2431, 1.5211, 1.3386, 1.6470], + device='cuda:0'), covar=tensor([0.1210, 0.1881, 0.1296, 0.1446, 0.1056, 0.1152, 0.2661, 0.0872], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0207, 0.0191, 0.0190, 0.0173, 0.0213, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:27:58,422 INFO [finetune.py:976] (0/7) Epoch 24, batch 4000, loss[loss=0.1611, simple_loss=0.2318, pruned_loss=0.04518, over 4752.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.241, pruned_loss=0.05005, over 954826.03 frames. ], batch size: 28, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:28:16,421 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.536e+01 1.548e+02 1.897e+02 2.315e+02 3.943e+02, threshold=3.793e+02, percent-clipped=5.0 +2023-03-27 05:28:31,229 INFO [finetune.py:976] (0/7) Epoch 24, batch 4050, loss[loss=0.184, simple_loss=0.2644, pruned_loss=0.05177, over 4846.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.246, pruned_loss=0.05242, over 952641.31 frames. ], batch size: 44, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:28:35,523 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4504, 2.2507, 1.6437, 0.9097, 1.8396, 1.9174, 1.8123, 2.0117], + device='cuda:0'), covar=tensor([0.1011, 0.0748, 0.1850, 0.2059, 0.1510, 0.2463, 0.2276, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0189, 0.0198, 0.0179, 0.0209, 0.0208, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:28:38,930 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.9037, 0.8648, 0.7833, 0.9659, 1.0231, 0.9877, 0.8688, 0.7736], + device='cuda:0'), covar=tensor([0.0435, 0.0267, 0.0563, 0.0266, 0.0262, 0.0364, 0.0306, 0.0365], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7655e-05, 8.1767e-05, 1.1436e-04, 8.5986e-05, 7.8522e-05, 8.4847e-05, + 7.6266e-05, 8.6120e-05], device='cuda:0') +2023-03-27 05:28:49,473 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5461, 2.2589, 2.8583, 1.7456, 2.4418, 2.7767, 2.0959, 2.9400], + device='cuda:0'), covar=tensor([0.1341, 0.2018, 0.1397, 0.2106, 0.1050, 0.1451, 0.2480, 0.0700], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0191, 0.0189, 0.0173, 0.0212, 0.0215, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:28:59,164 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:29:05,394 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3354, 1.5148, 1.6070, 0.7568, 1.6179, 1.8347, 1.8757, 1.4412], + device='cuda:0'), covar=tensor([0.0992, 0.0688, 0.0556, 0.0567, 0.0490, 0.0636, 0.0336, 0.0835], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0146, 0.0124, 0.0119, 0.0129, 0.0127, 0.0138, 0.0146], + device='cuda:0'), out_proj_covar=tensor([8.7707e-05, 1.0508e-04, 8.8607e-05, 8.3924e-05, 9.0415e-05, 9.0493e-05, + 9.8766e-05, 1.0411e-04], device='cuda:0') +2023-03-27 05:29:09,952 INFO [finetune.py:976] (0/7) Epoch 24, batch 4100, loss[loss=0.1589, simple_loss=0.2343, pruned_loss=0.04176, over 4779.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2479, pruned_loss=0.05258, over 952018.98 frames. ], batch size: 29, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:29:26,823 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-03-27 05:29:28,033 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 05:29:30,373 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8016, 1.2541, 1.8597, 1.8489, 1.6477, 1.5824, 1.7825, 1.7469], + device='cuda:0'), covar=tensor([0.3871, 0.3724, 0.3218, 0.3183, 0.4494, 0.3668, 0.4269, 0.3036], + device='cuda:0'), in_proj_covar=tensor([0.0260, 0.0244, 0.0265, 0.0290, 0.0289, 0.0266, 0.0296, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:29:32,641 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.978e+01 1.562e+02 1.866e+02 2.353e+02 4.250e+02, threshold=3.731e+02, percent-clipped=2.0 +2023-03-27 05:29:39,223 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135875.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:29:46,949 INFO [finetune.py:976] (0/7) Epoch 24, batch 4150, loss[loss=0.1261, simple_loss=0.2059, pruned_loss=0.02319, over 4795.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2483, pruned_loss=0.05173, over 953958.43 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:29:52,543 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 05:30:00,931 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.9969, 1.9647, 2.0702, 0.8868, 2.3328, 2.5686, 2.1664, 1.9423], + device='cuda:0'), covar=tensor([0.0984, 0.0778, 0.0532, 0.0754, 0.0523, 0.0693, 0.0572, 0.0780], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0125, 0.0120, 0.0130, 0.0128, 0.0139, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.8068e-05, 1.0564e-04, 8.9188e-05, 8.4334e-05, 9.1070e-05, 9.0898e-05, + 9.9351e-05, 1.0483e-04], device='cuda:0') +2023-03-27 05:30:19,800 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6750, 0.8007, 1.7428, 1.6928, 1.5443, 1.5216, 1.6119, 1.6609], + device='cuda:0'), covar=tensor([0.3396, 0.3882, 0.3271, 0.3425, 0.4546, 0.3566, 0.4025, 0.3050], + device='cuda:0'), in_proj_covar=tensor([0.0260, 0.0244, 0.0264, 0.0290, 0.0289, 0.0266, 0.0295, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:30:30,442 INFO [finetune.py:976] (0/7) Epoch 24, batch 4200, loss[loss=0.1604, simple_loss=0.2288, pruned_loss=0.04601, over 4852.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2488, pruned_loss=0.05186, over 955512.75 frames. ], batch size: 31, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:30:49,318 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.053e+01 1.587e+02 1.796e+02 2.438e+02 3.967e+02, threshold=3.591e+02, percent-clipped=1.0 +2023-03-27 05:31:00,631 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:03,038 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135986.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:03,577 INFO [finetune.py:976] (0/7) Epoch 24, batch 4250, loss[loss=0.1581, simple_loss=0.226, pruned_loss=0.04516, over 4930.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2461, pruned_loss=0.05125, over 957065.03 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:31:08,461 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:12,221 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-136000.pt +2023-03-27 05:31:45,375 INFO [finetune.py:976] (0/7) Epoch 24, batch 4300, loss[loss=0.1595, simple_loss=0.2319, pruned_loss=0.0435, over 4866.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2421, pruned_loss=0.04945, over 958528.38 frames. ], batch size: 44, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:31:49,201 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136043.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:32:02,576 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136055.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:32:06,222 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 05:32:14,139 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.579e+01 1.491e+02 1.827e+02 2.181e+02 5.621e+02, threshold=3.653e+02, percent-clipped=1.0 +2023-03-27 05:32:31,270 INFO [finetune.py:976] (0/7) Epoch 24, batch 4350, loss[loss=0.1998, simple_loss=0.2622, pruned_loss=0.06874, over 4863.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2405, pruned_loss=0.04887, over 959079.53 frames. ], batch size: 31, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:04,533 INFO [finetune.py:976] (0/7) Epoch 24, batch 4400, loss[loss=0.2066, simple_loss=0.2719, pruned_loss=0.07063, over 4899.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2401, pruned_loss=0.04928, over 956994.25 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:05,258 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5955, 1.1456, 0.8667, 1.4449, 1.9877, 1.0736, 1.3093, 1.5080], + device='cuda:0'), covar=tensor([0.1406, 0.2060, 0.1836, 0.1169, 0.1945, 0.2023, 0.1422, 0.1875], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 05:33:08,165 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:33:23,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.540e+02 1.819e+02 2.170e+02 3.954e+02, threshold=3.638e+02, percent-clipped=3.0 +2023-03-27 05:33:37,770 INFO [finetune.py:976] (0/7) Epoch 24, batch 4450, loss[loss=0.2248, simple_loss=0.298, pruned_loss=0.07585, over 4809.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2441, pruned_loss=0.05091, over 953777.09 frames. ], batch size: 51, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:46,175 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4107, 2.2980, 1.9621, 1.0649, 2.1913, 1.8047, 1.7098, 2.2016], + device='cuda:0'), covar=tensor([0.0873, 0.0745, 0.1501, 0.1933, 0.1266, 0.2236, 0.2293, 0.0923], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0191, 0.0199, 0.0180, 0.0209, 0.0208, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:33:48,777 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136203.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:34:13,601 INFO [finetune.py:976] (0/7) Epoch 24, batch 4500, loss[loss=0.1888, simple_loss=0.2594, pruned_loss=0.05906, over 4812.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2468, pruned_loss=0.05158, over 954754.40 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:34:13,792 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 05:34:39,505 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.509e+02 1.852e+02 2.239e+02 3.856e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-27 05:34:54,254 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:34:54,760 INFO [finetune.py:976] (0/7) Epoch 24, batch 4550, loss[loss=0.2033, simple_loss=0.2824, pruned_loss=0.06211, over 4817.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2476, pruned_loss=0.05176, over 954973.77 frames. ], batch size: 47, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:35:28,208 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136334.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:28,898 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:30,005 INFO [finetune.py:976] (0/7) Epoch 24, batch 4600, loss[loss=0.1971, simple_loss=0.2672, pruned_loss=0.06349, over 4886.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2473, pruned_loss=0.05172, over 953872.26 frames. ], batch size: 35, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:35:35,223 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136338.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:45,770 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:56,270 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.624e+02 1.856e+02 2.259e+02 4.732e+02, threshold=3.713e+02, percent-clipped=2.0 +2023-03-27 05:36:11,519 INFO [finetune.py:976] (0/7) Epoch 24, batch 4650, loss[loss=0.1855, simple_loss=0.2521, pruned_loss=0.05942, over 4872.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.246, pruned_loss=0.05188, over 954031.21 frames. ], batch size: 34, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:36:17,134 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136396.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:36:34,529 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7848, 2.6106, 2.1884, 1.0271, 2.3786, 2.0941, 1.9929, 2.4321], + device='cuda:0'), covar=tensor([0.0906, 0.0763, 0.1705, 0.2183, 0.1383, 0.2289, 0.2093, 0.1019], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0193, 0.0202, 0.0183, 0.0213, 0.0211, 0.0226, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:36:45,425 INFO [finetune.py:976] (0/7) Epoch 24, batch 4700, loss[loss=0.1482, simple_loss=0.2178, pruned_loss=0.03933, over 4863.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2438, pruned_loss=0.05135, over 953319.00 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:37:13,734 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.479e+02 1.754e+02 2.064e+02 3.231e+02, threshold=3.507e+02, percent-clipped=0.0 +2023-03-27 05:37:38,205 INFO [finetune.py:976] (0/7) Epoch 24, batch 4750, loss[loss=0.1479, simple_loss=0.2181, pruned_loss=0.03881, over 4816.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.242, pruned_loss=0.05066, over 954450.34 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:37:39,018 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 05:37:44,917 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:38:02,852 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8734, 1.3802, 1.8683, 1.8080, 1.5966, 1.6179, 1.7319, 1.7698], + device='cuda:0'), covar=tensor([0.4329, 0.4046, 0.3247, 0.3844, 0.4928, 0.4133, 0.4628, 0.3139], + device='cuda:0'), in_proj_covar=tensor([0.0261, 0.0244, 0.0265, 0.0290, 0.0289, 0.0266, 0.0296, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:38:10,346 INFO [finetune.py:976] (0/7) Epoch 24, batch 4800, loss[loss=0.2003, simple_loss=0.2824, pruned_loss=0.05905, over 4847.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2429, pruned_loss=0.05063, over 954633.05 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:38:17,001 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8834, 1.2470, 1.8862, 1.8704, 1.6733, 1.6387, 1.8227, 1.7286], + device='cuda:0'), covar=tensor([0.3720, 0.3988, 0.3160, 0.3352, 0.4768, 0.3558, 0.4192, 0.3160], + device='cuda:0'), in_proj_covar=tensor([0.0260, 0.0244, 0.0264, 0.0289, 0.0289, 0.0266, 0.0296, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:38:28,976 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.582e+02 2.021e+02 2.347e+02 5.093e+02, threshold=4.042e+02, percent-clipped=3.0 +2023-03-27 05:38:30,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9749, 1.7933, 2.2458, 1.4229, 1.9708, 2.2153, 1.7286, 2.3508], + device='cuda:0'), covar=tensor([0.1370, 0.2211, 0.1374, 0.2061, 0.1019, 0.1428, 0.2704, 0.1011], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0191, 0.0190, 0.0174, 0.0214, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:38:41,928 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4542, 2.2966, 2.0458, 2.5882, 2.9156, 2.4745, 2.4749, 1.8970], + device='cuda:0'), covar=tensor([0.2028, 0.1863, 0.1767, 0.1410, 0.1545, 0.0999, 0.1753, 0.1844], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0196, 0.0244, 0.0191, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:38:44,076 INFO [finetune.py:976] (0/7) Epoch 24, batch 4850, loss[loss=0.1513, simple_loss=0.2179, pruned_loss=0.04231, over 4736.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2453, pruned_loss=0.05093, over 953859.18 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 32.0 +2023-03-27 05:38:45,536 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-03-27 05:38:48,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1578, 2.1333, 2.2401, 1.5748, 2.1528, 2.3251, 2.3882, 1.7990], + device='cuda:0'), covar=tensor([0.0561, 0.0595, 0.0604, 0.0778, 0.0647, 0.0654, 0.0493, 0.1061], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0135, 0.0138, 0.0118, 0.0125, 0.0137, 0.0136, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:38:58,735 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7266, 1.6618, 1.5797, 1.6713, 1.2706, 4.1853, 1.5837, 1.9935], + device='cuda:0'), covar=tensor([0.3179, 0.2459, 0.2058, 0.2350, 0.1708, 0.0119, 0.2500, 0.1207], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0122, 0.0124, 0.0114, 0.0097, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:39:05,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-27 05:39:17,546 INFO [finetune.py:976] (0/7) Epoch 24, batch 4900, loss[loss=0.1993, simple_loss=0.2712, pruned_loss=0.06366, over 4904.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2458, pruned_loss=0.05056, over 951498.15 frames. ], batch size: 36, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:39:18,276 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:21,754 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:26,549 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136650.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:42,308 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.600e+02 1.925e+02 2.438e+02 3.559e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-27 05:39:54,858 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 05:39:57,216 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-27 05:39:59,908 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136686.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:00,476 INFO [finetune.py:976] (0/7) Epoch 24, batch 4950, loss[loss=0.1259, simple_loss=0.1949, pruned_loss=0.02846, over 4080.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2461, pruned_loss=0.05034, over 950795.30 frames. ], batch size: 17, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:40:03,943 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:05,836 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6889, 1.5193, 1.0829, 0.3165, 1.3022, 1.4975, 1.4524, 1.3973], + device='cuda:0'), covar=tensor([0.0913, 0.0887, 0.1456, 0.2024, 0.1357, 0.2515, 0.2460, 0.0949], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0190, 0.0198, 0.0180, 0.0209, 0.0208, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:40:08,761 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136698.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:12,472 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:33,840 INFO [finetune.py:976] (0/7) Epoch 24, batch 5000, loss[loss=0.1685, simple_loss=0.234, pruned_loss=0.05145, over 4926.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.245, pruned_loss=0.05022, over 948892.35 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:40:54,381 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0598, 3.5388, 3.6894, 3.9133, 3.8151, 3.6128, 4.1536, 1.3655], + device='cuda:0'), covar=tensor([0.0842, 0.0881, 0.1014, 0.1098, 0.1276, 0.1601, 0.0799, 0.5645], + device='cuda:0'), in_proj_covar=tensor([0.0346, 0.0245, 0.0282, 0.0292, 0.0337, 0.0284, 0.0305, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:41:02,611 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.520e+02 1.782e+02 2.173e+02 3.913e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-03-27 05:41:17,086 INFO [finetune.py:976] (0/7) Epoch 24, batch 5050, loss[loss=0.184, simple_loss=0.2538, pruned_loss=0.05708, over 4905.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.243, pruned_loss=0.04974, over 948667.10 frames. ], batch size: 37, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:41:25,199 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136798.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:30,672 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:49,357 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8695, 2.0613, 1.6868, 1.7586, 2.4190, 2.3851, 1.9699, 1.8831], + device='cuda:0'), covar=tensor([0.0427, 0.0400, 0.0545, 0.0377, 0.0246, 0.0614, 0.0448, 0.0451], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8200e-05, 8.1941e-05, 1.1409e-04, 8.5861e-05, 7.8592e-05, 8.4392e-05, + 7.6503e-05, 8.6301e-05], device='cuda:0') +2023-03-27 05:41:49,839 INFO [finetune.py:976] (0/7) Epoch 24, batch 5100, loss[loss=0.1392, simple_loss=0.2104, pruned_loss=0.03399, over 4903.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2396, pruned_loss=0.04869, over 951348.11 frames. ], batch size: 36, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:41:56,295 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136846.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:56,511 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-27 05:42:11,831 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.576e+02 1.921e+02 2.257e+02 4.191e+02, threshold=3.841e+02, percent-clipped=1.0 +2023-03-27 05:42:13,194 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136867.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:42:35,177 INFO [finetune.py:976] (0/7) Epoch 24, batch 5150, loss[loss=0.1894, simple_loss=0.2685, pruned_loss=0.05514, over 4758.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.24, pruned_loss=0.04892, over 952098.46 frames. ], batch size: 59, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:16,539 INFO [finetune.py:976] (0/7) Epoch 24, batch 5200, loss[loss=0.1798, simple_loss=0.2608, pruned_loss=0.04944, over 4910.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2439, pruned_loss=0.05019, over 952594.85 frames. ], batch size: 37, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:21,970 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5595, 1.5064, 1.4878, 1.5159, 1.2761, 3.4580, 1.4749, 1.7965], + device='cuda:0'), covar=tensor([0.3329, 0.2551, 0.2205, 0.2461, 0.1705, 0.0194, 0.2652, 0.1292], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:43:35,514 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.647e+02 1.940e+02 2.397e+02 3.428e+02, threshold=3.879e+02, percent-clipped=0.0 +2023-03-27 05:43:48,852 INFO [finetune.py:976] (0/7) Epoch 24, batch 5250, loss[loss=0.1535, simple_loss=0.2344, pruned_loss=0.03635, over 4759.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.247, pruned_loss=0.05114, over 952308.39 frames. ], batch size: 27, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:51,871 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136991.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:43:57,218 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136999.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:44:22,622 INFO [finetune.py:976] (0/7) Epoch 24, batch 5300, loss[loss=0.1756, simple_loss=0.2468, pruned_loss=0.05217, over 4879.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2494, pruned_loss=0.05237, over 951768.33 frames. ], batch size: 31, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:44:23,935 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137039.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:44:37,560 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1907, 1.3080, 1.3784, 0.6792, 1.3214, 1.5534, 1.5717, 1.2500], + device='cuda:0'), covar=tensor([0.0807, 0.0546, 0.0582, 0.0449, 0.0509, 0.0605, 0.0292, 0.0636], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0122, 0.0131, 0.0130, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9494e-05, 1.0741e-04, 9.0594e-05, 8.5773e-05, 9.1951e-05, 9.2336e-05, + 1.0134e-04, 1.0616e-04], device='cuda:0') +2023-03-27 05:44:42,412 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.600e+02 1.832e+02 2.198e+02 3.821e+02, threshold=3.665e+02, percent-clipped=0.0 +2023-03-27 05:45:05,791 INFO [finetune.py:976] (0/7) Epoch 24, batch 5350, loss[loss=0.1129, simple_loss=0.1826, pruned_loss=0.02163, over 4417.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.25, pruned_loss=0.05262, over 950289.10 frames. ], batch size: 19, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:45:38,834 INFO [finetune.py:976] (0/7) Epoch 24, batch 5400, loss[loss=0.2028, simple_loss=0.2681, pruned_loss=0.06873, over 4889.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2471, pruned_loss=0.05181, over 951684.34 frames. ], batch size: 32, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:45:57,159 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=137162.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:45:58,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.720e+01 1.449e+02 1.770e+02 2.209e+02 4.288e+02, threshold=3.541e+02, percent-clipped=1.0 +2023-03-27 05:46:16,975 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-27 05:46:17,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3454, 2.2121, 1.8211, 2.1214, 2.2670, 2.0023, 2.4900, 2.3612], + device='cuda:0'), covar=tensor([0.1368, 0.2026, 0.3028, 0.2341, 0.2411, 0.1601, 0.2708, 0.1638], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0253, 0.0249, 0.0205, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:46:22,349 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2627, 2.2305, 2.3055, 1.5224, 2.2072, 2.4650, 2.3369, 1.9318], + device='cuda:0'), covar=tensor([0.0588, 0.0605, 0.0660, 0.0861, 0.0613, 0.0691, 0.0605, 0.1088], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0134, 0.0136, 0.0116, 0.0123, 0.0135, 0.0135, 0.0157], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:46:22,823 INFO [finetune.py:976] (0/7) Epoch 24, batch 5450, loss[loss=0.1781, simple_loss=0.2413, pruned_loss=0.05747, over 4834.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2438, pruned_loss=0.05048, over 952490.60 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:46:52,450 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2898, 2.0956, 2.1600, 1.0053, 2.5348, 2.6578, 2.3529, 1.9444], + device='cuda:0'), covar=tensor([0.0853, 0.0742, 0.0552, 0.0720, 0.0501, 0.0614, 0.0401, 0.0844], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0128, 0.0123, 0.0132, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([9.0207e-05, 1.0810e-04, 9.1438e-05, 8.6397e-05, 9.2903e-05, 9.3140e-05, + 1.0202e-04, 1.0724e-04], device='cuda:0') +2023-03-27 05:46:53,073 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9688, 3.0670, 2.9357, 2.1636, 2.7966, 3.2964, 3.3156, 2.5508], + device='cuda:0'), covar=tensor([0.0605, 0.0576, 0.0644, 0.0814, 0.0712, 0.0694, 0.0493, 0.1017], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0134, 0.0136, 0.0117, 0.0124, 0.0136, 0.0136, 0.0158], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:46:55,997 INFO [finetune.py:976] (0/7) Epoch 24, batch 5500, loss[loss=0.2323, simple_loss=0.2725, pruned_loss=0.09607, over 4522.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2406, pruned_loss=0.04991, over 952382.35 frames. ], batch size: 20, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:46:56,841 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 05:47:13,426 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.493e+02 1.897e+02 2.213e+02 3.719e+02, threshold=3.794e+02, percent-clipped=2.0 +2023-03-27 05:47:25,000 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.00 vs. limit=5.0 +2023-03-27 05:47:36,857 INFO [finetune.py:976] (0/7) Epoch 24, batch 5550, loss[loss=0.2023, simple_loss=0.2853, pruned_loss=0.05972, over 4832.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2442, pruned_loss=0.05217, over 951661.40 frames. ], batch size: 40, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:47:47,667 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:47:50,334 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 05:48:20,577 INFO [finetune.py:976] (0/7) Epoch 24, batch 5600, loss[loss=0.1613, simple_loss=0.2431, pruned_loss=0.03976, over 4830.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2465, pruned_loss=0.05221, over 952212.51 frames. ], batch size: 47, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:48:26,397 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:48:37,702 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.584e+02 1.843e+02 2.259e+02 3.753e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-27 05:48:51,115 INFO [finetune.py:976] (0/7) Epoch 24, batch 5650, loss[loss=0.1884, simple_loss=0.2611, pruned_loss=0.05785, over 4825.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2477, pruned_loss=0.05223, over 951599.43 frames. ], batch size: 33, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:48:51,895 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 05:49:19,333 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9361, 2.7474, 2.2340, 1.3514, 2.4733, 2.3072, 2.0768, 2.5148], + device='cuda:0'), covar=tensor([0.0687, 0.0657, 0.1491, 0.1799, 0.1160, 0.1996, 0.2130, 0.0776], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0200, 0.0181, 0.0210, 0.0210, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:49:20,947 INFO [finetune.py:976] (0/7) Epoch 24, batch 5700, loss[loss=0.1479, simple_loss=0.2053, pruned_loss=0.04526, over 4156.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2432, pruned_loss=0.05085, over 936073.84 frames. ], batch size: 18, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:49:35,730 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:49:37,615 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-24.pt +2023-03-27 05:49:52,032 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.393e+01 1.414e+02 1.686e+02 2.138e+02 3.465e+02, threshold=3.373e+02, percent-clipped=0.0 +2023-03-27 05:49:52,048 INFO [finetune.py:976] (0/7) Epoch 25, batch 0, loss[loss=0.1484, simple_loss=0.229, pruned_loss=0.03396, over 4828.00 frames. ], tot_loss[loss=0.1484, simple_loss=0.229, pruned_loss=0.03396, over 4828.00 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:49:52,049 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 05:49:58,461 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6680, 3.5779, 3.3575, 1.5187, 3.5881, 2.7282, 0.6798, 2.4470], + device='cuda:0'), covar=tensor([0.1858, 0.1699, 0.1489, 0.3348, 0.1203, 0.1020, 0.3798, 0.1523], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0129, 0.0161, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 05:50:06,683 INFO [finetune.py:1010] (0/7) Epoch 25, validation: loss=0.1587, simple_loss=0.2267, pruned_loss=0.04536, over 2265189.00 frames. +2023-03-27 05:50:06,683 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 05:50:46,518 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137510.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:50:50,024 INFO [finetune.py:976] (0/7) Epoch 25, batch 50, loss[loss=0.16, simple_loss=0.2302, pruned_loss=0.04494, over 4815.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2497, pruned_loss=0.0536, over 213717.22 frames. ], batch size: 33, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:50:51,850 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7619, 1.7615, 1.6687, 1.7869, 1.2870, 3.5903, 1.3807, 1.7603], + device='cuda:0'), covar=tensor([0.2861, 0.2308, 0.1968, 0.2175, 0.1579, 0.0194, 0.2385, 0.1190], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 05:50:53,711 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-03-27 05:51:06,081 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8385, 1.3316, 0.7649, 1.6507, 2.1348, 1.3798, 1.6140, 1.6825], + device='cuda:0'), covar=tensor([0.1372, 0.1940, 0.1956, 0.1118, 0.1815, 0.1855, 0.1259, 0.1861], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:51:09,921 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0148, 1.6724, 2.3915, 4.0117, 2.6202, 2.8553, 1.0382, 3.4222], + device='cuda:0'), covar=tensor([0.2023, 0.1827, 0.1595, 0.0731, 0.0960, 0.1452, 0.2059, 0.0535], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0101, 0.0136, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:51:12,268 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6965, 1.5188, 1.5167, 0.7272, 1.7000, 1.8549, 1.8031, 1.4449], + device='cuda:0'), covar=tensor([0.0753, 0.0588, 0.0467, 0.0544, 0.0415, 0.0614, 0.0332, 0.0627], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9609e-05, 1.0756e-04, 9.0843e-05, 8.6264e-05, 9.2124e-05, 9.2654e-05, + 1.0144e-04, 1.0627e-04], device='cuda:0') +2023-03-27 05:51:25,237 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.580e+02 1.851e+02 2.170e+02 4.183e+02, threshold=3.702e+02, percent-clipped=2.0 +2023-03-27 05:51:25,253 INFO [finetune.py:976] (0/7) Epoch 25, batch 100, loss[loss=0.1395, simple_loss=0.2094, pruned_loss=0.03485, over 4841.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2416, pruned_loss=0.05177, over 378264.75 frames. ], batch size: 44, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:51:59,265 INFO [finetune.py:976] (0/7) Epoch 25, batch 150, loss[loss=0.191, simple_loss=0.25, pruned_loss=0.06597, over 4813.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2401, pruned_loss=0.05218, over 506708.87 frames. ], batch size: 45, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:52:09,253 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-27 05:52:33,558 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.544e+02 1.791e+02 2.141e+02 4.771e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-27 05:52:33,574 INFO [finetune.py:976] (0/7) Epoch 25, batch 200, loss[loss=0.1344, simple_loss=0.2079, pruned_loss=0.03051, over 4000.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2376, pruned_loss=0.04992, over 604532.42 frames. ], batch size: 17, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:52:49,602 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 05:53:26,995 INFO [finetune.py:976] (0/7) Epoch 25, batch 250, loss[loss=0.2095, simple_loss=0.2659, pruned_loss=0.07654, over 4927.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2428, pruned_loss=0.05188, over 683145.96 frames. ], batch size: 33, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:53:31,851 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9236, 1.8574, 1.5648, 1.6286, 1.7497, 1.7368, 1.7544, 2.4289], + device='cuda:0'), covar=tensor([0.3678, 0.3794, 0.3106, 0.3533, 0.3890, 0.2313, 0.3479, 0.1658], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0236, 0.0277, 0.0260, 0.0229, 0.0257, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:54:00,388 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.518e+01 1.603e+02 1.998e+02 2.287e+02 4.515e+02, threshold=3.995e+02, percent-clipped=2.0 +2023-03-27 05:54:00,404 INFO [finetune.py:976] (0/7) Epoch 25, batch 300, loss[loss=0.1818, simple_loss=0.2561, pruned_loss=0.05373, over 4774.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2483, pruned_loss=0.05289, over 743324.73 frames. ], batch size: 28, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:54:01,149 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8194, 1.4117, 1.8732, 1.9016, 1.6581, 1.6348, 1.8441, 1.7594], + device='cuda:0'), covar=tensor([0.4321, 0.4162, 0.3200, 0.3711, 0.4938, 0.4001, 0.4438, 0.3001], + device='cuda:0'), in_proj_covar=tensor([0.0262, 0.0246, 0.0266, 0.0291, 0.0291, 0.0267, 0.0297, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 05:54:33,830 INFO [finetune.py:976] (0/7) Epoch 25, batch 350, loss[loss=0.1969, simple_loss=0.2643, pruned_loss=0.06473, over 4902.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2499, pruned_loss=0.05338, over 790918.60 frames. ], batch size: 36, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:55:00,501 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-27 05:55:07,123 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.541e+02 1.822e+02 2.129e+02 2.910e+02, threshold=3.644e+02, percent-clipped=0.0 +2023-03-27 05:55:07,139 INFO [finetune.py:976] (0/7) Epoch 25, batch 400, loss[loss=0.147, simple_loss=0.2231, pruned_loss=0.03545, over 4208.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2492, pruned_loss=0.05239, over 827778.24 frames. ], batch size: 66, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:55:31,421 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137892.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:55:33,184 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137895.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:55:53,569 INFO [finetune.py:976] (0/7) Epoch 25, batch 450, loss[loss=0.1732, simple_loss=0.2431, pruned_loss=0.05165, over 4828.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2472, pruned_loss=0.05153, over 857982.76 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:55:54,831 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3154, 1.6314, 0.6880, 1.9472, 2.4880, 1.8721, 1.8947, 1.9516], + device='cuda:0'), covar=tensor([0.1401, 0.2027, 0.2121, 0.1218, 0.1822, 0.1781, 0.1400, 0.2065], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0118, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 05:56:19,216 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137953.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:56:20,970 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137956.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:56:26,870 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.489e+02 1.801e+02 2.267e+02 5.324e+02, threshold=3.602e+02, percent-clipped=3.0 +2023-03-27 05:56:26,886 INFO [finetune.py:976] (0/7) Epoch 25, batch 500, loss[loss=0.1904, simple_loss=0.264, pruned_loss=0.05841, over 4812.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2452, pruned_loss=0.05069, over 879848.01 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:56:50,816 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-138000.pt +2023-03-27 05:56:57,123 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 05:56:58,235 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5718, 1.7002, 1.4006, 1.6444, 1.9784, 1.8591, 1.6379, 1.5105], + device='cuda:0'), covar=tensor([0.0372, 0.0313, 0.0619, 0.0302, 0.0219, 0.0459, 0.0347, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0106, 0.0145, 0.0112, 0.0101, 0.0114, 0.0102, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7783e-05, 8.1419e-05, 1.1367e-04, 8.5842e-05, 7.8596e-05, 8.3952e-05, + 7.5641e-05, 8.5857e-05], device='cuda:0') +2023-03-27 05:57:01,547 INFO [finetune.py:976] (0/7) Epoch 25, batch 550, loss[loss=0.1735, simple_loss=0.2424, pruned_loss=0.05235, over 4787.00 frames. ], tot_loss[loss=0.17, simple_loss=0.241, pruned_loss=0.04953, over 897881.98 frames. ], batch size: 29, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:57:34,657 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.805e+01 1.427e+02 1.740e+02 1.994e+02 3.808e+02, threshold=3.480e+02, percent-clipped=1.0 +2023-03-27 05:57:34,673 INFO [finetune.py:976] (0/7) Epoch 25, batch 600, loss[loss=0.2184, simple_loss=0.2826, pruned_loss=0.07708, over 4817.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2397, pruned_loss=0.04859, over 911902.81 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:58:07,364 INFO [finetune.py:976] (0/7) Epoch 25, batch 650, loss[loss=0.159, simple_loss=0.2304, pruned_loss=0.04378, over 4813.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2434, pruned_loss=0.04962, over 921455.55 frames. ], batch size: 25, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:58:34,860 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4171, 1.5723, 1.5921, 0.8719, 1.6981, 1.9359, 1.9696, 1.4195], + device='cuda:0'), covar=tensor([0.0921, 0.0588, 0.0614, 0.0562, 0.0458, 0.0554, 0.0300, 0.0697], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0122, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9281e-05, 1.0735e-04, 9.0633e-05, 8.5889e-05, 9.1829e-05, 9.2007e-05, + 1.0089e-04, 1.0580e-04], device='cuda:0') +2023-03-27 05:58:59,110 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.613e+02 1.882e+02 2.325e+02 5.163e+02, threshold=3.765e+02, percent-clipped=4.0 +2023-03-27 05:58:59,125 INFO [finetune.py:976] (0/7) Epoch 25, batch 700, loss[loss=0.1628, simple_loss=0.233, pruned_loss=0.04632, over 4760.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2441, pruned_loss=0.04951, over 929026.71 frames. ], batch size: 27, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:59:26,087 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 05:59:32,527 INFO [finetune.py:976] (0/7) Epoch 25, batch 750, loss[loss=0.1475, simple_loss=0.2339, pruned_loss=0.0306, over 4746.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2451, pruned_loss=0.04977, over 935154.15 frames. ], batch size: 27, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:59:53,523 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138248.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:59:55,842 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138251.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:05,193 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.386e+01 1.513e+02 1.803e+02 2.270e+02 6.862e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-27 06:00:05,209 INFO [finetune.py:976] (0/7) Epoch 25, batch 800, loss[loss=0.1653, simple_loss=0.2489, pruned_loss=0.04086, over 4865.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2449, pruned_loss=0.04943, over 939363.80 frames. ], batch size: 44, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:00:08,353 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138270.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:19,839 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-03-27 06:00:38,347 INFO [finetune.py:976] (0/7) Epoch 25, batch 850, loss[loss=0.1712, simple_loss=0.2423, pruned_loss=0.04998, over 4870.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2443, pruned_loss=0.04971, over 944710.05 frames. ], batch size: 34, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:00:44,520 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138322.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:45,786 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1846, 2.2163, 1.9711, 2.2810, 2.1292, 2.1644, 2.1255, 2.9901], + device='cuda:0'), covar=tensor([0.3571, 0.4320, 0.3099, 0.4034, 0.4399, 0.2278, 0.4263, 0.1430], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0264, 0.0235, 0.0276, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:00:54,532 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138331.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:01:24,849 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.388e+02 1.706e+02 2.091e+02 3.563e+02, threshold=3.412e+02, percent-clipped=0.0 +2023-03-27 06:01:24,865 INFO [finetune.py:976] (0/7) Epoch 25, batch 900, loss[loss=0.181, simple_loss=0.2409, pruned_loss=0.06054, over 4779.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2416, pruned_loss=0.04879, over 948060.31 frames. ], batch size: 28, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:01:34,650 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7807, 1.7093, 1.6699, 1.7168, 1.3701, 3.7729, 1.5144, 1.9212], + device='cuda:0'), covar=tensor([0.3271, 0.2493, 0.2087, 0.2395, 0.1677, 0.0189, 0.2443, 0.1281], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:01:36,365 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138383.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:01:55,430 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2345, 2.3528, 1.6524, 2.1667, 2.1497, 1.8494, 2.9022, 2.3062], + device='cuda:0'), covar=tensor([0.1401, 0.1849, 0.3291, 0.2881, 0.2821, 0.1823, 0.2370, 0.1788], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0191, 0.0236, 0.0256, 0.0251, 0.0207, 0.0216, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:01:57,678 INFO [finetune.py:976] (0/7) Epoch 25, batch 950, loss[loss=0.187, simple_loss=0.2652, pruned_loss=0.05442, over 4829.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2409, pruned_loss=0.04912, over 950446.78 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:02:03,943 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9476, 1.8154, 1.6415, 1.6813, 2.1843, 2.2131, 1.7764, 1.7052], + device='cuda:0'), covar=tensor([0.0361, 0.0380, 0.0564, 0.0387, 0.0203, 0.0383, 0.0421, 0.0450], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0113, 0.0102, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7886e-05, 8.1737e-05, 1.1419e-04, 8.5897e-05, 7.8502e-05, 8.3867e-05, + 7.5927e-05, 8.6010e-05], device='cuda:0') +2023-03-27 06:02:30,846 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.447e+02 1.781e+02 2.298e+02 4.571e+02, threshold=3.563e+02, percent-clipped=3.0 +2023-03-27 06:02:30,862 INFO [finetune.py:976] (0/7) Epoch 25, batch 1000, loss[loss=0.1766, simple_loss=0.2561, pruned_loss=0.0486, over 4812.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2427, pruned_loss=0.04988, over 952595.94 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:02:35,736 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9628, 1.9683, 1.6064, 1.8196, 1.8078, 1.8356, 1.8544, 2.5234], + device='cuda:0'), covar=tensor([0.3717, 0.3809, 0.3246, 0.3589, 0.3730, 0.2313, 0.3569, 0.1636], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0234, 0.0275, 0.0258, 0.0228, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:03:03,759 INFO [finetune.py:976] (0/7) Epoch 25, batch 1050, loss[loss=0.1466, simple_loss=0.2183, pruned_loss=0.03747, over 4158.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2438, pruned_loss=0.04988, over 953568.21 frames. ], batch size: 18, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:03:07,282 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5000, 2.4839, 2.6214, 1.4337, 2.9299, 3.0692, 2.7861, 2.3512], + device='cuda:0'), covar=tensor([0.1076, 0.0880, 0.0431, 0.0686, 0.0726, 0.0649, 0.0398, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0122, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9189e-05, 1.0684e-04, 9.0393e-05, 8.5855e-05, 9.1832e-05, 9.2029e-05, + 1.0062e-04, 1.0546e-04], device='cuda:0') +2023-03-27 06:03:25,372 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:03:27,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:03:39,079 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.570e+02 1.874e+02 2.177e+02 7.699e+02, threshold=3.747e+02, percent-clipped=3.0 +2023-03-27 06:03:39,095 INFO [finetune.py:976] (0/7) Epoch 25, batch 1100, loss[loss=0.2002, simple_loss=0.2719, pruned_loss=0.06424, over 4836.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2465, pruned_loss=0.05086, over 954795.14 frames. ], batch size: 47, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:04:17,867 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138596.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:04:19,632 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138599.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:04:29,870 INFO [finetune.py:976] (0/7) Epoch 25, batch 1150, loss[loss=0.1905, simple_loss=0.2619, pruned_loss=0.05951, over 4910.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.248, pruned_loss=0.05112, over 954181.38 frames. ], batch size: 33, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:04:39,031 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:01,251 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-27 06:05:03,419 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.832e+01 1.520e+02 1.733e+02 2.222e+02 3.582e+02, threshold=3.466e+02, percent-clipped=0.0 +2023-03-27 06:05:03,435 INFO [finetune.py:976] (0/7) Epoch 25, batch 1200, loss[loss=0.1806, simple_loss=0.2592, pruned_loss=0.05098, over 4822.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.246, pruned_loss=0.05006, over 954457.33 frames. ], batch size: 33, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:05:09,908 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 06:05:13,801 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138678.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:15,626 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138681.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:18,698 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138686.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:25,921 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138697.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:37,221 INFO [finetune.py:976] (0/7) Epoch 25, batch 1250, loss[loss=0.1656, simple_loss=0.2241, pruned_loss=0.05358, over 4868.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2439, pruned_loss=0.05031, over 955227.36 frames. ], batch size: 31, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:05:55,611 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138742.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:59,139 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138747.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:06:05,707 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138758.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:06:12,321 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.434e+02 1.700e+02 2.124e+02 3.591e+02, threshold=3.400e+02, percent-clipped=1.0 +2023-03-27 06:06:12,337 INFO [finetune.py:976] (0/7) Epoch 25, batch 1300, loss[loss=0.1828, simple_loss=0.2477, pruned_loss=0.05897, over 4914.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2415, pruned_loss=0.04987, over 954988.61 frames. ], batch size: 36, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:06:57,378 INFO [finetune.py:976] (0/7) Epoch 25, batch 1350, loss[loss=0.1973, simple_loss=0.2736, pruned_loss=0.06049, over 4938.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2412, pruned_loss=0.04993, over 953956.68 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:07:31,278 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.573e+02 1.999e+02 2.321e+02 4.595e+02, threshold=3.999e+02, percent-clipped=3.0 +2023-03-27 06:07:31,294 INFO [finetune.py:976] (0/7) Epoch 25, batch 1400, loss[loss=0.1937, simple_loss=0.2567, pruned_loss=0.06535, over 4776.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2431, pruned_loss=0.05037, over 952644.99 frames. ], batch size: 26, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:07:37,335 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5803, 1.4443, 1.3845, 0.8355, 1.5662, 1.6943, 1.7426, 1.3749], + device='cuda:0'), covar=tensor([0.0831, 0.0629, 0.0485, 0.0489, 0.0470, 0.0475, 0.0301, 0.0531], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0150, 0.0128, 0.0123, 0.0132, 0.0131, 0.0143, 0.0149], + device='cuda:0'), out_proj_covar=tensor([9.0075e-05, 1.0785e-04, 9.1346e-05, 8.6950e-05, 9.2839e-05, 9.2997e-05, + 1.0213e-04, 1.0668e-04], device='cuda:0') +2023-03-27 06:07:46,612 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4413, 1.3361, 1.7334, 1.7803, 1.4947, 3.1837, 1.3049, 1.4576], + device='cuda:0'), covar=tensor([0.1020, 0.1834, 0.1134, 0.0903, 0.1662, 0.0264, 0.1524, 0.1832], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:08:01,082 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138909.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:04,578 INFO [finetune.py:976] (0/7) Epoch 25, batch 1450, loss[loss=0.1997, simple_loss=0.2631, pruned_loss=0.06816, over 4813.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2456, pruned_loss=0.05068, over 953723.42 frames. ], batch size: 38, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:08:11,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6079, 2.5010, 2.0966, 1.0756, 2.2523, 1.9218, 1.9099, 2.2868], + device='cuda:0'), covar=tensor([0.0864, 0.0809, 0.1714, 0.2201, 0.1504, 0.2428, 0.2095, 0.1066], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0191, 0.0200, 0.0180, 0.0208, 0.0209, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:08:11,763 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138926.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:15,135 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0571, 0.9827, 1.0249, 0.4562, 0.9899, 1.1856, 1.2082, 1.0097], + device='cuda:0'), covar=tensor([0.0862, 0.0682, 0.0582, 0.0513, 0.0556, 0.0684, 0.0418, 0.0685], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0131, 0.0130, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9607e-05, 1.0709e-04, 9.0805e-05, 8.6405e-05, 9.2241e-05, 9.2393e-05, + 1.0150e-04, 1.0585e-04], device='cuda:0') +2023-03-27 06:08:23,885 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5072, 1.4176, 1.2177, 1.5743, 1.6079, 1.5620, 0.9774, 1.2259], + device='cuda:0'), covar=tensor([0.2447, 0.2183, 0.2303, 0.1792, 0.1676, 0.1402, 0.2818, 0.2125], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0213, 0.0216, 0.0199, 0.0245, 0.0192, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:08:38,070 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.768e+01 1.472e+02 1.793e+02 2.201e+02 3.947e+02, threshold=3.587e+02, percent-clipped=0.0 +2023-03-27 06:08:38,085 INFO [finetune.py:976] (0/7) Epoch 25, batch 1500, loss[loss=0.2016, simple_loss=0.268, pruned_loss=0.06759, over 4817.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2472, pruned_loss=0.0512, over 953187.98 frames. ], batch size: 38, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:08:41,667 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138970.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 06:08:44,018 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:46,498 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:23,189 INFO [finetune.py:976] (0/7) Epoch 25, batch 1550, loss[loss=0.1445, simple_loss=0.2129, pruned_loss=0.03802, over 4927.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2467, pruned_loss=0.05103, over 952403.01 frames. ], batch size: 42, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:09:34,901 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:46,771 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139037.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:50,283 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139042.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:57,446 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139053.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:10:02,891 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9602, 1.4625, 2.0179, 1.9853, 1.7630, 1.7242, 1.9473, 1.8884], + device='cuda:0'), covar=tensor([0.4056, 0.3600, 0.3086, 0.3376, 0.4534, 0.3594, 0.3998, 0.2987], + device='cuda:0'), in_proj_covar=tensor([0.0261, 0.0245, 0.0265, 0.0290, 0.0291, 0.0267, 0.0297, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:10:05,110 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.522e+01 1.375e+02 1.683e+02 2.077e+02 3.862e+02, threshold=3.366e+02, percent-clipped=3.0 +2023-03-27 06:10:05,126 INFO [finetune.py:976] (0/7) Epoch 25, batch 1600, loss[loss=0.228, simple_loss=0.2846, pruned_loss=0.08571, over 4252.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2438, pruned_loss=0.05041, over 953171.39 frames. ], batch size: 65, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:10:38,954 INFO [finetune.py:976] (0/7) Epoch 25, batch 1650, loss[loss=0.1532, simple_loss=0.2189, pruned_loss=0.04377, over 4832.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2401, pruned_loss=0.04893, over 952514.79 frames. ], batch size: 25, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:11:12,573 INFO [finetune.py:976] (0/7) Epoch 25, batch 1700, loss[loss=0.1713, simple_loss=0.2321, pruned_loss=0.05524, over 4901.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2382, pruned_loss=0.04817, over 953268.23 frames. ], batch size: 32, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:11:13,175 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.831e+01 1.435e+02 1.759e+02 2.193e+02 3.727e+02, threshold=3.518e+02, percent-clipped=3.0 +2023-03-27 06:11:21,160 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9913, 1.9490, 2.0778, 1.3592, 1.9765, 2.1415, 1.9806, 1.6583], + device='cuda:0'), covar=tensor([0.0558, 0.0687, 0.0630, 0.0828, 0.0757, 0.0669, 0.0622, 0.1156], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0137, 0.0140, 0.0120, 0.0127, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:11:56,392 INFO [finetune.py:976] (0/7) Epoch 25, batch 1750, loss[loss=0.1752, simple_loss=0.265, pruned_loss=0.04273, over 4751.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.241, pruned_loss=0.04911, over 953727.64 frames. ], batch size: 54, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:12:09,785 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 06:12:11,571 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0939, 2.1251, 1.7578, 2.1093, 2.0454, 2.0315, 2.0386, 2.8882], + device='cuda:0'), covar=tensor([0.3879, 0.4781, 0.3568, 0.4229, 0.4805, 0.2733, 0.4276, 0.1783], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0264, 0.0235, 0.0275, 0.0258, 0.0229, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:12:20,705 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:12:39,090 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 06:12:39,404 INFO [finetune.py:976] (0/7) Epoch 25, batch 1800, loss[loss=0.1431, simple_loss=0.2344, pruned_loss=0.02591, over 4818.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2431, pruned_loss=0.04967, over 951849.02 frames. ], batch size: 40, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:12:39,466 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139265.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:12:39,956 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.974e+01 1.553e+02 1.833e+02 2.131e+02 4.022e+02, threshold=3.667e+02, percent-clipped=3.0 +2023-03-27 06:12:46,678 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139276.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:12:55,157 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139289.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:12:57,120 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 06:13:02,269 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:02,338 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 06:13:13,308 INFO [finetune.py:976] (0/7) Epoch 25, batch 1850, loss[loss=0.2213, simple_loss=0.2951, pruned_loss=0.07377, over 4818.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2452, pruned_loss=0.05012, over 954717.42 frames. ], batch size: 40, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:13:27,397 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139337.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:27,444 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139337.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:13:30,852 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139342.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:34,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6359, 2.4346, 2.0016, 2.8913, 2.5736, 2.2302, 3.0991, 2.6202], + device='cuda:0'), covar=tensor([0.1264, 0.2176, 0.3028, 0.2288, 0.2547, 0.1651, 0.2534, 0.1878], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0189, 0.0234, 0.0253, 0.0249, 0.0204, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:13:36,152 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:38,380 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139353.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:45,593 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 06:13:46,544 INFO [finetune.py:976] (0/7) Epoch 25, batch 1900, loss[loss=0.1813, simple_loss=0.2605, pruned_loss=0.05102, over 4787.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2464, pruned_loss=0.05071, over 952606.07 frames. ], batch size: 25, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:13:47,140 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.553e+02 1.835e+02 2.150e+02 3.557e+02, threshold=3.671e+02, percent-clipped=0.0 +2023-03-27 06:13:59,663 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139385.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:03,122 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:10,502 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139401.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:19,883 INFO [finetune.py:976] (0/7) Epoch 25, batch 1950, loss[loss=0.1736, simple_loss=0.2415, pruned_loss=0.0528, over 4814.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2448, pruned_loss=0.05018, over 950540.83 frames. ], batch size: 25, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:15:11,616 INFO [finetune.py:976] (0/7) Epoch 25, batch 2000, loss[loss=0.13, simple_loss=0.1965, pruned_loss=0.03172, over 4060.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2425, pruned_loss=0.04954, over 950647.85 frames. ], batch size: 17, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:15:12,708 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.253e+01 1.362e+02 1.721e+02 2.187e+02 3.038e+02, threshold=3.442e+02, percent-clipped=0.0 +2023-03-27 06:15:15,219 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5608, 1.4775, 1.4294, 1.4924, 0.8990, 2.8368, 1.0091, 1.5015], + device='cuda:0'), covar=tensor([0.3248, 0.2435, 0.2077, 0.2442, 0.1867, 0.0279, 0.2791, 0.1232], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:15:38,522 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-27 06:15:45,238 INFO [finetune.py:976] (0/7) Epoch 25, batch 2050, loss[loss=0.1418, simple_loss=0.209, pruned_loss=0.03734, over 4829.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2397, pruned_loss=0.04895, over 951659.39 frames. ], batch size: 25, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:16:18,436 INFO [finetune.py:976] (0/7) Epoch 25, batch 2100, loss[loss=0.1841, simple_loss=0.2636, pruned_loss=0.05234, over 4807.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2405, pruned_loss=0.0496, over 953055.05 frames. ], batch size: 45, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:16:18,550 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139565.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:16:20,121 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.352e+01 1.449e+02 1.714e+02 2.109e+02 3.824e+02, threshold=3.428e+02, percent-clipped=2.0 +2023-03-27 06:16:22,145 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1622, 1.8402, 2.2231, 2.1676, 1.9010, 1.8872, 2.1567, 2.0311], + device='cuda:0'), covar=tensor([0.3985, 0.4144, 0.2919, 0.3703, 0.4752, 0.3856, 0.4572, 0.3105], + device='cuda:0'), in_proj_covar=tensor([0.0263, 0.0246, 0.0266, 0.0291, 0.0291, 0.0268, 0.0297, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:16:22,741 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139571.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:30,365 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139582.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:38,025 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139594.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:50,718 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139613.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:52,382 INFO [finetune.py:976] (0/7) Epoch 25, batch 2150, loss[loss=0.1908, simple_loss=0.2711, pruned_loss=0.0552, over 4831.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2435, pruned_loss=0.05049, over 953777.41 frames. ], batch size: 51, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:17:09,458 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.1061, 4.4730, 4.6865, 4.9460, 4.8706, 4.5256, 5.2404, 1.5214], + device='cuda:0'), covar=tensor([0.0687, 0.0870, 0.0888, 0.0831, 0.1088, 0.1801, 0.0541, 0.6207], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0249, 0.0284, 0.0298, 0.0340, 0.0288, 0.0308, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:17:09,460 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139632.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:17:09,491 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:19,885 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 06:17:21,659 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:27,566 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139645.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:35,169 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-27 06:17:43,824 INFO [finetune.py:976] (0/7) Epoch 25, batch 2200, loss[loss=0.2577, simple_loss=0.3206, pruned_loss=0.09741, over 4893.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2452, pruned_loss=0.05103, over 953148.00 frames. ], batch size: 35, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:17:45,448 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.512e+02 1.789e+02 2.111e+02 3.462e+02, threshold=3.578e+02, percent-clipped=1.0 +2023-03-27 06:18:03,279 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7502, 1.3632, 1.0194, 1.5760, 2.1109, 1.4076, 1.5247, 1.6684], + device='cuda:0'), covar=tensor([0.1499, 0.2024, 0.1663, 0.1182, 0.1923, 0.1896, 0.1421, 0.1940], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0109, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 06:18:17,068 INFO [finetune.py:976] (0/7) Epoch 25, batch 2250, loss[loss=0.1689, simple_loss=0.2555, pruned_loss=0.04115, over 4809.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2471, pruned_loss=0.05133, over 954063.26 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:18:46,193 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-03-27 06:18:50,822 INFO [finetune.py:976] (0/7) Epoch 25, batch 2300, loss[loss=0.1762, simple_loss=0.2459, pruned_loss=0.05321, over 4896.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.248, pruned_loss=0.05123, over 953715.48 frames. ], batch size: 36, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:18:52,010 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.533e+02 1.822e+02 2.118e+02 3.916e+02, threshold=3.645e+02, percent-clipped=1.0 +2023-03-27 06:18:53,806 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139769.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:18:56,839 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-27 06:19:06,804 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139789.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:16,872 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139804.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:18,113 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:23,929 INFO [finetune.py:976] (0/7) Epoch 25, batch 2350, loss[loss=0.1581, simple_loss=0.236, pruned_loss=0.04008, over 4867.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2465, pruned_loss=0.05129, over 953132.49 frames. ], batch size: 31, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:19:32,258 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0349, 1.8727, 1.6271, 1.6087, 1.7363, 1.7317, 1.8067, 2.4758], + device='cuda:0'), covar=tensor([0.3410, 0.3681, 0.3028, 0.3484, 0.3726, 0.2249, 0.3501, 0.1597], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0234, 0.0275, 0.0258, 0.0228, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:19:35,053 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:35,635 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8992, 1.5960, 2.3942, 1.4760, 2.0951, 2.1587, 1.4160, 2.3024], + device='cuda:0'), covar=tensor([0.1407, 0.2345, 0.1190, 0.1950, 0.1026, 0.1550, 0.3255, 0.0975], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0209, 0.0194, 0.0192, 0.0176, 0.0215, 0.0219, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:19:54,881 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:54,890 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:06,567 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-27 06:20:07,501 INFO [finetune.py:976] (0/7) Epoch 25, batch 2400, loss[loss=0.1638, simple_loss=0.2324, pruned_loss=0.0476, over 4849.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2439, pruned_loss=0.05076, over 953603.38 frames. ], batch size: 44, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:20:07,625 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:10,610 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.173e+01 1.419e+02 1.768e+02 2.081e+02 3.267e+02, threshold=3.536e+02, percent-clipped=0.0 +2023-03-27 06:20:10,736 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139867.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:35,935 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139894.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:35,993 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6086, 2.4170, 2.0215, 2.7739, 2.5076, 2.2641, 3.0659, 2.5947], + device='cuda:0'), covar=tensor([0.1264, 0.2217, 0.2953, 0.2515, 0.2444, 0.1573, 0.2724, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0192, 0.0238, 0.0256, 0.0252, 0.0207, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:20:47,283 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:50,038 INFO [finetune.py:976] (0/7) Epoch 25, batch 2450, loss[loss=0.1966, simple_loss=0.2578, pruned_loss=0.06765, over 4827.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2418, pruned_loss=0.0504, over 952644.07 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:20:57,354 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139927.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:01,387 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139932.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 06:21:03,882 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 06:21:05,440 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:05,483 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8300, 1.8426, 1.6705, 1.9693, 1.7530, 4.5107, 1.7860, 2.2856], + device='cuda:0'), covar=tensor([0.3239, 0.2376, 0.2046, 0.2215, 0.1369, 0.0142, 0.2325, 0.1125], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:21:08,384 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139942.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:10,175 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:22,671 INFO [finetune.py:976] (0/7) Epoch 25, batch 2500, loss[loss=0.1679, simple_loss=0.2446, pruned_loss=0.04567, over 4768.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2427, pruned_loss=0.05107, over 951917.31 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:21:24,359 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.523e+02 1.884e+02 2.422e+02 3.755e+02, threshold=3.768e+02, percent-clipped=3.0 +2023-03-27 06:21:32,355 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139980.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:42,127 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139993.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:45,777 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5360, 1.1240, 0.9324, 1.4210, 1.9196, 1.1246, 1.4185, 1.4445], + device='cuda:0'), covar=tensor([0.1669, 0.2220, 0.1834, 0.1281, 0.2069, 0.2017, 0.1443, 0.2054], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 06:21:46,439 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-140000.pt +2023-03-27 06:21:48,255 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1265, 1.4370, 0.8736, 1.8668, 2.3643, 1.7872, 1.7803, 1.8536], + device='cuda:0'), covar=tensor([0.1524, 0.2121, 0.2022, 0.1245, 0.1814, 0.1774, 0.1487, 0.1940], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 06:21:55,419 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 06:21:57,548 INFO [finetune.py:976] (0/7) Epoch 25, batch 2550, loss[loss=0.2057, simple_loss=0.2837, pruned_loss=0.06383, over 4898.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2442, pruned_loss=0.05077, over 950971.80 frames. ], batch size: 43, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:22:07,688 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2401, 2.2983, 2.0255, 2.4055, 2.1029, 4.8576, 2.0194, 2.6083], + device='cuda:0'), covar=tensor([0.2745, 0.2064, 0.1805, 0.1940, 0.1200, 0.0100, 0.2155, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:22:11,736 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 06:22:36,076 INFO [finetune.py:976] (0/7) Epoch 25, batch 2600, loss[loss=0.2103, simple_loss=0.2792, pruned_loss=0.07077, over 4851.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.246, pruned_loss=0.0514, over 950917.30 frames. ], batch size: 44, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:22:42,052 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.547e+02 1.969e+02 2.320e+02 4.703e+02, threshold=3.938e+02, percent-clipped=1.0 +2023-03-27 06:22:56,604 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-27 06:23:14,140 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140103.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:19,335 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140110.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:22,259 INFO [finetune.py:976] (0/7) Epoch 25, batch 2650, loss[loss=0.1974, simple_loss=0.2689, pruned_loss=0.06296, over 4899.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2472, pruned_loss=0.05128, over 951364.93 frames. ], batch size: 37, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:23:25,377 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7145, 1.3979, 2.1126, 1.3738, 1.8895, 1.9086, 1.3296, 2.1223], + device='cuda:0'), covar=tensor([0.1331, 0.2251, 0.1105, 0.1710, 0.0954, 0.1449, 0.2844, 0.0887], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0207, 0.0192, 0.0191, 0.0176, 0.0214, 0.0217, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:23:28,792 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:41,809 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:51,792 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140160.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:53,481 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140162.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:55,149 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:55,632 INFO [finetune.py:976] (0/7) Epoch 25, batch 2700, loss[loss=0.1462, simple_loss=0.2282, pruned_loss=0.03212, over 4741.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2457, pruned_loss=0.05018, over 951296.80 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:23:56,848 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.476e+02 1.708e+02 2.136e+02 4.297e+02, threshold=3.417e+02, percent-clipped=1.0 +2023-03-27 06:23:59,440 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:00,578 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3132, 3.7617, 3.9232, 4.0971, 4.1193, 3.8054, 4.3842, 1.3962], + device='cuda:0'), covar=tensor([0.0732, 0.0842, 0.0952, 0.1123, 0.1082, 0.1538, 0.0722, 0.5596], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0247, 0.0281, 0.0295, 0.0336, 0.0284, 0.0306, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:24:22,839 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140206.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:26,682 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 06:24:28,675 INFO [finetune.py:976] (0/7) Epoch 25, batch 2750, loss[loss=0.1336, simple_loss=0.207, pruned_loss=0.03011, over 4830.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2437, pruned_loss=0.0499, over 953580.36 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:24:36,535 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140227.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:43,710 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:58,284 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140259.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:01,787 INFO [finetune.py:976] (0/7) Epoch 25, batch 2800, loss[loss=0.1498, simple_loss=0.2224, pruned_loss=0.03862, over 4834.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2405, pruned_loss=0.04865, over 955618.03 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:25:02,941 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.479e+02 1.751e+02 2.221e+02 3.486e+02, threshold=3.502e+02, percent-clipped=1.0 +2023-03-27 06:25:10,750 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140275.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:22,261 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:53,836 INFO [finetune.py:976] (0/7) Epoch 25, batch 2850, loss[loss=0.1807, simple_loss=0.2536, pruned_loss=0.05394, over 4820.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.24, pruned_loss=0.04896, over 953450.67 frames. ], batch size: 45, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:25:56,986 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140320.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:26:27,553 INFO [finetune.py:976] (0/7) Epoch 25, batch 2900, loss[loss=0.2111, simple_loss=0.29, pruned_loss=0.06616, over 4732.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2424, pruned_loss=0.04955, over 951900.13 frames. ], batch size: 59, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:26:28,760 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.583e+02 1.866e+02 2.190e+02 4.311e+02, threshold=3.732e+02, percent-clipped=1.0 +2023-03-27 06:26:40,787 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6934, 1.2655, 0.8540, 1.5664, 1.9745, 1.3509, 1.5197, 1.5852], + device='cuda:0'), covar=tensor([0.1587, 0.2213, 0.2138, 0.1307, 0.2198, 0.2131, 0.1444, 0.2042], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 06:26:41,426 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9521, 1.1367, 1.9052, 1.8246, 1.7490, 1.6822, 1.7632, 1.9522], + device='cuda:0'), covar=tensor([0.3755, 0.3752, 0.3498, 0.3866, 0.5221, 0.4200, 0.4519, 0.3352], + device='cuda:0'), in_proj_covar=tensor([0.0263, 0.0247, 0.0266, 0.0292, 0.0292, 0.0268, 0.0297, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:26:41,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140387.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:26:52,264 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5585, 2.3764, 2.1277, 2.6408, 2.5064, 2.2967, 2.7914, 2.5751], + device='cuda:0'), covar=tensor([0.1281, 0.2086, 0.2757, 0.1921, 0.2352, 0.1583, 0.2311, 0.1639], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0189, 0.0234, 0.0252, 0.0247, 0.0204, 0.0213, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:27:01,088 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.77 vs. limit=5.0 +2023-03-27 06:27:01,453 INFO [finetune.py:976] (0/7) Epoch 25, batch 2950, loss[loss=0.1719, simple_loss=0.2555, pruned_loss=0.04412, over 4907.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2447, pruned_loss=0.04986, over 952772.00 frames. ], batch size: 37, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:27:02,760 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0323, 1.6468, 2.4793, 1.6877, 2.1245, 2.3819, 1.6149, 2.4547], + device='cuda:0'), covar=tensor([0.1435, 0.2457, 0.1298, 0.1855, 0.1063, 0.1301, 0.3190, 0.0936], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0209, 0.0193, 0.0192, 0.0176, 0.0215, 0.0218, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:27:07,555 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:21,219 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:23,056 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140448.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:30,132 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140459.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:30,774 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:32,483 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:33,171 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-27 06:27:34,685 INFO [finetune.py:976] (0/7) Epoch 25, batch 3000, loss[loss=0.1798, simple_loss=0.2512, pruned_loss=0.05422, over 4859.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2463, pruned_loss=0.05066, over 954134.72 frames. ], batch size: 31, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:27:34,686 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 06:27:48,789 INFO [finetune.py:1010] (0/7) Epoch 25, validation: loss=0.1571, simple_loss=0.2254, pruned_loss=0.04443, over 2265189.00 frames. +2023-03-27 06:27:48,790 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 06:27:49,489 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140466.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:49,621 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-03-27 06:27:50,498 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.568e+02 1.888e+02 2.214e+02 4.503e+02, threshold=3.776e+02, percent-clipped=3.0 +2023-03-27 06:27:54,106 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6657, 1.5031, 2.0234, 2.0348, 1.6760, 3.5761, 1.4722, 1.6140], + device='cuda:0'), covar=tensor([0.0909, 0.1760, 0.1132, 0.0843, 0.1597, 0.0247, 0.1554, 0.1796], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:27:59,512 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140473.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:14,352 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140489.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:20,342 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:29,178 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:30,342 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140508.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:30,996 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:31,579 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140510.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:32,830 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4037, 1.3960, 1.9200, 1.7242, 1.5614, 3.4144, 1.4124, 1.5803], + device='cuda:0'), covar=tensor([0.1012, 0.1876, 0.1177, 0.0980, 0.1702, 0.0212, 0.1538, 0.1825], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:28:34,543 INFO [finetune.py:976] (0/7) Epoch 25, batch 3050, loss[loss=0.1729, simple_loss=0.2492, pruned_loss=0.04833, over 4801.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2476, pruned_loss=0.05052, over 954841.65 frames. ], batch size: 40, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:28:58,612 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140550.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:00,938 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140554.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:08,056 INFO [finetune.py:976] (0/7) Epoch 25, batch 3100, loss[loss=0.1433, simple_loss=0.2195, pruned_loss=0.03356, over 4776.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.246, pruned_loss=0.04987, over 956857.98 frames. ], batch size: 26, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:29:09,242 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.313e+01 1.495e+02 1.767e+02 2.180e+02 4.499e+02, threshold=3.535e+02, percent-clipped=1.0 +2023-03-27 06:29:12,195 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140570.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:42,051 INFO [finetune.py:976] (0/7) Epoch 25, batch 3150, loss[loss=0.1243, simple_loss=0.2032, pruned_loss=0.02268, over 4935.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2423, pruned_loss=0.04907, over 957444.44 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:29:42,121 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140615.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:46,962 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4934, 2.3720, 1.9115, 2.4857, 2.3598, 2.0863, 2.6893, 2.4689], + device='cuda:0'), covar=tensor([0.1367, 0.2095, 0.2950, 0.2461, 0.2554, 0.1740, 0.2764, 0.1679], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0190, 0.0235, 0.0253, 0.0248, 0.0205, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:30:15,049 INFO [finetune.py:976] (0/7) Epoch 25, batch 3200, loss[loss=0.2031, simple_loss=0.2586, pruned_loss=0.07383, over 4822.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2398, pruned_loss=0.04867, over 956781.95 frames. ], batch size: 40, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:30:16,220 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.929e+01 1.486e+02 1.750e+02 2.144e+02 4.466e+02, threshold=3.500e+02, percent-clipped=2.0 +2023-03-27 06:30:49,672 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 06:31:06,567 INFO [finetune.py:976] (0/7) Epoch 25, batch 3250, loss[loss=0.1691, simple_loss=0.2439, pruned_loss=0.0471, over 4826.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2409, pruned_loss=0.04973, over 956157.63 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:31:11,473 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140723.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:24,215 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-27 06:31:25,167 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140743.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:35,350 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140759.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:39,361 INFO [finetune.py:976] (0/7) Epoch 25, batch 3300, loss[loss=0.1868, simple_loss=0.255, pruned_loss=0.05929, over 4826.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2432, pruned_loss=0.0505, over 955965.46 frames. ], batch size: 40, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:31:40,541 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:41,064 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.629e+02 1.945e+02 2.397e+02 4.021e+02, threshold=3.889e+02, percent-clipped=5.0 +2023-03-27 06:31:43,078 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.13 vs. limit=5.0 +2023-03-27 06:31:53,078 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:56,525 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4771, 1.4420, 2.0092, 1.8354, 1.6412, 3.5491, 1.4866, 1.6434], + device='cuda:0'), covar=tensor([0.0951, 0.1796, 0.1102, 0.0936, 0.1565, 0.0201, 0.1434, 0.1758], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:32:08,088 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:09,344 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140809.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:12,394 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140814.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:12,930 INFO [finetune.py:976] (0/7) Epoch 25, batch 3350, loss[loss=0.1618, simple_loss=0.244, pruned_loss=0.03983, over 4887.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2462, pruned_loss=0.05174, over 955866.22 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:32:20,668 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5286, 3.4184, 3.2470, 1.4121, 3.4756, 2.6238, 1.1662, 2.3597], + device='cuda:0'), covar=tensor([0.2476, 0.1966, 0.1519, 0.3203, 0.1278, 0.1055, 0.3489, 0.1413], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0178, 0.0161, 0.0130, 0.0160, 0.0123, 0.0147, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 06:32:33,946 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140845.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:44,304 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0664, 2.0913, 1.4498, 2.0313, 2.0458, 1.6913, 2.6671, 2.0826], + device='cuda:0'), covar=tensor([0.1462, 0.1998, 0.3521, 0.3071, 0.2937, 0.1827, 0.2295, 0.2001], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0188, 0.0233, 0.0251, 0.0246, 0.0204, 0.0212, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:32:46,606 INFO [finetune.py:976] (0/7) Epoch 25, batch 3400, loss[loss=0.1835, simple_loss=0.2613, pruned_loss=0.05285, over 4805.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2471, pruned_loss=0.05178, over 955757.14 frames. ], batch size: 40, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:32:46,674 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:47,794 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.564e+02 1.878e+02 2.236e+02 3.278e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-27 06:32:49,697 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:33:19,362 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4755, 3.8731, 4.0846, 4.3063, 4.2866, 3.9296, 4.5865, 1.3887], + device='cuda:0'), covar=tensor([0.0809, 0.0803, 0.0789, 0.1005, 0.1163, 0.1434, 0.0637, 0.5629], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0247, 0.0282, 0.0296, 0.0337, 0.0286, 0.0307, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:33:38,065 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8512, 1.5211, 2.3375, 1.4984, 1.9867, 2.1186, 1.4306, 2.1657], + device='cuda:0'), covar=tensor([0.1293, 0.2179, 0.1024, 0.1708, 0.0913, 0.1347, 0.2784, 0.0998], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0190, 0.0190, 0.0174, 0.0213, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:33:39,167 INFO [finetune.py:976] (0/7) Epoch 25, batch 3450, loss[loss=0.1629, simple_loss=0.2319, pruned_loss=0.047, over 4777.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2461, pruned_loss=0.05104, over 955529.29 frames. ], batch size: 29, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:33:39,278 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140915.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:33:40,469 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140917.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:01,681 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:11,807 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140963.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:12,940 INFO [finetune.py:976] (0/7) Epoch 25, batch 3500, loss[loss=0.1976, simple_loss=0.2645, pruned_loss=0.06539, over 4900.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2435, pruned_loss=0.0503, over 956231.06 frames. ], batch size: 36, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:34:14,173 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.468e+02 1.748e+02 2.204e+02 3.629e+02, threshold=3.496e+02, percent-clipped=0.0 +2023-03-27 06:34:20,896 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:41,916 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:46,048 INFO [finetune.py:976] (0/7) Epoch 25, batch 3550, loss[loss=0.1446, simple_loss=0.2213, pruned_loss=0.034, over 4823.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2413, pruned_loss=0.04965, over 954050.46 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:34:49,865 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-27 06:35:04,135 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141043.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:19,345 INFO [finetune.py:976] (0/7) Epoch 25, batch 3600, loss[loss=0.1634, simple_loss=0.243, pruned_loss=0.04192, over 4908.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2391, pruned_loss=0.04919, over 955017.42 frames. ], batch size: 37, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:35:20,525 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.456e+02 1.796e+02 2.356e+02 3.995e+02, threshold=3.592e+02, percent-clipped=1.0 +2023-03-27 06:35:28,401 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141079.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:36,613 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141091.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:37,332 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-27 06:35:46,121 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2894, 2.2907, 1.9828, 2.5242, 2.2954, 2.1834, 2.2190, 3.0589], + device='cuda:0'), covar=tensor([0.3613, 0.4840, 0.3295, 0.3788, 0.4038, 0.2599, 0.4089, 0.1500], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0276, 0.0259, 0.0229, 0.0255, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:36:00,471 INFO [finetune.py:976] (0/7) Epoch 25, batch 3650, loss[loss=0.225, simple_loss=0.2947, pruned_loss=0.07765, over 4909.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2405, pruned_loss=0.04959, over 954088.85 frames. ], batch size: 36, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:36:30,931 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141143.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:32,098 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:45,006 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7689, 1.8067, 1.5965, 2.0518, 2.0513, 2.0099, 1.5165, 1.4997], + device='cuda:0'), covar=tensor([0.2065, 0.1734, 0.1759, 0.1424, 0.1713, 0.1110, 0.2155, 0.1819], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0215, 0.0197, 0.0245, 0.0192, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:36:46,081 INFO [finetune.py:976] (0/7) Epoch 25, batch 3700, loss[loss=0.1963, simple_loss=0.259, pruned_loss=0.06676, over 4813.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2432, pruned_loss=0.05054, over 954084.32 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:36:46,165 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:46,182 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:47,288 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.708e+02 2.029e+02 2.382e+02 3.628e+02, threshold=4.058e+02, percent-clipped=1.0 +2023-03-27 06:37:03,881 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141193.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:11,204 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141204.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:18,115 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:19,743 INFO [finetune.py:976] (0/7) Epoch 25, batch 3750, loss[loss=0.1437, simple_loss=0.2254, pruned_loss=0.031, over 4864.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2444, pruned_loss=0.0505, over 954552.91 frames. ], batch size: 31, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:37:52,662 INFO [finetune.py:976] (0/7) Epoch 25, batch 3800, loss[loss=0.1857, simple_loss=0.2664, pruned_loss=0.05245, over 4801.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2455, pruned_loss=0.05052, over 955075.50 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:37:54,343 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.508e+02 1.827e+02 2.217e+02 6.513e+02, threshold=3.654e+02, percent-clipped=2.0 +2023-03-27 06:37:58,031 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:19,442 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141303.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:26,938 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5638, 1.4732, 1.9316, 3.3995, 2.2696, 2.4753, 0.7361, 2.8174], + device='cuda:0'), covar=tensor([0.1723, 0.1371, 0.1391, 0.0506, 0.0772, 0.1488, 0.1966, 0.0412], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0102, 0.0136, 0.0123, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 06:38:32,077 INFO [finetune.py:976] (0/7) Epoch 25, batch 3850, loss[loss=0.1599, simple_loss=0.2269, pruned_loss=0.04644, over 4819.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2454, pruned_loss=0.05113, over 954772.01 frames. ], batch size: 33, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:38:49,966 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:57,263 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5795, 1.5844, 1.3415, 1.5692, 1.9349, 1.8111, 1.5395, 1.4273], + device='cuda:0'), covar=tensor([0.0362, 0.0342, 0.0606, 0.0317, 0.0215, 0.0484, 0.0352, 0.0427], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0107, 0.0145, 0.0111, 0.0100, 0.0114, 0.0103, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7825e-05, 8.1683e-05, 1.1336e-04, 8.5251e-05, 7.7699e-05, 8.4305e-05, + 7.6375e-05, 8.5392e-05], device='cuda:0') +2023-03-27 06:38:57,281 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2746, 2.2202, 1.7252, 2.2917, 2.1818, 1.9251, 2.5245, 2.2647], + device='cuda:0'), covar=tensor([0.1244, 0.1950, 0.2864, 0.2406, 0.2381, 0.1573, 0.2600, 0.1636], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0191, 0.0235, 0.0253, 0.0249, 0.0206, 0.0213, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:39:16,916 INFO [finetune.py:976] (0/7) Epoch 25, batch 3900, loss[loss=0.1448, simple_loss=0.2265, pruned_loss=0.03152, over 4828.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2434, pruned_loss=0.05057, over 956062.38 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:39:18,105 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.504e+02 1.773e+02 2.110e+02 6.012e+02, threshold=3.546e+02, percent-clipped=1.0 +2023-03-27 06:39:25,836 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9890, 2.1728, 1.7226, 1.8936, 2.4541, 2.6095, 2.1299, 2.0890], + device='cuda:0'), covar=tensor([0.0466, 0.0370, 0.0654, 0.0348, 0.0267, 0.0434, 0.0394, 0.0389], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8092e-05, 8.1913e-05, 1.1369e-04, 8.5441e-05, 7.7918e-05, 8.4623e-05, + 7.6618e-05, 8.5683e-05], device='cuda:0') +2023-03-27 06:39:26,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141379.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:39:33,595 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:39:49,667 INFO [finetune.py:976] (0/7) Epoch 25, batch 3950, loss[loss=0.1677, simple_loss=0.2332, pruned_loss=0.05106, over 4814.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2409, pruned_loss=0.05009, over 955133.38 frames. ], batch size: 51, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:39:51,035 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141417.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:39:58,915 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141427.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:07,856 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141441.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:19,338 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6851, 4.0743, 4.2882, 4.4286, 4.4672, 4.2138, 4.7665, 1.5590], + device='cuda:0'), covar=tensor([0.0757, 0.0828, 0.0800, 0.1103, 0.1186, 0.1555, 0.0619, 0.5714], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0248, 0.0283, 0.0296, 0.0338, 0.0288, 0.0309, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:40:23,345 INFO [finetune.py:976] (0/7) Epoch 25, batch 4000, loss[loss=0.1481, simple_loss=0.2124, pruned_loss=0.04191, over 4820.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2411, pruned_loss=0.05023, over 953207.59 frames. ], batch size: 25, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:40:23,423 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141465.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:24,526 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.514e+02 1.750e+02 2.113e+02 3.817e+02, threshold=3.500e+02, percent-clipped=1.0 +2023-03-27 06:40:33,185 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141478.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 06:40:46,340 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141499.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:48,236 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141502.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:55,308 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:56,982 INFO [finetune.py:976] (0/7) Epoch 25, batch 4050, loss[loss=0.1647, simple_loss=0.2268, pruned_loss=0.05129, over 3993.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05131, over 953106.34 frames. ], batch size: 17, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:41:49,013 INFO [finetune.py:976] (0/7) Epoch 25, batch 4100, loss[loss=0.1302, simple_loss=0.2038, pruned_loss=0.02825, over 4754.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2467, pruned_loss=0.05196, over 953015.74 frames. ], batch size: 26, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:41:50,186 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.605e+02 1.887e+02 2.173e+02 5.231e+02, threshold=3.774e+02, percent-clipped=3.0 +2023-03-27 06:41:54,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141573.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:41:57,565 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-27 06:42:14,764 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141603.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:22,402 INFO [finetune.py:976] (0/7) Epoch 25, batch 4150, loss[loss=0.1339, simple_loss=0.2097, pruned_loss=0.02911, over 4802.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2482, pruned_loss=0.05243, over 953705.99 frames. ], batch size: 25, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:42:26,080 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141621.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:46,644 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141651.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:55,969 INFO [finetune.py:976] (0/7) Epoch 25, batch 4200, loss[loss=0.1634, simple_loss=0.232, pruned_loss=0.04742, over 4295.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2481, pruned_loss=0.05176, over 954915.93 frames. ], batch size: 66, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:42:57,194 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.553e+02 1.832e+02 2.223e+02 5.119e+02, threshold=3.664e+02, percent-clipped=2.0 +2023-03-27 06:43:09,606 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:43:29,321 INFO [finetune.py:976] (0/7) Epoch 25, batch 4250, loss[loss=0.168, simple_loss=0.2364, pruned_loss=0.04976, over 4820.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2453, pruned_loss=0.05104, over 954389.22 frames. ], batch size: 41, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:44:21,245 INFO [finetune.py:976] (0/7) Epoch 25, batch 4300, loss[loss=0.1405, simple_loss=0.2059, pruned_loss=0.03757, over 4905.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2421, pruned_loss=0.05018, over 953990.92 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:44:22,423 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.814e+01 1.357e+02 1.630e+02 2.025e+02 3.929e+02, threshold=3.260e+02, percent-clipped=1.0 +2023-03-27 06:44:26,617 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141773.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:44:43,603 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141797.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:44:44,857 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141799.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:44:46,876 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8691, 1.8096, 1.6478, 2.0591, 2.3142, 2.1049, 1.6006, 1.5981], + device='cuda:0'), covar=tensor([0.2325, 0.2038, 0.2025, 0.1720, 0.1566, 0.1122, 0.2362, 0.1948], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0214, 0.0197, 0.0245, 0.0191, 0.0216, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:44:55,201 INFO [finetune.py:976] (0/7) Epoch 25, batch 4350, loss[loss=0.1833, simple_loss=0.2521, pruned_loss=0.0573, over 4910.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2386, pruned_loss=0.04882, over 954550.25 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:45:17,467 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141847.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:45:28,845 INFO [finetune.py:976] (0/7) Epoch 25, batch 4400, loss[loss=0.1746, simple_loss=0.2493, pruned_loss=0.04997, over 4801.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2407, pruned_loss=0.05005, over 955105.10 frames. ], batch size: 51, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:45:30,033 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.595e+02 1.814e+02 2.202e+02 4.275e+02, threshold=3.628e+02, percent-clipped=6.0 +2023-03-27 06:45:48,598 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 06:46:01,884 INFO [finetune.py:976] (0/7) Epoch 25, batch 4450, loss[loss=0.1506, simple_loss=0.2421, pruned_loss=0.02959, over 4753.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2436, pruned_loss=0.05049, over 954672.77 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:46:02,609 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141916.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:46:04,537 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-03-27 06:46:06,892 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5809, 1.4745, 1.4828, 1.5208, 1.1573, 2.9731, 1.0149, 1.4724], + device='cuda:0'), covar=tensor([0.3176, 0.2473, 0.2125, 0.2296, 0.1631, 0.0288, 0.2737, 0.1295], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 06:46:09,960 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141928.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:46:23,731 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.4174, 4.7706, 4.9992, 5.2189, 5.1350, 4.8543, 5.5450, 1.7208], + device='cuda:0'), covar=tensor([0.0729, 0.0758, 0.0734, 0.0954, 0.1167, 0.1422, 0.0552, 0.5643], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0248, 0.0282, 0.0296, 0.0338, 0.0288, 0.0306, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:46:46,889 INFO [finetune.py:976] (0/7) Epoch 25, batch 4500, loss[loss=0.2362, simple_loss=0.2942, pruned_loss=0.08912, over 4801.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2452, pruned_loss=0.05095, over 953121.55 frames. ], batch size: 40, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:46:52,641 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.869e+01 1.594e+02 1.826e+02 2.236e+02 4.959e+02, threshold=3.653e+02, percent-clipped=2.0 +2023-03-27 06:47:02,919 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141977.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:08,156 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:10,666 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141989.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:18,895 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-142000.pt +2023-03-27 06:47:30,042 INFO [finetune.py:976] (0/7) Epoch 25, batch 4550, loss[loss=0.1777, simple_loss=0.2548, pruned_loss=0.05032, over 4742.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2449, pruned_loss=0.05031, over 953133.85 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:47:41,388 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:48:03,343 INFO [finetune.py:976] (0/7) Epoch 25, batch 4600, loss[loss=0.1852, simple_loss=0.2504, pruned_loss=0.05997, over 4846.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2446, pruned_loss=0.04983, over 953891.40 frames. ], batch size: 44, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:48:04,586 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.716e+01 1.569e+02 1.799e+02 2.271e+02 3.318e+02, threshold=3.598e+02, percent-clipped=0.0 +2023-03-27 06:48:07,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4281, 1.3539, 1.5971, 2.4833, 1.6954, 2.3301, 1.0280, 2.1739], + device='cuda:0'), covar=tensor([0.1675, 0.1322, 0.1141, 0.0693, 0.0876, 0.1113, 0.1400, 0.0544], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0101, 0.0136, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 06:48:08,700 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142073.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:48:23,781 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142097.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:48:32,116 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5438, 2.6571, 2.5660, 1.7451, 2.3451, 2.7452, 2.8090, 2.1938], + device='cuda:0'), covar=tensor([0.0552, 0.0594, 0.0734, 0.0920, 0.1199, 0.0663, 0.0522, 0.1025], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0136, 0.0140, 0.0119, 0.0127, 0.0138, 0.0139, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:48:36,589 INFO [finetune.py:976] (0/7) Epoch 25, batch 4650, loss[loss=0.1935, simple_loss=0.2549, pruned_loss=0.06603, over 4917.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2423, pruned_loss=0.04922, over 954438.55 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:48:40,333 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142121.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:48:47,503 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2617, 2.0418, 1.6824, 1.6194, 2.4542, 2.8021, 2.1831, 2.1027], + device='cuda:0'), covar=tensor([0.0368, 0.0405, 0.0891, 0.0453, 0.0293, 0.0383, 0.0392, 0.0450], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0145, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8275e-05, 8.1544e-05, 1.1356e-04, 8.5214e-05, 7.7992e-05, 8.4926e-05, + 7.6603e-05, 8.5671e-05], device='cuda:0') +2023-03-27 06:48:57,214 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:49:19,997 INFO [finetune.py:976] (0/7) Epoch 25, batch 4700, loss[loss=0.1289, simple_loss=0.1935, pruned_loss=0.03221, over 4791.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2394, pruned_loss=0.04856, over 955808.95 frames. ], batch size: 29, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:49:21,183 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.163e+01 1.384e+02 1.765e+02 2.088e+02 3.764e+02, threshold=3.531e+02, percent-clipped=1.0 +2023-03-27 06:50:00,898 INFO [finetune.py:976] (0/7) Epoch 25, batch 4750, loss[loss=0.1494, simple_loss=0.225, pruned_loss=0.03692, over 4888.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2382, pruned_loss=0.04824, over 955332.59 frames. ], batch size: 32, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:50:34,333 INFO [finetune.py:976] (0/7) Epoch 25, batch 4800, loss[loss=0.2315, simple_loss=0.3086, pruned_loss=0.07718, over 4816.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2421, pruned_loss=0.04955, over 954209.29 frames. ], batch size: 40, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:50:35,545 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.947e+01 1.535e+02 1.762e+02 2.238e+02 3.446e+02, threshold=3.524e+02, percent-clipped=1.0 +2023-03-27 06:50:39,630 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:50:47,477 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142284.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:51:07,058 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1942, 2.2320, 1.7908, 2.2287, 2.1139, 2.1017, 2.1307, 2.8922], + device='cuda:0'), covar=tensor([0.3496, 0.4436, 0.3368, 0.4034, 0.4493, 0.2383, 0.4044, 0.1632], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0235, 0.0275, 0.0258, 0.0228, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:51:07,519 INFO [finetune.py:976] (0/7) Epoch 25, batch 4850, loss[loss=0.1667, simple_loss=0.2445, pruned_loss=0.04445, over 4840.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2452, pruned_loss=0.05031, over 953177.94 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:51:39,151 INFO [finetune.py:976] (0/7) Epoch 25, batch 4900, loss[loss=0.1922, simple_loss=0.2697, pruned_loss=0.05732, over 4895.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2455, pruned_loss=0.05069, over 950939.24 frames. ], batch size: 36, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:51:40,866 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.551e+02 1.812e+02 2.135e+02 6.918e+02, threshold=3.624e+02, percent-clipped=2.0 +2023-03-27 06:52:31,143 INFO [finetune.py:976] (0/7) Epoch 25, batch 4950, loss[loss=0.1712, simple_loss=0.2473, pruned_loss=0.0475, over 4805.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.0503, over 953051.32 frames. ], batch size: 40, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:52:50,126 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 06:53:03,769 INFO [finetune.py:976] (0/7) Epoch 25, batch 5000, loss[loss=0.1369, simple_loss=0.2086, pruned_loss=0.03255, over 4794.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2446, pruned_loss=0.05003, over 952102.00 frames. ], batch size: 29, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:53:04,975 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.190e+01 1.432e+02 1.813e+02 2.155e+02 3.992e+02, threshold=3.625e+02, percent-clipped=1.0 +2023-03-27 06:53:21,194 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-27 06:53:36,413 INFO [finetune.py:976] (0/7) Epoch 25, batch 5050, loss[loss=0.1454, simple_loss=0.2221, pruned_loss=0.03431, over 4850.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2422, pruned_loss=0.04975, over 950303.06 frames. ], batch size: 49, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:54:09,847 INFO [finetune.py:976] (0/7) Epoch 25, batch 5100, loss[loss=0.2191, simple_loss=0.2789, pruned_loss=0.0796, over 4904.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2384, pruned_loss=0.04827, over 950330.24 frames. ], batch size: 36, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:54:11,047 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.212e+01 1.519e+02 1.807e+02 2.247e+02 4.075e+02, threshold=3.613e+02, percent-clipped=2.0 +2023-03-27 06:54:14,210 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:14,849 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142573.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:24,167 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142584.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:59,671 INFO [finetune.py:976] (0/7) Epoch 25, batch 5150, loss[loss=0.1495, simple_loss=0.2121, pruned_loss=0.04347, over 4795.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.239, pruned_loss=0.04859, over 952119.22 frames. ], batch size: 26, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:55:03,299 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142620.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:10,595 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:12,832 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142634.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:33,009 INFO [finetune.py:976] (0/7) Epoch 25, batch 5200, loss[loss=0.1876, simple_loss=0.2732, pruned_loss=0.05099, over 4809.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.242, pruned_loss=0.04944, over 950745.57 frames. ], batch size: 45, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:55:34,193 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.563e+02 1.762e+02 2.093e+02 3.679e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-03-27 06:55:59,594 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1232, 2.9046, 2.5631, 1.3509, 2.7633, 2.1739, 2.0900, 2.4615], + device='cuda:0'), covar=tensor([0.0990, 0.0846, 0.1974, 0.2315, 0.1768, 0.2339, 0.2475, 0.1316], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0194, 0.0203, 0.0184, 0.0212, 0.0214, 0.0227, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:56:06,169 INFO [finetune.py:976] (0/7) Epoch 25, batch 5250, loss[loss=0.2166, simple_loss=0.2859, pruned_loss=0.07359, over 4902.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2443, pruned_loss=0.05006, over 951355.60 frames. ], batch size: 37, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:56:12,144 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142724.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:56:18,243 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3026, 1.3060, 1.4899, 1.0546, 1.3043, 1.4785, 1.3077, 1.5922], + device='cuda:0'), covar=tensor([0.1026, 0.2000, 0.1144, 0.1380, 0.0820, 0.1036, 0.2610, 0.0695], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0204, 0.0190, 0.0188, 0.0173, 0.0211, 0.0213, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 06:56:36,987 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-27 06:56:39,095 INFO [finetune.py:976] (0/7) Epoch 25, batch 5300, loss[loss=0.186, simple_loss=0.2591, pruned_loss=0.05648, over 4776.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2458, pruned_loss=0.05071, over 952019.24 frames. ], batch size: 29, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:56:40,274 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.558e+02 1.826e+02 2.127e+02 3.045e+02, threshold=3.651e+02, percent-clipped=0.0 +2023-03-27 06:56:51,748 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142785.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:56:58,199 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-03-27 06:57:19,998 INFO [finetune.py:976] (0/7) Epoch 25, batch 5350, loss[loss=0.1526, simple_loss=0.2101, pruned_loss=0.0475, over 3977.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2462, pruned_loss=0.05035, over 953457.90 frames. ], batch size: 17, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:57:25,295 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1968, 2.2089, 1.8843, 1.7971, 2.6630, 2.7673, 2.2509, 2.1796], + device='cuda:0'), covar=tensor([0.0373, 0.0342, 0.0582, 0.0371, 0.0207, 0.0444, 0.0286, 0.0388], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0111, 0.0101, 0.0115, 0.0103, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7880e-05, 8.1480e-05, 1.1366e-04, 8.5156e-05, 7.8299e-05, 8.4950e-05, + 7.6521e-05, 8.5515e-05], device='cuda:0') +2023-03-27 06:57:27,160 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 06:58:06,041 INFO [finetune.py:976] (0/7) Epoch 25, batch 5400, loss[loss=0.1638, simple_loss=0.2402, pruned_loss=0.04369, over 4836.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.243, pruned_loss=0.04915, over 954072.21 frames. ], batch size: 39, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:58:07,256 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.487e+02 1.682e+02 2.190e+02 4.832e+02, threshold=3.364e+02, percent-clipped=1.0 +2023-03-27 06:58:15,835 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-27 06:58:38,663 INFO [finetune.py:976] (0/7) Epoch 25, batch 5450, loss[loss=0.127, simple_loss=0.1955, pruned_loss=0.0293, over 4731.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2409, pruned_loss=0.04841, over 956179.05 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:58:47,634 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:59:11,889 INFO [finetune.py:976] (0/7) Epoch 25, batch 5500, loss[loss=0.1761, simple_loss=0.2452, pruned_loss=0.0535, over 4812.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2385, pruned_loss=0.04801, over 957431.19 frames. ], batch size: 45, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 06:59:13,716 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.198e+01 1.372e+02 1.708e+02 2.223e+02 4.314e+02, threshold=3.415e+02, percent-clipped=3.0 +2023-03-27 06:59:19,159 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4482, 1.5853, 1.2636, 1.4363, 1.8343, 1.7324, 1.4541, 1.3483], + device='cuda:0'), covar=tensor([0.0402, 0.0302, 0.0579, 0.0338, 0.0208, 0.0560, 0.0368, 0.0415], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0111, 0.0101, 0.0116, 0.0104, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8220e-05, 8.1740e-05, 1.1402e-04, 8.5260e-05, 7.8377e-05, 8.5454e-05, + 7.6959e-05, 8.5680e-05], device='cuda:0') +2023-03-27 06:59:46,280 INFO [finetune.py:976] (0/7) Epoch 25, batch 5550, loss[loss=0.1941, simple_loss=0.2688, pruned_loss=0.05971, over 4926.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2406, pruned_loss=0.04879, over 957958.45 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:00:24,804 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8481, 3.7152, 3.4239, 1.7670, 3.7743, 2.9294, 0.9165, 2.6587], + device='cuda:0'), covar=tensor([0.2338, 0.2436, 0.1868, 0.3727, 0.1166, 0.1059, 0.4852, 0.1611], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0179, 0.0162, 0.0131, 0.0161, 0.0124, 0.0148, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 07:00:29,923 INFO [finetune.py:976] (0/7) Epoch 25, batch 5600, loss[loss=0.2159, simple_loss=0.2741, pruned_loss=0.07885, over 4760.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2443, pruned_loss=0.04944, over 956426.05 frames. ], batch size: 54, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:00:31,669 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.665e+01 1.700e+02 1.937e+02 2.306e+02 4.675e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-27 07:00:38,726 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:01:00,341 INFO [finetune.py:976] (0/7) Epoch 25, batch 5650, loss[loss=0.1824, simple_loss=0.2585, pruned_loss=0.05319, over 4797.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2471, pruned_loss=0.05015, over 957993.78 frames. ], batch size: 29, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:01:23,401 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8641, 3.4349, 3.0857, 2.0312, 3.3930, 2.8838, 2.6987, 3.1153], + device='cuda:0'), covar=tensor([0.0686, 0.0796, 0.1599, 0.1891, 0.1297, 0.1700, 0.1771, 0.0969], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0193, 0.0201, 0.0183, 0.0211, 0.0213, 0.0225, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:01:29,839 INFO [finetune.py:976] (0/7) Epoch 25, batch 5700, loss[loss=0.1376, simple_loss=0.1977, pruned_loss=0.03874, over 3998.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2418, pruned_loss=0.04868, over 938756.95 frames. ], batch size: 17, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:01:31,567 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.213e+01 1.339e+02 1.671e+02 2.043e+02 4.216e+02, threshold=3.342e+02, percent-clipped=1.0 +2023-03-27 07:01:33,573 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-03-27 07:01:35,817 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4132, 1.5222, 1.6219, 0.9304, 1.5769, 1.8003, 1.8153, 1.4990], + device='cuda:0'), covar=tensor([0.0868, 0.0736, 0.0619, 0.0570, 0.0578, 0.0647, 0.0444, 0.0656], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0146, 0.0126, 0.0121, 0.0129, 0.0128, 0.0140, 0.0147], + device='cuda:0'), out_proj_covar=tensor([8.8016e-05, 1.0488e-04, 9.0215e-05, 8.4865e-05, 8.9921e-05, 9.0713e-05, + 9.9719e-05, 1.0468e-04], device='cuda:0') +2023-03-27 07:01:46,294 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-25.pt +2023-03-27 07:01:58,298 INFO [finetune.py:976] (0/7) Epoch 26, batch 0, loss[loss=0.1867, simple_loss=0.2634, pruned_loss=0.05506, over 4810.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2634, pruned_loss=0.05506, over 4810.00 frames. ], batch size: 38, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:01:58,299 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 07:02:00,832 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8381, 1.7045, 2.0012, 1.3753, 1.7075, 2.0304, 1.6787, 2.1481], + device='cuda:0'), covar=tensor([0.1252, 0.2388, 0.1471, 0.1743, 0.1023, 0.1472, 0.2934, 0.0916], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0190, 0.0190, 0.0173, 0.0212, 0.0215, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:02:02,845 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4728, 1.3358, 1.3190, 1.4335, 1.6759, 1.6273, 1.4480, 1.2810], + device='cuda:0'), covar=tensor([0.0453, 0.0363, 0.0650, 0.0344, 0.0308, 0.0521, 0.0373, 0.0498], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0102, 0.0116, 0.0104, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8500e-05, 8.1959e-05, 1.1422e-04, 8.5531e-05, 7.8734e-05, 8.5958e-05, + 7.7183e-05, 8.6065e-05], device='cuda:0') +2023-03-27 07:02:05,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1001, 1.8839, 1.8023, 1.7983, 1.8403, 1.8536, 1.8685, 2.4713], + device='cuda:0'), covar=tensor([0.3384, 0.4312, 0.3070, 0.3462, 0.4020, 0.2311, 0.3784, 0.1732], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0237, 0.0278, 0.0260, 0.0230, 0.0257, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:02:14,281 INFO [finetune.py:1010] (0/7) Epoch 26, validation: loss=0.1591, simple_loss=0.2269, pruned_loss=0.04565, over 2265189.00 frames. +2023-03-27 07:02:14,281 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 07:02:27,542 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9201, 1.6806, 2.1029, 1.3803, 1.8893, 2.1739, 1.5668, 2.1457], + device='cuda:0'), covar=tensor([0.1242, 0.2089, 0.1440, 0.1899, 0.1060, 0.1330, 0.2903, 0.0957], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0191, 0.0190, 0.0174, 0.0213, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:02:43,913 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143229.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:03:00,613 INFO [finetune.py:976] (0/7) Epoch 26, batch 50, loss[loss=0.1743, simple_loss=0.2447, pruned_loss=0.05195, over 4869.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2512, pruned_loss=0.05281, over 216064.08 frames. ], batch size: 34, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:03:18,452 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.460e+02 1.766e+02 2.058e+02 4.416e+02, threshold=3.532e+02, percent-clipped=3.0 +2023-03-27 07:03:24,477 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:03:34,116 INFO [finetune.py:976] (0/7) Epoch 26, batch 100, loss[loss=0.1951, simple_loss=0.2578, pruned_loss=0.06622, over 4843.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2428, pruned_loss=0.04959, over 381751.31 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:03:43,418 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5126, 1.0370, 0.7572, 1.3610, 1.9442, 0.8169, 1.2479, 1.3788], + device='cuda:0'), covar=tensor([0.1580, 0.2195, 0.1853, 0.1295, 0.1946, 0.2025, 0.1513, 0.1908], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:04:07,506 INFO [finetune.py:976] (0/7) Epoch 26, batch 150, loss[loss=0.1612, simple_loss=0.2278, pruned_loss=0.04736, over 4820.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2389, pruned_loss=0.0489, over 509982.82 frames. ], batch size: 30, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:04:15,602 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-03-27 07:04:25,693 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.744e+01 1.335e+02 1.679e+02 2.114e+02 2.886e+02, threshold=3.358e+02, percent-clipped=0.0 +2023-03-27 07:04:33,623 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:04:41,259 INFO [finetune.py:976] (0/7) Epoch 26, batch 200, loss[loss=0.1941, simple_loss=0.2433, pruned_loss=0.07244, over 4116.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2371, pruned_loss=0.04816, over 608426.51 frames. ], batch size: 65, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:04:46,840 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143400.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:05:05,300 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143428.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:05:05,377 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1768, 2.1778, 1.8105, 2.3140, 2.2089, 1.9777, 2.5327, 2.2866], + device='cuda:0'), covar=tensor([0.1177, 0.2027, 0.2570, 0.2289, 0.2135, 0.1432, 0.3040, 0.1508], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0191, 0.0237, 0.0256, 0.0252, 0.0208, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:05:22,139 INFO [finetune.py:976] (0/7) Epoch 26, batch 250, loss[loss=0.1977, simple_loss=0.2775, pruned_loss=0.05897, over 4828.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2404, pruned_loss=0.04911, over 683859.30 frames. ], batch size: 40, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:05:48,862 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143461.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:05:51,331 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4883, 2.4418, 2.0373, 1.0342, 2.2563, 1.8810, 1.8177, 2.1694], + device='cuda:0'), covar=tensor([0.0954, 0.0810, 0.1940, 0.2371, 0.1587, 0.2502, 0.2294, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0193, 0.0202, 0.0183, 0.0212, 0.0213, 0.0226, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:05:53,067 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.741e+01 1.618e+02 1.961e+02 2.394e+02 5.476e+02, threshold=3.922e+02, percent-clipped=2.0 +2023-03-27 07:05:59,875 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143473.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:06:12,193 INFO [finetune.py:976] (0/7) Epoch 26, batch 300, loss[loss=0.1813, simple_loss=0.2515, pruned_loss=0.05555, over 4921.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2445, pruned_loss=0.04999, over 744272.14 frames. ], batch size: 42, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:06:27,531 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6113, 1.5854, 1.9087, 1.2999, 1.6673, 1.9240, 1.5016, 2.0840], + device='cuda:0'), covar=tensor([0.1340, 0.2106, 0.1296, 0.1640, 0.1011, 0.1321, 0.2778, 0.0908], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0204, 0.0190, 0.0189, 0.0173, 0.0212, 0.0215, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:06:40,202 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143534.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:06:44,912 INFO [finetune.py:976] (0/7) Epoch 26, batch 350, loss[loss=0.171, simple_loss=0.2505, pruned_loss=0.04578, over 4841.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2462, pruned_loss=0.05072, over 790807.52 frames. ], batch size: 30, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:07:03,059 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.441e+02 1.724e+02 2.077e+02 3.544e+02, threshold=3.448e+02, percent-clipped=0.0 +2023-03-27 07:07:11,598 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9743, 4.7572, 4.6191, 2.6892, 4.8641, 3.7220, 0.9102, 3.4649], + device='cuda:0'), covar=tensor([0.1971, 0.1400, 0.1109, 0.2570, 0.0748, 0.0710, 0.4108, 0.1006], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0178, 0.0161, 0.0130, 0.0160, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 07:07:18,098 INFO [finetune.py:976] (0/7) Epoch 26, batch 400, loss[loss=0.2113, simple_loss=0.2754, pruned_loss=0.07358, over 4806.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2476, pruned_loss=0.05104, over 827106.07 frames. ], batch size: 41, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:07:54,070 INFO [finetune.py:976] (0/7) Epoch 26, batch 450, loss[loss=0.2182, simple_loss=0.2941, pruned_loss=0.07117, over 4882.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2461, pruned_loss=0.05068, over 854754.35 frames. ], batch size: 35, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:08:22,193 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.245e+01 1.544e+02 1.809e+02 2.165e+02 3.752e+02, threshold=3.619e+02, percent-clipped=3.0 +2023-03-27 07:08:37,487 INFO [finetune.py:976] (0/7) Epoch 26, batch 500, loss[loss=0.138, simple_loss=0.2163, pruned_loss=0.02984, over 4910.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.243, pruned_loss=0.04924, over 876765.33 frames. ], batch size: 29, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:08:59,263 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0217, 1.7113, 2.3358, 4.0341, 2.7929, 2.7319, 0.8785, 3.4591], + device='cuda:0'), covar=tensor([0.1759, 0.1439, 0.1504, 0.0514, 0.0771, 0.1606, 0.2131, 0.0378], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0136, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 07:09:11,115 INFO [finetune.py:976] (0/7) Epoch 26, batch 550, loss[loss=0.1493, simple_loss=0.2186, pruned_loss=0.04004, over 4753.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2402, pruned_loss=0.04855, over 893823.45 frames. ], batch size: 27, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:09:20,249 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143756.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:09:28,915 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.175e+01 1.443e+02 1.723e+02 1.984e+02 5.074e+02, threshold=3.446e+02, percent-clipped=2.0 +2023-03-27 07:09:44,563 INFO [finetune.py:976] (0/7) Epoch 26, batch 600, loss[loss=0.235, simple_loss=0.2996, pruned_loss=0.08522, over 4250.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2403, pruned_loss=0.04884, over 903260.15 frames. ], batch size: 65, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:09:56,240 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1682, 2.1587, 2.1779, 1.4782, 2.0746, 2.2514, 2.2965, 1.7546], + device='cuda:0'), covar=tensor([0.0570, 0.0553, 0.0669, 0.0865, 0.0767, 0.0639, 0.0558, 0.1096], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0119, 0.0128, 0.0138, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:10:09,699 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143829.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:10:14,577 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0667, 1.5786, 0.8115, 1.8462, 2.2526, 1.7607, 1.6229, 1.9242], + device='cuda:0'), covar=tensor([0.1467, 0.1912, 0.1999, 0.1252, 0.2051, 0.2037, 0.1394, 0.1955], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:10:17,582 INFO [finetune.py:976] (0/7) Epoch 26, batch 650, loss[loss=0.1748, simple_loss=0.2514, pruned_loss=0.04908, over 4832.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2439, pruned_loss=0.05001, over 915013.73 frames. ], batch size: 49, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:10:18,994 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-03-27 07:10:23,068 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.6388, 3.9961, 4.1984, 4.4339, 4.4252, 4.0887, 4.7043, 1.9165], + device='cuda:0'), covar=tensor([0.0674, 0.0827, 0.0732, 0.0814, 0.1112, 0.1452, 0.0660, 0.4689], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0249, 0.0282, 0.0298, 0.0338, 0.0289, 0.0308, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:10:41,260 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.600e+02 1.821e+02 2.293e+02 5.159e+02, threshold=3.642e+02, percent-clipped=4.0 +2023-03-27 07:11:12,407 INFO [finetune.py:976] (0/7) Epoch 26, batch 700, loss[loss=0.1339, simple_loss=0.1998, pruned_loss=0.03397, over 4724.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2441, pruned_loss=0.05, over 923592.83 frames. ], batch size: 23, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:11:41,697 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5464, 2.4817, 1.9404, 0.9294, 2.1047, 1.9705, 1.8959, 2.2159], + device='cuda:0'), covar=tensor([0.1006, 0.0783, 0.1610, 0.2045, 0.1446, 0.2278, 0.1967, 0.0949], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0193, 0.0201, 0.0183, 0.0211, 0.0213, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:11:48,835 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 07:11:51,903 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.56 vs. limit=2.0 +2023-03-27 07:11:52,958 INFO [finetune.py:976] (0/7) Epoch 26, batch 750, loss[loss=0.1475, simple_loss=0.2193, pruned_loss=0.03789, over 4777.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2444, pruned_loss=0.05, over 928841.27 frames. ], batch size: 51, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:12:09,869 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.567e+02 1.788e+02 2.169e+02 3.888e+02, threshold=3.576e+02, percent-clipped=1.0 +2023-03-27 07:12:23,533 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8006, 1.5218, 1.2717, 1.3860, 1.8446, 1.9514, 1.6550, 1.4173], + device='cuda:0'), covar=tensor([0.0311, 0.0367, 0.0818, 0.0406, 0.0245, 0.0440, 0.0376, 0.0482], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0107, 0.0147, 0.0111, 0.0102, 0.0116, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8859e-05, 8.1923e-05, 1.1446e-04, 8.5337e-05, 7.9135e-05, 8.5613e-05, + 7.6399e-05, 8.6010e-05], device='cuda:0') +2023-03-27 07:12:26,436 INFO [finetune.py:976] (0/7) Epoch 26, batch 800, loss[loss=0.1662, simple_loss=0.2411, pruned_loss=0.04567, over 4860.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2431, pruned_loss=0.04898, over 936143.62 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:12:31,388 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-144000.pt +2023-03-27 07:12:48,819 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8624, 1.6155, 1.4885, 1.2697, 1.5804, 1.5744, 1.5838, 2.1914], + device='cuda:0'), covar=tensor([0.3534, 0.3694, 0.2919, 0.3575, 0.3712, 0.2318, 0.3322, 0.1706], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0236, 0.0278, 0.0259, 0.0230, 0.0258, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:12:55,321 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.5588, 2.9955, 2.8024, 1.4215, 3.1156, 2.5152, 2.3514, 2.8328], + device='cuda:0'), covar=tensor([0.0580, 0.0926, 0.1522, 0.2161, 0.1249, 0.1739, 0.1917, 0.0971], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0194, 0.0202, 0.0183, 0.0212, 0.0213, 0.0226, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:12:59,649 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-03-27 07:13:00,662 INFO [finetune.py:976] (0/7) Epoch 26, batch 850, loss[loss=0.1656, simple_loss=0.2467, pruned_loss=0.04229, over 4727.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2417, pruned_loss=0.04836, over 941679.69 frames. ], batch size: 59, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:13:09,704 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144056.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:13:16,921 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.460e+02 1.746e+02 2.115e+02 7.519e+02, threshold=3.492e+02, percent-clipped=2.0 +2023-03-27 07:13:43,956 INFO [finetune.py:976] (0/7) Epoch 26, batch 900, loss[loss=0.1571, simple_loss=0.2257, pruned_loss=0.04421, over 4933.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2395, pruned_loss=0.0483, over 946096.50 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:13:51,325 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144104.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:14:05,690 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1066, 1.9222, 1.7241, 1.7570, 1.8536, 1.8401, 1.8970, 2.5132], + device='cuda:0'), covar=tensor([0.3852, 0.4017, 0.3308, 0.3386, 0.3828, 0.2527, 0.3534, 0.1802], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0237, 0.0278, 0.0260, 0.0230, 0.0258, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:14:07,459 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144129.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:14:16,858 INFO [finetune.py:976] (0/7) Epoch 26, batch 950, loss[loss=0.1565, simple_loss=0.2272, pruned_loss=0.0429, over 4712.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2387, pruned_loss=0.04829, over 948325.26 frames. ], batch size: 23, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:14:33,135 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.468e+02 1.742e+02 2.065e+02 3.876e+02, threshold=3.485e+02, percent-clipped=2.0 +2023-03-27 07:14:39,231 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144177.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:14:50,350 INFO [finetune.py:976] (0/7) Epoch 26, batch 1000, loss[loss=0.2368, simple_loss=0.3049, pruned_loss=0.08441, over 4758.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2423, pruned_loss=0.0494, over 950231.41 frames. ], batch size: 54, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:14:59,105 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 07:15:20,614 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7633, 1.7439, 1.4921, 1.8928, 2.1863, 1.8679, 1.5716, 1.4536], + device='cuda:0'), covar=tensor([0.2064, 0.1881, 0.1867, 0.1573, 0.1680, 0.1191, 0.2296, 0.1851], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0215, 0.0198, 0.0244, 0.0191, 0.0216, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:15:22,310 INFO [finetune.py:976] (0/7) Epoch 26, batch 1050, loss[loss=0.2056, simple_loss=0.2809, pruned_loss=0.06511, over 4813.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2445, pruned_loss=0.05011, over 949878.86 frames. ], batch size: 45, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:15:40,002 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.483e+02 1.785e+02 2.219e+02 5.161e+02, threshold=3.570e+02, percent-clipped=2.0 +2023-03-27 07:16:01,565 INFO [finetune.py:976] (0/7) Epoch 26, batch 1100, loss[loss=0.2198, simple_loss=0.2825, pruned_loss=0.07857, over 4889.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.246, pruned_loss=0.05058, over 950717.20 frames. ], batch size: 35, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:16:03,512 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1372, 2.1372, 1.8057, 2.2802, 2.1588, 1.9451, 2.4604, 2.2345], + device='cuda:0'), covar=tensor([0.1231, 0.1793, 0.2536, 0.2073, 0.2094, 0.1412, 0.2639, 0.1421], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0188, 0.0233, 0.0251, 0.0247, 0.0205, 0.0212, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:16:45,110 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-27 07:16:45,525 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1795, 1.8476, 2.3968, 1.5902, 2.1839, 2.5276, 1.7384, 2.5806], + device='cuda:0'), covar=tensor([0.1371, 0.2265, 0.1546, 0.2220, 0.1086, 0.1494, 0.2904, 0.0918], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0205, 0.0191, 0.0190, 0.0174, 0.0212, 0.0216, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:16:56,473 INFO [finetune.py:976] (0/7) Epoch 26, batch 1150, loss[loss=0.158, simple_loss=0.2374, pruned_loss=0.03929, over 4916.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2468, pruned_loss=0.05049, over 952369.01 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:17:13,862 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.499e+02 1.760e+02 2.197e+02 4.327e+02, threshold=3.521e+02, percent-clipped=2.0 +2023-03-27 07:17:28,601 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2995, 2.2393, 2.2445, 1.7381, 2.0447, 2.5094, 2.3996, 1.8906], + device='cuda:0'), covar=tensor([0.0569, 0.0629, 0.0720, 0.0854, 0.1621, 0.0559, 0.0503, 0.1017], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0120, 0.0128, 0.0139, 0.0141, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:17:30,185 INFO [finetune.py:976] (0/7) Epoch 26, batch 1200, loss[loss=0.1876, simple_loss=0.2563, pruned_loss=0.05951, over 4815.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2448, pruned_loss=0.0497, over 952871.28 frames. ], batch size: 40, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:17:39,539 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 07:17:43,720 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144411.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:18:03,373 INFO [finetune.py:976] (0/7) Epoch 26, batch 1250, loss[loss=0.1465, simple_loss=0.2295, pruned_loss=0.03172, over 4822.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2443, pruned_loss=0.05037, over 954975.18 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:18:11,744 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-27 07:18:21,713 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.894e+01 1.606e+02 1.805e+02 2.274e+02 3.881e+02, threshold=3.611e+02, percent-clipped=1.0 +2023-03-27 07:18:24,300 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:18:37,275 INFO [finetune.py:976] (0/7) Epoch 26, batch 1300, loss[loss=0.1907, simple_loss=0.2589, pruned_loss=0.06122, over 4826.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2417, pruned_loss=0.04941, over 956089.02 frames. ], batch size: 39, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:19:00,390 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5095, 2.3212, 1.7468, 0.8343, 2.0451, 2.0577, 1.9038, 2.1215], + device='cuda:0'), covar=tensor([0.0916, 0.0827, 0.1524, 0.2048, 0.1348, 0.2297, 0.2178, 0.0882], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0211, 0.0211, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:19:21,353 INFO [finetune.py:976] (0/7) Epoch 26, batch 1350, loss[loss=0.1446, simple_loss=0.2182, pruned_loss=0.03546, over 4789.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2425, pruned_loss=0.05015, over 954066.36 frames. ], batch size: 29, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:19:39,470 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.499e+02 1.803e+02 2.073e+02 4.281e+02, threshold=3.607e+02, percent-clipped=1.0 +2023-03-27 07:19:52,753 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4953, 1.3569, 1.2453, 1.4735, 1.6128, 1.5649, 1.0431, 1.2764], + device='cuda:0'), covar=tensor([0.2025, 0.1927, 0.1811, 0.1606, 0.1594, 0.1183, 0.2413, 0.1758], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0216, 0.0199, 0.0245, 0.0191, 0.0216, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:19:54,498 INFO [finetune.py:976] (0/7) Epoch 26, batch 1400, loss[loss=0.158, simple_loss=0.2105, pruned_loss=0.05277, over 4016.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2436, pruned_loss=0.05002, over 952034.11 frames. ], batch size: 17, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:20:21,867 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-27 07:20:22,400 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1834, 1.9536, 1.7222, 1.6688, 1.9153, 1.8994, 1.9430, 2.6485], + device='cuda:0'), covar=tensor([0.3628, 0.4420, 0.3397, 0.3854, 0.3715, 0.2698, 0.3657, 0.1699], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0236, 0.0277, 0.0258, 0.0229, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:20:27,738 INFO [finetune.py:976] (0/7) Epoch 26, batch 1450, loss[loss=0.2238, simple_loss=0.2848, pruned_loss=0.08145, over 4868.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2464, pruned_loss=0.05163, over 953105.05 frames. ], batch size: 34, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:20:45,824 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.092e+01 1.571e+02 1.855e+02 2.334e+02 4.645e+02, threshold=3.710e+02, percent-clipped=2.0 +2023-03-27 07:21:01,002 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 07:21:01,362 INFO [finetune.py:976] (0/7) Epoch 26, batch 1500, loss[loss=0.1398, simple_loss=0.2081, pruned_loss=0.0357, over 4485.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2469, pruned_loss=0.05114, over 952537.40 frames. ], batch size: 19, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:21:50,337 INFO [finetune.py:976] (0/7) Epoch 26, batch 1550, loss[loss=0.1786, simple_loss=0.2427, pruned_loss=0.05729, over 4866.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2465, pruned_loss=0.05091, over 952991.84 frames. ], batch size: 34, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:21:58,267 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.06 vs. limit=5.0 +2023-03-27 07:22:18,425 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=144767.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:22:18,964 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.547e+02 1.850e+02 2.044e+02 4.068e+02, threshold=3.700e+02, percent-clipped=1.0 +2023-03-27 07:22:30,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7658, 1.5873, 1.8583, 1.2631, 1.7667, 1.9069, 1.5630, 2.1131], + device='cuda:0'), covar=tensor([0.1202, 0.1969, 0.1240, 0.1730, 0.0875, 0.1204, 0.2610, 0.0828], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0204, 0.0189, 0.0188, 0.0172, 0.0211, 0.0213, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:22:35,464 INFO [finetune.py:976] (0/7) Epoch 26, batch 1600, loss[loss=0.2018, simple_loss=0.2485, pruned_loss=0.07758, over 4252.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.245, pruned_loss=0.0503, over 951578.24 frames. ], batch size: 65, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:22:43,142 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 07:23:09,211 INFO [finetune.py:976] (0/7) Epoch 26, batch 1650, loss[loss=0.2095, simple_loss=0.2862, pruned_loss=0.06636, over 4902.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2441, pruned_loss=0.05057, over 954874.36 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:23:17,750 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3003, 1.4478, 1.4454, 0.8444, 1.4890, 1.6232, 1.7274, 1.3760], + device='cuda:0'), covar=tensor([0.0991, 0.0615, 0.0588, 0.0529, 0.0477, 0.0687, 0.0315, 0.0671], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0147, 0.0128, 0.0122, 0.0130, 0.0129, 0.0142, 0.0148], + device='cuda:0'), out_proj_covar=tensor([8.9098e-05, 1.0608e-04, 9.0942e-05, 8.5846e-05, 9.0926e-05, 9.1489e-05, + 1.0098e-04, 1.0605e-04], device='cuda:0') +2023-03-27 07:23:26,320 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.503e+01 1.460e+02 1.738e+02 2.010e+02 3.428e+02, threshold=3.475e+02, percent-clipped=0.0 +2023-03-27 07:23:42,420 INFO [finetune.py:976] (0/7) Epoch 26, batch 1700, loss[loss=0.1955, simple_loss=0.2585, pruned_loss=0.06627, over 4759.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2416, pruned_loss=0.04952, over 957075.90 frames. ], batch size: 54, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:24:25,823 INFO [finetune.py:976] (0/7) Epoch 26, batch 1750, loss[loss=0.1939, simple_loss=0.2671, pruned_loss=0.06037, over 4850.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2444, pruned_loss=0.05055, over 957358.31 frames. ], batch size: 49, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:24:42,914 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.967e+01 1.592e+02 1.823e+02 2.389e+02 4.337e+02, threshold=3.645e+02, percent-clipped=3.0 +2023-03-27 07:24:45,902 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6598, 1.1437, 0.8938, 1.5438, 1.9909, 1.3589, 1.4147, 1.4652], + device='cuda:0'), covar=tensor([0.1445, 0.2124, 0.1871, 0.1160, 0.1975, 0.1921, 0.1461, 0.2017], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:24:59,584 INFO [finetune.py:976] (0/7) Epoch 26, batch 1800, loss[loss=0.1682, simple_loss=0.2383, pruned_loss=0.04911, over 4880.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2486, pruned_loss=0.05174, over 957398.61 frames. ], batch size: 32, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:25:02,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144996.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:23,061 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:33,497 INFO [finetune.py:976] (0/7) Epoch 26, batch 1850, loss[loss=0.195, simple_loss=0.2617, pruned_loss=0.06416, over 4795.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2479, pruned_loss=0.05144, over 955608.65 frames. ], batch size: 45, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:25:43,141 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145057.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:25:48,485 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145066.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:49,079 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145067.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:49,580 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.533e+02 1.829e+02 2.183e+02 4.392e+02, threshold=3.659e+02, percent-clipped=3.0 +2023-03-27 07:25:55,607 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 07:26:03,668 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:06,519 INFO [finetune.py:976] (0/7) Epoch 26, batch 1900, loss[loss=0.1997, simple_loss=0.2689, pruned_loss=0.06522, over 4888.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2476, pruned_loss=0.05087, over 957550.87 frames. ], batch size: 43, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:26:21,294 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145115.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:29,577 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:45,616 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6078, 1.5156, 1.4883, 1.6129, 1.1427, 3.3119, 1.2780, 1.7016], + device='cuda:0'), covar=tensor([0.3460, 0.2667, 0.2233, 0.2448, 0.1802, 0.0204, 0.2614, 0.1332], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:26:46,244 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1152, 2.0033, 1.7936, 2.0577, 1.8498, 1.8312, 1.8823, 2.6378], + device='cuda:0'), covar=tensor([0.3341, 0.3572, 0.3003, 0.3473, 0.3836, 0.2244, 0.3648, 0.1459], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0264, 0.0235, 0.0275, 0.0258, 0.0228, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:26:46,687 INFO [finetune.py:976] (0/7) Epoch 26, batch 1950, loss[loss=0.1685, simple_loss=0.2267, pruned_loss=0.05514, over 4847.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2471, pruned_loss=0.05121, over 956132.12 frames. ], batch size: 49, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:26:57,619 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7807, 1.3808, 0.8592, 1.6350, 2.2488, 1.3347, 1.6721, 1.5751], + device='cuda:0'), covar=tensor([0.1468, 0.1916, 0.1864, 0.1209, 0.1817, 0.1820, 0.1387, 0.1967], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:27:06,387 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4361, 1.3987, 1.5959, 2.4730, 1.6784, 2.2289, 0.9546, 2.1740], + device='cuda:0'), covar=tensor([0.1543, 0.1250, 0.1043, 0.0682, 0.0893, 0.1203, 0.1458, 0.0529], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0164, 0.0101, 0.0135, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 07:27:15,912 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.577e+02 1.831e+02 2.188e+02 4.363e+02, threshold=3.662e+02, percent-clipped=3.0 +2023-03-27 07:27:33,091 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0103, 2.2522, 1.9090, 1.8719, 2.6620, 2.7122, 2.1683, 2.1025], + device='cuda:0'), covar=tensor([0.0557, 0.0312, 0.0603, 0.0348, 0.0218, 0.0391, 0.0409, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0108, 0.0149, 0.0112, 0.0102, 0.0117, 0.0104, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.9592e-05, 8.2567e-05, 1.1592e-04, 8.6029e-05, 7.9356e-05, 8.6271e-05, + 7.7143e-05, 8.6779e-05], device='cuda:0') +2023-03-27 07:27:40,103 INFO [finetune.py:976] (0/7) Epoch 26, batch 2000, loss[loss=0.1768, simple_loss=0.2532, pruned_loss=0.05023, over 4828.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2437, pruned_loss=0.05025, over 955543.34 frames. ], batch size: 39, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:28:00,000 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7735, 1.7633, 1.5486, 1.9702, 2.2469, 1.9773, 1.5537, 1.5075], + device='cuda:0'), covar=tensor([0.2378, 0.2090, 0.2130, 0.1703, 0.1676, 0.1186, 0.2435, 0.2068], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0215, 0.0198, 0.0244, 0.0191, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:28:13,277 INFO [finetune.py:976] (0/7) Epoch 26, batch 2050, loss[loss=0.1586, simple_loss=0.2303, pruned_loss=0.04342, over 4914.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2404, pruned_loss=0.04907, over 955454.18 frames. ], batch size: 36, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:28:19,939 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4606, 1.6427, 1.6153, 0.8691, 1.7022, 1.9022, 1.9219, 1.4990], + device='cuda:0'), covar=tensor([0.0864, 0.0592, 0.0562, 0.0559, 0.0462, 0.0599, 0.0304, 0.0690], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0123, 0.0130, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.8954e-05, 1.0618e-04, 9.1031e-05, 8.6206e-05, 9.0856e-05, 9.1632e-05, + 1.0083e-04, 1.0626e-04], device='cuda:0') +2023-03-27 07:28:30,378 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.187e+01 1.436e+02 1.792e+02 2.264e+02 4.038e+02, threshold=3.583e+02, percent-clipped=1.0 +2023-03-27 07:28:45,865 INFO [finetune.py:976] (0/7) Epoch 26, batch 2100, loss[loss=0.2013, simple_loss=0.2738, pruned_loss=0.0644, over 4857.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2403, pruned_loss=0.04952, over 955428.90 frames. ], batch size: 49, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:29:03,661 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.5624, 3.2027, 2.9506, 1.6256, 3.1335, 2.6058, 2.4093, 2.8043], + device='cuda:0'), covar=tensor([0.0857, 0.0694, 0.1434, 0.1866, 0.1370, 0.1763, 0.1787, 0.0990], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0192, 0.0200, 0.0181, 0.0209, 0.0209, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:29:09,172 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2134, 1.4017, 1.4272, 0.7499, 1.3947, 1.6002, 1.6470, 1.3718], + device='cuda:0'), covar=tensor([0.0942, 0.0677, 0.0651, 0.0564, 0.0617, 0.0709, 0.0466, 0.0778], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0123, 0.0130, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9137e-05, 1.0644e-04, 9.1238e-05, 8.6419e-05, 9.1066e-05, 9.1707e-05, + 1.0097e-04, 1.0656e-04], device='cuda:0') +2023-03-27 07:29:19,662 INFO [finetune.py:976] (0/7) Epoch 26, batch 2150, loss[loss=0.1284, simple_loss=0.2134, pruned_loss=0.02164, over 4784.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.242, pruned_loss=0.04932, over 954105.00 frames. ], batch size: 29, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:29:28,937 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145352.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:29:39,076 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0163, 1.8684, 1.6420, 1.7910, 1.7868, 1.7670, 1.8607, 2.4991], + device='cuda:0'), covar=tensor([0.3544, 0.3937, 0.3181, 0.3675, 0.3971, 0.2523, 0.3428, 0.1665], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0237, 0.0278, 0.0260, 0.0230, 0.0258, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:29:43,500 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8458, 3.3569, 3.5957, 3.7284, 3.6134, 3.3989, 3.9022, 1.2140], + device='cuda:0'), covar=tensor([0.0897, 0.0935, 0.0915, 0.1080, 0.1348, 0.1657, 0.0855, 0.5783], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0249, 0.0280, 0.0295, 0.0337, 0.0287, 0.0304, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:29:47,542 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.494e+02 1.709e+02 2.298e+02 6.165e+02, threshold=3.419e+02, percent-clipped=3.0 +2023-03-27 07:29:49,487 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:29:56,783 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145383.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:02,627 INFO [finetune.py:976] (0/7) Epoch 26, batch 2200, loss[loss=0.1934, simple_loss=0.273, pruned_loss=0.0569, over 4809.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05002, over 956057.75 frames. ], batch size: 40, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:30:03,785 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4763, 1.3781, 2.0153, 3.1442, 2.0493, 2.2270, 1.1663, 2.7088], + device='cuda:0'), covar=tensor([0.1695, 0.1347, 0.1107, 0.0503, 0.0801, 0.1377, 0.1511, 0.0388], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0101, 0.0134, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 07:30:22,243 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 07:30:23,255 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145422.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:29,365 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:36,242 INFO [finetune.py:976] (0/7) Epoch 26, batch 2250, loss[loss=0.191, simple_loss=0.2577, pruned_loss=0.06213, over 4832.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2456, pruned_loss=0.05024, over 953951.09 frames. ], batch size: 49, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:30:37,628 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3582, 1.2897, 1.2087, 1.2377, 1.6239, 1.5320, 1.3454, 1.2018], + device='cuda:0'), covar=tensor([0.0393, 0.0324, 0.0648, 0.0368, 0.0256, 0.0423, 0.0378, 0.0444], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0109, 0.0149, 0.0113, 0.0103, 0.0118, 0.0104, 0.0115], + device='cuda:0'), out_proj_covar=tensor([8.0326e-05, 8.3024e-05, 1.1643e-04, 8.6571e-05, 7.9891e-05, 8.6772e-05, + 7.7549e-05, 8.7307e-05], device='cuda:0') +2023-03-27 07:30:44,584 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5818, 1.4763, 2.1769, 1.8360, 1.7590, 4.0784, 1.6160, 1.6048], + device='cuda:0'), covar=tensor([0.1009, 0.1870, 0.1307, 0.0945, 0.1552, 0.0185, 0.1403, 0.1815], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0072, 0.0076, 0.0090, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:30:53,946 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.189e+01 1.488e+02 1.760e+02 2.143e+02 3.776e+02, threshold=3.521e+02, percent-clipped=2.0 +2023-03-27 07:31:08,986 INFO [finetune.py:976] (0/7) Epoch 26, batch 2300, loss[loss=0.1798, simple_loss=0.2445, pruned_loss=0.05751, over 4811.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2452, pruned_loss=0.04951, over 954747.96 frames. ], batch size: 39, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:31:41,340 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7281, 3.9639, 3.6970, 1.8282, 3.9782, 2.8997, 0.8369, 2.8254], + device='cuda:0'), covar=tensor([0.2128, 0.1526, 0.1558, 0.3250, 0.0979, 0.1054, 0.4433, 0.1292], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0179, 0.0160, 0.0130, 0.0161, 0.0124, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 07:31:42,487 INFO [finetune.py:976] (0/7) Epoch 26, batch 2350, loss[loss=0.1769, simple_loss=0.2417, pruned_loss=0.05609, over 4827.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2433, pruned_loss=0.04888, over 954565.82 frames. ], batch size: 39, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:31:53,140 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0715, 2.0179, 2.0372, 1.2512, 2.0544, 2.0698, 2.0464, 1.7321], + device='cuda:0'), covar=tensor([0.0509, 0.0675, 0.0692, 0.0936, 0.0752, 0.0656, 0.0608, 0.1108], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0119, 0.0128, 0.0138, 0.0139, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:32:06,623 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.544e+02 1.840e+02 2.226e+02 4.643e+02, threshold=3.680e+02, percent-clipped=1.0 +2023-03-27 07:32:34,348 INFO [finetune.py:976] (0/7) Epoch 26, batch 2400, loss[loss=0.1813, simple_loss=0.2496, pruned_loss=0.05655, over 4846.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2409, pruned_loss=0.04882, over 955372.05 frames. ], batch size: 47, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:32:35,678 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145594.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:32:58,801 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3978, 3.8447, 4.0394, 4.2361, 4.1661, 3.8202, 4.4656, 1.4053], + device='cuda:0'), covar=tensor([0.0758, 0.0799, 0.0836, 0.0892, 0.1272, 0.1516, 0.0672, 0.5672], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0249, 0.0280, 0.0296, 0.0338, 0.0287, 0.0304, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:32:58,943 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-03-27 07:33:15,387 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6940, 3.6371, 3.4880, 1.8932, 3.7355, 2.8325, 0.8374, 2.6395], + device='cuda:0'), covar=tensor([0.2664, 0.1915, 0.1646, 0.3003, 0.0950, 0.1015, 0.4459, 0.1372], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0179, 0.0160, 0.0131, 0.0161, 0.0124, 0.0149, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 07:33:17,100 INFO [finetune.py:976] (0/7) Epoch 26, batch 2450, loss[loss=0.1658, simple_loss=0.2402, pruned_loss=0.04573, over 4871.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2388, pruned_loss=0.0482, over 955887.75 frames. ], batch size: 31, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:33:17,845 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:23,873 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145652.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:33:26,195 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145655.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:33:34,919 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.521e+01 1.449e+02 1.824e+02 2.163e+02 4.630e+02, threshold=3.648e+02, percent-clipped=2.0 +2023-03-27 07:33:45,640 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145683.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:51,043 INFO [finetune.py:976] (0/7) Epoch 26, batch 2500, loss[loss=0.1549, simple_loss=0.2365, pruned_loss=0.03669, over 4896.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.239, pruned_loss=0.04819, over 956108.09 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:33:55,945 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145700.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:58,923 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:09,999 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0074, 1.8530, 1.7178, 1.7756, 1.7920, 1.7467, 1.8391, 2.4005], + device='cuda:0'), covar=tensor([0.2969, 0.3009, 0.2543, 0.2598, 0.3191, 0.2055, 0.2839, 0.1354], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0235, 0.0275, 0.0258, 0.0228, 0.0256, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:34:11,629 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145722.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:15,148 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145727.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:17,567 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145731.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:24,602 INFO [finetune.py:976] (0/7) Epoch 26, batch 2550, loss[loss=0.1452, simple_loss=0.2213, pruned_loss=0.03461, over 4704.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2409, pruned_loss=0.04845, over 954632.11 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:34:42,325 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.531e+01 1.551e+02 1.807e+02 2.106e+02 4.459e+02, threshold=3.615e+02, percent-clipped=2.0 +2023-03-27 07:34:43,538 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145770.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:35:08,874 INFO [finetune.py:976] (0/7) Epoch 26, batch 2600, loss[loss=0.149, simple_loss=0.2218, pruned_loss=0.03811, over 4884.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2443, pruned_loss=0.04981, over 955110.63 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:35:28,211 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145821.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:35:35,875 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1455, 1.2727, 1.3534, 0.7133, 1.3348, 1.5824, 1.5868, 1.2833], + device='cuda:0'), covar=tensor([0.0917, 0.0608, 0.0540, 0.0497, 0.0509, 0.0537, 0.0352, 0.0655], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0122, 0.0130, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.8925e-05, 1.0617e-04, 9.1291e-05, 8.6019e-05, 9.0989e-05, 9.1421e-05, + 1.0087e-04, 1.0643e-04], device='cuda:0') +2023-03-27 07:35:42,707 INFO [finetune.py:976] (0/7) Epoch 26, batch 2650, loss[loss=0.1709, simple_loss=0.2599, pruned_loss=0.04097, over 4774.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2457, pruned_loss=0.05028, over 954172.06 frames. ], batch size: 28, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:35:55,353 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5914, 1.5661, 2.0473, 1.8116, 1.7518, 3.0986, 1.5326, 1.6663], + device='cuda:0'), covar=tensor([0.0941, 0.1550, 0.1263, 0.0867, 0.1305, 0.0327, 0.1316, 0.1587], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:36:00,019 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.986e+01 1.512e+02 1.783e+02 2.110e+02 4.476e+02, threshold=3.566e+02, percent-clipped=1.0 +2023-03-27 07:36:09,553 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145882.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:36:16,424 INFO [finetune.py:976] (0/7) Epoch 26, batch 2700, loss[loss=0.1569, simple_loss=0.2291, pruned_loss=0.04232, over 4921.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2447, pruned_loss=0.0495, over 954166.01 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:36:48,573 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8757, 1.4771, 0.7963, 1.6489, 2.2127, 1.4892, 1.6029, 1.6444], + device='cuda:0'), covar=tensor([0.1362, 0.1888, 0.1975, 0.1198, 0.1830, 0.1790, 0.1352, 0.1991], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:36:49,651 INFO [finetune.py:976] (0/7) Epoch 26, batch 2750, loss[loss=0.1723, simple_loss=0.2385, pruned_loss=0.05304, over 4900.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2414, pruned_loss=0.0483, over 955281.68 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:36:53,367 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8474, 1.4335, 1.9900, 1.8629, 1.6728, 1.6713, 1.8351, 1.9020], + device='cuda:0'), covar=tensor([0.4233, 0.4132, 0.3229, 0.3896, 0.4737, 0.3940, 0.4550, 0.2939], + device='cuda:0'), in_proj_covar=tensor([0.0266, 0.0247, 0.0266, 0.0294, 0.0294, 0.0271, 0.0300, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:36:55,116 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145950.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:36:56,579 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-27 07:37:00,657 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6077, 1.4508, 1.5233, 0.7959, 1.6761, 1.9128, 1.6592, 1.4463], + device='cuda:0'), covar=tensor([0.1047, 0.1014, 0.0667, 0.0677, 0.0647, 0.0658, 0.0583, 0.0779], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0122, 0.0130, 0.0129, 0.0142, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.9061e-05, 1.0631e-04, 9.1093e-05, 8.5941e-05, 9.1062e-05, 9.1470e-05, + 1.0093e-04, 1.0625e-04], device='cuda:0') +2023-03-27 07:37:07,588 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.569e+02 1.804e+02 2.200e+02 3.850e+02, threshold=3.609e+02, percent-clipped=2.0 +2023-03-27 07:37:29,448 INFO [finetune.py:976] (0/7) Epoch 26, batch 2800, loss[loss=0.1743, simple_loss=0.235, pruned_loss=0.05678, over 4901.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2392, pruned_loss=0.0481, over 956896.23 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:37:38,978 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145999.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:37:40,153 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-146000.pt +2023-03-27 07:38:14,436 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:38:16,767 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4883, 1.3581, 1.5231, 0.8791, 1.4801, 1.5487, 1.5599, 1.3053], + device='cuda:0'), covar=tensor([0.0540, 0.0803, 0.0699, 0.0882, 0.1019, 0.0669, 0.0608, 0.1211], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0120, 0.0128, 0.0138, 0.0140, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:38:17,418 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2093, 2.1121, 1.7385, 1.8384, 2.1867, 1.9335, 2.3100, 2.2190], + device='cuda:0'), covar=tensor([0.1316, 0.1991, 0.3059, 0.2546, 0.2581, 0.1721, 0.3154, 0.1751], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0189, 0.0235, 0.0251, 0.0249, 0.0206, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:38:21,624 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-27 07:38:24,986 INFO [finetune.py:976] (0/7) Epoch 26, batch 2850, loss[loss=0.1949, simple_loss=0.2522, pruned_loss=0.06876, over 4721.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2385, pruned_loss=0.04778, over 956829.52 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:38:37,436 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8899, 3.7305, 3.5903, 1.7258, 3.9188, 2.9659, 1.0236, 2.7244], + device='cuda:0'), covar=tensor([0.2219, 0.2122, 0.1656, 0.3626, 0.0940, 0.0982, 0.4584, 0.1546], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0177, 0.0158, 0.0129, 0.0159, 0.0122, 0.0147, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 07:38:42,241 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.185e+01 1.523e+02 1.819e+02 2.110e+02 4.930e+02, threshold=3.638e+02, percent-clipped=2.0 +2023-03-27 07:38:46,430 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:38:58,297 INFO [finetune.py:976] (0/7) Epoch 26, batch 2900, loss[loss=0.151, simple_loss=0.222, pruned_loss=0.03997, over 4787.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2401, pruned_loss=0.04824, over 957567.26 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:39:31,499 INFO [finetune.py:976] (0/7) Epoch 26, batch 2950, loss[loss=0.1567, simple_loss=0.2348, pruned_loss=0.03935, over 4818.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2425, pruned_loss=0.04813, over 959414.55 frames. ], batch size: 39, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:39:36,383 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146149.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:39:49,295 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.550e+02 1.859e+02 2.106e+02 3.478e+02, threshold=3.719e+02, percent-clipped=0.0 +2023-03-27 07:39:54,199 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146177.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:40:04,822 INFO [finetune.py:976] (0/7) Epoch 26, batch 3000, loss[loss=0.1631, simple_loss=0.2453, pruned_loss=0.04044, over 4921.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2431, pruned_loss=0.04799, over 958507.13 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:40:04,823 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 07:40:19,938 INFO [finetune.py:1010] (0/7) Epoch 26, validation: loss=0.1577, simple_loss=0.2252, pruned_loss=0.04507, over 2265189.00 frames. +2023-03-27 07:40:19,938 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 07:40:35,400 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146210.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:40:40,169 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5089, 1.4101, 1.5198, 0.7695, 1.5158, 1.5630, 1.5151, 1.2953], + device='cuda:0'), covar=tensor([0.0617, 0.0854, 0.0776, 0.1010, 0.0958, 0.0748, 0.0636, 0.1359], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0120, 0.0128, 0.0138, 0.0139, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:40:53,996 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4795, 1.3684, 1.9228, 1.8280, 1.5387, 3.5505, 1.3349, 1.5345], + device='cuda:0'), covar=tensor([0.0969, 0.1910, 0.1116, 0.0933, 0.1724, 0.0206, 0.1581, 0.1860], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 07:40:56,711 INFO [finetune.py:976] (0/7) Epoch 26, batch 3050, loss[loss=0.134, simple_loss=0.2091, pruned_loss=0.02944, over 4697.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2446, pruned_loss=0.04848, over 958347.24 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:41:02,118 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146250.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:41:14,894 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.646e+01 1.408e+02 1.795e+02 2.163e+02 4.679e+02, threshold=3.589e+02, percent-clipped=3.0 +2023-03-27 07:41:29,847 INFO [finetune.py:976] (0/7) Epoch 26, batch 3100, loss[loss=0.1805, simple_loss=0.2422, pruned_loss=0.05944, over 4904.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04851, over 956305.92 frames. ], batch size: 36, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:41:34,029 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146298.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:41:34,709 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:42:02,609 INFO [finetune.py:976] (0/7) Epoch 26, batch 3150, loss[loss=0.1671, simple_loss=0.2349, pruned_loss=0.04967, over 4738.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2418, pruned_loss=0.04858, over 955952.24 frames. ], batch size: 59, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:42:06,558 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:42:21,143 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.464e+02 1.851e+02 2.163e+02 3.423e+02, threshold=3.701e+02, percent-clipped=0.0 +2023-03-27 07:42:26,726 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1997, 1.9792, 1.7608, 1.8034, 1.8924, 1.9360, 1.9331, 2.6878], + device='cuda:0'), covar=tensor([0.3396, 0.3991, 0.3149, 0.3742, 0.4038, 0.2361, 0.3653, 0.1612], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0237, 0.0277, 0.0259, 0.0230, 0.0257, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:42:31,611 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9247, 1.2393, 1.9077, 1.9029, 1.7093, 1.6544, 1.8400, 1.8487], + device='cuda:0'), covar=tensor([0.3494, 0.3524, 0.2845, 0.3178, 0.4190, 0.3395, 0.3648, 0.2616], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0248, 0.0268, 0.0295, 0.0296, 0.0272, 0.0302, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:42:38,056 INFO [finetune.py:976] (0/7) Epoch 26, batch 3200, loss[loss=0.1688, simple_loss=0.2398, pruned_loss=0.0489, over 4789.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.239, pruned_loss=0.04795, over 956060.39 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:42:40,013 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.4417, 3.1213, 2.8909, 1.4962, 3.0602, 2.4916, 2.4192, 2.9250], + device='cuda:0'), covar=tensor([0.0887, 0.0720, 0.1721, 0.2160, 0.1579, 0.2116, 0.2043, 0.1039], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0191, 0.0199, 0.0181, 0.0208, 0.0209, 0.0222, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:43:35,343 INFO [finetune.py:976] (0/7) Epoch 26, batch 3250, loss[loss=0.2055, simple_loss=0.2682, pruned_loss=0.07134, over 4145.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2391, pruned_loss=0.04798, over 954693.77 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:43:43,813 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5798, 1.4832, 1.4475, 1.4856, 1.2357, 3.3706, 1.3463, 1.6495], + device='cuda:0'), covar=tensor([0.3317, 0.2560, 0.2164, 0.2404, 0.1717, 0.0208, 0.2741, 0.1341], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0115, 0.0120, 0.0123, 0.0112, 0.0095, 0.0093, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:43:53,733 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.451e+02 1.808e+02 2.175e+02 4.535e+02, threshold=3.616e+02, percent-clipped=3.0 +2023-03-27 07:43:58,665 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146477.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:44:08,661 INFO [finetune.py:976] (0/7) Epoch 26, batch 3300, loss[loss=0.1336, simple_loss=0.2128, pruned_loss=0.02714, over 4705.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.241, pruned_loss=0.04841, over 953119.33 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:44:17,138 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:44:30,705 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146525.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:44:41,549 INFO [finetune.py:976] (0/7) Epoch 26, batch 3350, loss[loss=0.1774, simple_loss=0.2499, pruned_loss=0.05245, over 4788.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2447, pruned_loss=0.04979, over 952484.74 frames. ], batch size: 51, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:45:00,361 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.852e+01 1.562e+02 1.841e+02 2.282e+02 4.006e+02, threshold=3.682e+02, percent-clipped=1.0 +2023-03-27 07:45:14,937 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146591.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:45:15,431 INFO [finetune.py:976] (0/7) Epoch 26, batch 3400, loss[loss=0.1962, simple_loss=0.2633, pruned_loss=0.06451, over 4894.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2464, pruned_loss=0.05, over 952942.05 frames. ], batch size: 36, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:45:48,130 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:45:58,655 INFO [finetune.py:976] (0/7) Epoch 26, batch 3450, loss[loss=0.1534, simple_loss=0.2205, pruned_loss=0.04313, over 4873.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2452, pruned_loss=0.04923, over 953870.10 frames. ], batch size: 31, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:46:04,892 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146652.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:46:17,058 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.450e+02 1.708e+02 2.017e+02 4.995e+02, threshold=3.417e+02, percent-clipped=1.0 +2023-03-27 07:46:22,481 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6021, 1.0736, 0.7995, 1.3623, 2.0659, 0.8888, 1.3038, 1.3077], + device='cuda:0'), covar=tensor([0.1546, 0.2220, 0.1818, 0.1324, 0.1859, 0.1956, 0.1575, 0.2100], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:46:27,959 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0827, 2.7157, 2.6321, 1.3401, 2.7361, 2.1833, 2.1319, 2.5478], + device='cuda:0'), covar=tensor([0.1040, 0.0882, 0.1634, 0.2285, 0.1513, 0.2577, 0.2187, 0.1217], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0200, 0.0182, 0.0210, 0.0211, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:46:28,570 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146687.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:46:32,390 INFO [finetune.py:976] (0/7) Epoch 26, batch 3500, loss[loss=0.169, simple_loss=0.246, pruned_loss=0.046, over 4758.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2432, pruned_loss=0.0488, over 954482.03 frames. ], batch size: 28, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:05,274 INFO [finetune.py:976] (0/7) Epoch 26, batch 3550, loss[loss=0.1677, simple_loss=0.2404, pruned_loss=0.0475, over 4907.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2398, pruned_loss=0.04772, over 955312.56 frames. ], batch size: 36, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:22,722 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.430e+02 1.748e+02 2.315e+02 5.079e+02, threshold=3.497e+02, percent-clipped=6.0 +2023-03-27 07:47:32,138 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7528, 1.3190, 0.8032, 1.5176, 2.1357, 1.0872, 1.5234, 1.5864], + device='cuda:0'), covar=tensor([0.1428, 0.1899, 0.1858, 0.1119, 0.1821, 0.1922, 0.1332, 0.1848], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 07:47:38,113 INFO [finetune.py:976] (0/7) Epoch 26, batch 3600, loss[loss=0.1481, simple_loss=0.2296, pruned_loss=0.03326, over 4858.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2371, pruned_loss=0.04678, over 954240.68 frames. ], batch size: 44, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:47,319 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146805.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:08,912 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 07:48:24,412 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0213, 1.6382, 2.2927, 1.4089, 2.0988, 2.2104, 1.5638, 2.2449], + device='cuda:0'), covar=tensor([0.1135, 0.1926, 0.1301, 0.2010, 0.0876, 0.1383, 0.2604, 0.0841], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0192, 0.0189, 0.0174, 0.0212, 0.0216, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:48:26,739 INFO [finetune.py:976] (0/7) Epoch 26, batch 3650, loss[loss=0.1713, simple_loss=0.2489, pruned_loss=0.04682, over 4920.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2386, pruned_loss=0.04718, over 955863.62 frames. ], batch size: 42, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:48:28,100 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146844.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:37,462 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146853.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:53,566 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.488e+01 1.511e+02 1.813e+02 2.229e+02 3.524e+02, threshold=3.627e+02, percent-clipped=1.0 +2023-03-27 07:49:12,952 INFO [finetune.py:976] (0/7) Epoch 26, batch 3700, loss[loss=0.1957, simple_loss=0.2672, pruned_loss=0.06207, over 4812.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2411, pruned_loss=0.04832, over 953586.76 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:49:21,428 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146905.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:49:27,361 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3084, 2.9145, 3.0765, 3.2388, 3.0592, 2.8492, 3.3503, 1.0072], + device='cuda:0'), covar=tensor([0.1096, 0.1200, 0.1115, 0.1255, 0.1722, 0.1879, 0.1096, 0.5754], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0248, 0.0281, 0.0294, 0.0337, 0.0286, 0.0303, 0.0300], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:49:33,428 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-27 07:49:46,521 INFO [finetune.py:976] (0/7) Epoch 26, batch 3750, loss[loss=0.1619, simple_loss=0.2343, pruned_loss=0.04474, over 4904.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2438, pruned_loss=0.04976, over 952219.67 frames. ], batch size: 46, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:49:49,607 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:49:57,545 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-27 07:50:03,845 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.820e+01 1.503e+02 1.791e+02 2.461e+02 5.017e+02, threshold=3.581e+02, percent-clipped=5.0 +2023-03-27 07:50:09,251 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7805, 1.7433, 1.6397, 1.7726, 1.4652, 4.3149, 1.6223, 1.9564], + device='cuda:0'), covar=tensor([0.3321, 0.2500, 0.2134, 0.2251, 0.1490, 0.0137, 0.2376, 0.1249], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:50:12,661 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146982.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:50:14,435 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7427, 2.6417, 2.3893, 2.9391, 2.6714, 2.5987, 2.6039, 3.4954], + device='cuda:0'), covar=tensor([0.3424, 0.4835, 0.3122, 0.3538, 0.3786, 0.2492, 0.3516, 0.1616], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0237, 0.0278, 0.0260, 0.0230, 0.0258, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:50:19,582 INFO [finetune.py:976] (0/7) Epoch 26, batch 3800, loss[loss=0.1689, simple_loss=0.2506, pruned_loss=0.04359, over 4919.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2451, pruned_loss=0.05013, over 953892.21 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:50:55,309 INFO [finetune.py:976] (0/7) Epoch 26, batch 3850, loss[loss=0.1382, simple_loss=0.2093, pruned_loss=0.03354, over 4938.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2443, pruned_loss=0.05025, over 953667.92 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:51:21,146 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.517e+02 1.854e+02 2.266e+02 5.483e+02, threshold=3.707e+02, percent-clipped=2.0 +2023-03-27 07:51:37,019 INFO [finetune.py:976] (0/7) Epoch 26, batch 3900, loss[loss=0.1609, simple_loss=0.2333, pruned_loss=0.04425, over 4910.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2417, pruned_loss=0.04934, over 954985.88 frames. ], batch size: 35, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:00,907 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-27 07:52:09,602 INFO [finetune.py:976] (0/7) Epoch 26, batch 3950, loss[loss=0.1689, simple_loss=0.2316, pruned_loss=0.05307, over 4093.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.239, pruned_loss=0.04811, over 956130.43 frames. ], batch size: 17, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:27,910 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.492e+02 1.682e+02 1.976e+02 2.814e+02, threshold=3.365e+02, percent-clipped=0.0 +2023-03-27 07:52:42,786 INFO [finetune.py:976] (0/7) Epoch 26, batch 4000, loss[loss=0.1255, simple_loss=0.21, pruned_loss=0.02052, over 4708.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2382, pruned_loss=0.04805, over 951600.89 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:43,108 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-27 07:52:48,868 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147200.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:15,800 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:16,326 INFO [finetune.py:976] (0/7) Epoch 26, batch 4050, loss[loss=0.198, simple_loss=0.2707, pruned_loss=0.06268, over 4908.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2427, pruned_loss=0.04982, over 951418.89 frames. ], batch size: 36, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:53:21,646 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147247.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:30,812 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.2724, 4.6683, 4.8512, 5.1467, 4.9983, 4.6657, 5.3795, 1.6604], + device='cuda:0'), covar=tensor([0.0676, 0.0776, 0.0708, 0.0865, 0.1182, 0.1500, 0.0441, 0.5692], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0249, 0.0282, 0.0296, 0.0339, 0.0288, 0.0304, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:53:48,586 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.660e+02 1.920e+02 2.375e+02 4.575e+02, threshold=3.840e+02, percent-clipped=6.0 +2023-03-27 07:53:58,888 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:08,318 INFO [finetune.py:976] (0/7) Epoch 26, batch 4100, loss[loss=0.1697, simple_loss=0.2481, pruned_loss=0.04566, over 4816.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2447, pruned_loss=0.05014, over 953515.98 frames. ], batch size: 45, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:54:13,961 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147295.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:18,774 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147302.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:24,614 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147310.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:37,221 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147330.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:44,910 INFO [finetune.py:976] (0/7) Epoch 26, batch 4150, loss[loss=0.2083, simple_loss=0.2846, pruned_loss=0.06599, over 4924.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2459, pruned_loss=0.05029, over 954118.20 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:55:03,459 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.520e+02 1.873e+02 2.208e+02 5.004e+02, threshold=3.746e+02, percent-clipped=1.0 +2023-03-27 07:55:04,808 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:18,277 INFO [finetune.py:976] (0/7) Epoch 26, batch 4200, loss[loss=0.1643, simple_loss=0.2384, pruned_loss=0.04506, over 4798.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2448, pruned_loss=0.04954, over 954792.31 frames. ], batch size: 45, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:55:28,551 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147407.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:46,455 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147434.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:51,221 INFO [finetune.py:976] (0/7) Epoch 26, batch 4250, loss[loss=0.1541, simple_loss=0.2319, pruned_loss=0.03817, over 4824.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.243, pruned_loss=0.04887, over 956060.99 frames. ], batch size: 40, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:56:15,675 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6372, 1.5408, 1.4503, 1.5404, 1.1417, 3.0883, 1.3326, 1.6450], + device='cuda:0'), covar=tensor([0.3006, 0.2324, 0.2016, 0.2160, 0.1714, 0.0242, 0.2554, 0.1193], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 07:56:16,336 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:56:16,791 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.473e+02 1.815e+02 2.255e+02 8.587e+02, threshold=3.630e+02, percent-clipped=2.0 +2023-03-27 07:56:34,761 INFO [finetune.py:976] (0/7) Epoch 26, batch 4300, loss[loss=0.1316, simple_loss=0.199, pruned_loss=0.03208, over 4723.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.24, pruned_loss=0.04826, over 955968.84 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:56:36,725 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147495.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:56:40,206 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147500.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:57:08,536 INFO [finetune.py:976] (0/7) Epoch 26, batch 4350, loss[loss=0.16, simple_loss=0.2194, pruned_loss=0.05031, over 4820.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2373, pruned_loss=0.04717, over 957109.26 frames. ], batch size: 30, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:57:12,262 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:57:26,970 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.870e+01 1.462e+02 1.667e+02 1.917e+02 5.708e+02, threshold=3.333e+02, percent-clipped=2.0 +2023-03-27 07:57:29,998 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8827, 1.7241, 1.8820, 1.1925, 1.9200, 1.9906, 1.8943, 1.5437], + device='cuda:0'), covar=tensor([0.0561, 0.0801, 0.0723, 0.0897, 0.0901, 0.0656, 0.0621, 0.1267], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0135, 0.0139, 0.0118, 0.0126, 0.0137, 0.0138, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:57:42,399 INFO [finetune.py:976] (0/7) Epoch 26, batch 4400, loss[loss=0.1891, simple_loss=0.2613, pruned_loss=0.05842, over 4824.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.238, pruned_loss=0.04687, over 958223.83 frames. ], batch size: 39, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:57:45,500 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147597.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:58:16,299 INFO [finetune.py:976] (0/7) Epoch 26, batch 4450, loss[loss=0.1683, simple_loss=0.2524, pruned_loss=0.04213, over 4820.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2413, pruned_loss=0.04761, over 957927.75 frames. ], batch size: 40, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:58:21,325 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8902, 1.4658, 1.9971, 1.9310, 1.7450, 1.7191, 1.9187, 1.9083], + device='cuda:0'), covar=tensor([0.4027, 0.4187, 0.3290, 0.3860, 0.4954, 0.3895, 0.4531, 0.2927], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0247, 0.0267, 0.0294, 0.0294, 0.0270, 0.0301, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:58:27,307 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:58:34,814 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:58:36,564 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.542e+02 1.864e+02 2.192e+02 3.736e+02, threshold=3.727e+02, percent-clipped=4.0 +2023-03-27 07:58:36,699 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2809, 2.3000, 1.9902, 1.1248, 2.0900, 1.7917, 1.6823, 2.1876], + device='cuda:0'), covar=tensor([0.1018, 0.0685, 0.1382, 0.1895, 0.1418, 0.2303, 0.2299, 0.0859], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0193, 0.0201, 0.0183, 0.0211, 0.0212, 0.0226, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:59:06,168 INFO [finetune.py:976] (0/7) Epoch 26, batch 4500, loss[loss=0.1308, simple_loss=0.2025, pruned_loss=0.0295, over 4787.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.243, pruned_loss=0.04872, over 955991.56 frames. ], batch size: 29, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:59:15,027 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1497, 1.9923, 1.7315, 1.8789, 1.8585, 1.8589, 1.9655, 2.6087], + device='cuda:0'), covar=tensor([0.3366, 0.3890, 0.2966, 0.3447, 0.3879, 0.2148, 0.3421, 0.1568], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0237, 0.0277, 0.0260, 0.0230, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 07:59:37,020 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147720.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:59:52,138 INFO [finetune.py:976] (0/7) Epoch 26, batch 4550, loss[loss=0.2011, simple_loss=0.2581, pruned_loss=0.0721, over 4243.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2448, pruned_loss=0.04935, over 956655.44 frames. ], batch size: 65, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:00:04,820 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147763.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:05,976 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4389, 1.4106, 1.3862, 1.3666, 0.8841, 2.0737, 0.7136, 1.1157], + device='cuda:0'), covar=tensor([0.2872, 0.2169, 0.1837, 0.2099, 0.1541, 0.0328, 0.2317, 0.1166], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:00:09,335 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.722e+01 1.488e+02 1.756e+02 2.293e+02 4.562e+02, threshold=3.512e+02, percent-clipped=2.0 +2023-03-27 08:00:24,542 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147790.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:25,690 INFO [finetune.py:976] (0/7) Epoch 26, batch 4600, loss[loss=0.1297, simple_loss=0.2026, pruned_loss=0.02842, over 4791.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2444, pruned_loss=0.04924, over 957182.03 frames. ], batch size: 29, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:00:39,902 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:48,869 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 08:00:52,225 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2354, 1.9998, 1.8789, 2.2110, 2.7762, 2.2594, 2.3473, 1.8087], + device='cuda:0'), covar=tensor([0.1817, 0.1879, 0.1731, 0.1455, 0.1490, 0.1070, 0.1706, 0.1621], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0211, 0.0215, 0.0199, 0.0246, 0.0191, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:00:59,126 INFO [finetune.py:976] (0/7) Epoch 26, batch 4650, loss[loss=0.1286, simple_loss=0.2017, pruned_loss=0.02772, over 4862.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.241, pruned_loss=0.04816, over 957935.00 frames. ], batch size: 31, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:01:08,850 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147857.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:15,988 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.955e+01 1.458e+02 1.712e+02 2.175e+02 4.467e+02, threshold=3.424e+02, percent-clipped=3.0 +2023-03-27 08:01:21,845 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147876.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:39,635 INFO [finetune.py:976] (0/7) Epoch 26, batch 4700, loss[loss=0.1554, simple_loss=0.2272, pruned_loss=0.04184, over 4904.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2386, pruned_loss=0.04765, over 957259.58 frames. ], batch size: 35, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:01:46,514 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147897.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:55,415 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:59,698 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:16,231 INFO [finetune.py:976] (0/7) Epoch 26, batch 4750, loss[loss=0.1871, simple_loss=0.2584, pruned_loss=0.05795, over 4899.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2361, pruned_loss=0.04681, over 957492.61 frames. ], batch size: 43, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:02:18,601 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:27,475 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1398, 4.8752, 4.5367, 2.8226, 5.0373, 3.8510, 0.9669, 3.4593], + device='cuda:0'), covar=tensor([0.2232, 0.3096, 0.1591, 0.3040, 0.0826, 0.0905, 0.4968, 0.1622], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0179, 0.0160, 0.0129, 0.0160, 0.0124, 0.0147, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 08:02:29,345 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6525, 2.5176, 2.6481, 1.9883, 2.8418, 2.9811, 2.8219, 2.0215], + device='cuda:0'), covar=tensor([0.0714, 0.0866, 0.0876, 0.0997, 0.0644, 0.0771, 0.0786, 0.1683], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0136, 0.0141, 0.0120, 0.0127, 0.0139, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:02:32,358 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147966.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:34,089 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.470e+02 1.654e+02 2.049e+02 2.990e+02, threshold=3.309e+02, percent-clipped=0.0 +2023-03-27 08:02:35,998 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147972.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:41,796 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3654, 1.3662, 1.6091, 2.4733, 1.6771, 2.2353, 0.8953, 2.1959], + device='cuda:0'), covar=tensor([0.1726, 0.1420, 0.1197, 0.0799, 0.0939, 0.1256, 0.1592, 0.0616], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0134, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 08:02:50,079 INFO [finetune.py:976] (0/7) Epoch 26, batch 4800, loss[loss=0.2104, simple_loss=0.2839, pruned_loss=0.06847, over 4886.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2393, pruned_loss=0.04763, over 956491.99 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:02:51,386 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4443, 3.8543, 4.1010, 4.1600, 4.2208, 3.9536, 4.4825, 1.8249], + device='cuda:0'), covar=tensor([0.0697, 0.0868, 0.0764, 0.0864, 0.1087, 0.1483, 0.0684, 0.5540], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0248, 0.0283, 0.0296, 0.0339, 0.0287, 0.0307, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:02:56,041 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-148000.pt +2023-03-27 08:03:02,657 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2196, 2.0734, 1.7611, 2.0254, 1.9078, 1.9742, 1.9330, 2.6924], + device='cuda:0'), covar=tensor([0.3908, 0.4181, 0.3330, 0.4094, 0.4438, 0.2567, 0.3859, 0.1748], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0266, 0.0238, 0.0278, 0.0261, 0.0231, 0.0259, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:03:06,224 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:06,843 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148015.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:16,561 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-03-27 08:03:24,537 INFO [finetune.py:976] (0/7) Epoch 26, batch 4850, loss[loss=0.1908, simple_loss=0.262, pruned_loss=0.05977, over 4824.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2432, pruned_loss=0.04906, over 954560.39 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:03:39,246 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148063.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:42,785 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.530e+02 1.908e+02 2.333e+02 3.886e+02, threshold=3.817e+02, percent-clipped=4.0 +2023-03-27 08:04:04,122 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:04:05,256 INFO [finetune.py:976] (0/7) Epoch 26, batch 4900, loss[loss=0.1998, simple_loss=0.2686, pruned_loss=0.06551, over 4784.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.245, pruned_loss=0.04921, over 955271.17 frames. ], batch size: 29, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:04:30,817 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148111.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:04:48,566 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 08:04:57,724 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148138.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:05:00,565 INFO [finetune.py:976] (0/7) Epoch 26, batch 4950, loss[loss=0.1582, simple_loss=0.2346, pruned_loss=0.04092, over 4747.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2466, pruned_loss=0.05004, over 954705.03 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:05:18,900 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.030e+01 1.572e+02 1.871e+02 2.257e+02 5.603e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-27 08:05:20,229 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:05:21,483 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.3023, 2.9592, 3.0878, 3.2381, 3.0453, 2.8854, 3.3696, 0.9283], + device='cuda:0'), covar=tensor([0.1212, 0.1121, 0.1119, 0.1284, 0.2028, 0.2028, 0.1186, 0.6102], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0250, 0.0285, 0.0299, 0.0341, 0.0289, 0.0308, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:05:33,993 INFO [finetune.py:976] (0/7) Epoch 26, batch 5000, loss[loss=0.1513, simple_loss=0.2301, pruned_loss=0.03624, over 4935.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2443, pruned_loss=0.04937, over 954046.82 frames. ], batch size: 38, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:05:48,841 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:07,384 INFO [finetune.py:976] (0/7) Epoch 26, batch 5050, loss[loss=0.1621, simple_loss=0.2329, pruned_loss=0.04563, over 4895.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2423, pruned_loss=0.04911, over 955617.76 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:06:25,049 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148267.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:26,158 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.490e+02 1.796e+02 2.082e+02 4.496e+02, threshold=3.592e+02, percent-clipped=3.0 +2023-03-27 08:06:37,572 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:40,483 INFO [finetune.py:976] (0/7) Epoch 26, batch 5100, loss[loss=0.1949, simple_loss=0.2539, pruned_loss=0.06796, over 4860.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2388, pruned_loss=0.04799, over 954645.01 frames. ], batch size: 31, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:07:00,790 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1668, 1.9225, 1.7351, 1.7372, 1.8457, 1.8748, 1.9298, 2.5970], + device='cuda:0'), covar=tensor([0.3479, 0.4272, 0.3246, 0.3628, 0.3773, 0.2449, 0.3618, 0.1619], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0265, 0.0237, 0.0276, 0.0260, 0.0230, 0.0258, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:07:02,510 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:22,781 INFO [finetune.py:976] (0/7) Epoch 26, batch 5150, loss[loss=0.1847, simple_loss=0.262, pruned_loss=0.05371, over 4862.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2398, pruned_loss=0.0488, over 953834.85 frames. ], batch size: 34, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:07:27,034 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148348.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:36,968 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148363.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:41,444 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.549e+02 1.828e+02 2.233e+02 5.689e+02, threshold=3.657e+02, percent-clipped=4.0 +2023-03-27 08:07:55,902 INFO [finetune.py:976] (0/7) Epoch 26, batch 5200, loss[loss=0.1636, simple_loss=0.2252, pruned_loss=0.05103, over 4760.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.244, pruned_loss=0.04994, over 954507.11 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 08:08:15,076 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-03-27 08:08:21,961 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148429.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:08:21,990 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1443, 2.2289, 1.8337, 2.2048, 2.1270, 2.0885, 2.1106, 2.8296], + device='cuda:0'), covar=tensor([0.3756, 0.4305, 0.3455, 0.4045, 0.4090, 0.2530, 0.4353, 0.1719], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0265, 0.0237, 0.0276, 0.0260, 0.0229, 0.0258, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:08:29,695 INFO [finetune.py:976] (0/7) Epoch 26, batch 5250, loss[loss=0.1999, simple_loss=0.2725, pruned_loss=0.06367, over 4897.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2459, pruned_loss=0.05013, over 954881.01 frames. ], batch size: 35, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:08:44,236 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 08:08:48,645 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.514e+02 1.742e+02 2.193e+02 4.299e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 08:08:49,331 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:08:53,899 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3369, 2.4582, 2.2894, 1.7937, 2.2025, 2.6083, 2.5471, 1.9615], + device='cuda:0'), covar=tensor([0.0603, 0.0613, 0.0784, 0.0922, 0.1160, 0.0678, 0.0603, 0.1148], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0137, 0.0142, 0.0120, 0.0129, 0.0140, 0.0141, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:09:02,318 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148490.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:03,426 INFO [finetune.py:976] (0/7) Epoch 26, batch 5300, loss[loss=0.192, simple_loss=0.2758, pruned_loss=0.05413, over 4781.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2467, pruned_loss=0.05019, over 953423.03 frames. ], batch size: 51, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:09:07,097 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:20,037 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:21,826 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.0481, 4.7503, 4.6079, 2.4702, 4.8373, 3.5759, 0.9843, 3.4215], + device='cuda:0'), covar=tensor([0.2129, 0.2066, 0.1212, 0.2885, 0.0896, 0.0915, 0.4237, 0.1317], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0179, 0.0160, 0.0129, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 08:09:28,366 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148519.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:54,000 INFO [finetune.py:976] (0/7) Epoch 26, batch 5350, loss[loss=0.1893, simple_loss=0.2541, pruned_loss=0.06225, over 4903.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2453, pruned_loss=0.0496, over 953807.04 frames. ], batch size: 37, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:09:55,949 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7440, 3.7451, 3.6081, 1.5059, 3.8250, 2.8258, 0.7766, 2.6309], + device='cuda:0'), covar=tensor([0.2344, 0.2243, 0.1703, 0.3837, 0.1083, 0.1074, 0.4691, 0.1624], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0178, 0.0160, 0.0129, 0.0160, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 08:10:12,432 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148559.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:13,587 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148561.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:17,777 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148567.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:19,467 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.412e+02 1.653e+02 1.941e+02 3.220e+02, threshold=3.306e+02, percent-clipped=0.0 +2023-03-27 08:10:34,702 INFO [finetune.py:976] (0/7) Epoch 26, batch 5400, loss[loss=0.1615, simple_loss=0.2418, pruned_loss=0.04063, over 4779.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.243, pruned_loss=0.04897, over 955552.45 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:10:37,808 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1243, 2.0643, 1.6582, 2.1285, 2.0616, 1.7707, 2.4209, 2.1447], + device='cuda:0'), covar=tensor([0.1271, 0.1934, 0.2844, 0.2319, 0.2555, 0.1729, 0.2849, 0.1654], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0189, 0.0237, 0.0253, 0.0249, 0.0206, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:10:49,743 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148615.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:07,906 INFO [finetune.py:976] (0/7) Epoch 26, batch 5450, loss[loss=0.1762, simple_loss=0.2491, pruned_loss=0.05164, over 4910.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2402, pruned_loss=0.04836, over 955201.16 frames. ], batch size: 37, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:11:08,563 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:13,485 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148651.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:25,813 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.571e+02 1.855e+02 2.174e+02 4.125e+02, threshold=3.711e+02, percent-clipped=2.0 +2023-03-27 08:11:41,098 INFO [finetune.py:976] (0/7) Epoch 26, batch 5500, loss[loss=0.1319, simple_loss=0.2059, pruned_loss=0.02891, over 4771.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.237, pruned_loss=0.04735, over 954921.56 frames. ], batch size: 26, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:11:43,099 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2777, 1.7550, 2.2559, 2.2322, 2.0645, 2.0380, 2.1859, 2.1469], + device='cuda:0'), covar=tensor([0.3630, 0.3820, 0.3085, 0.3635, 0.4858, 0.3673, 0.4396, 0.2798], + device='cuda:0'), in_proj_covar=tensor([0.0264, 0.0246, 0.0266, 0.0294, 0.0294, 0.0270, 0.0300, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:11:53,951 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148712.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:56,391 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:24,626 INFO [finetune.py:976] (0/7) Epoch 26, batch 5550, loss[loss=0.1657, simple_loss=0.2571, pruned_loss=0.03719, over 4899.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2386, pruned_loss=0.04743, over 955962.04 frames. ], batch size: 43, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:12:42,332 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.664e+01 1.578e+02 1.917e+02 2.288e+02 4.413e+02, threshold=3.834e+02, percent-clipped=2.0 +2023-03-27 08:12:47,501 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:52,459 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148785.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:54,307 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.26 vs. limit=5.0 +2023-03-27 08:12:56,500 INFO [finetune.py:976] (0/7) Epoch 26, batch 5600, loss[loss=0.1736, simple_loss=0.2327, pruned_loss=0.05731, over 4776.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2425, pruned_loss=0.04851, over 956225.28 frames. ], batch size: 26, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:13:10,205 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 08:13:25,691 INFO [finetune.py:976] (0/7) Epoch 26, batch 5650, loss[loss=0.1734, simple_loss=0.2569, pruned_loss=0.04495, over 4858.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2443, pruned_loss=0.04846, over 955330.86 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:13:28,067 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0043, 1.3226, 0.8268, 1.8035, 2.3832, 1.7272, 1.6404, 1.8506], + device='cuda:0'), covar=tensor([0.1453, 0.2143, 0.2041, 0.1264, 0.1870, 0.1960, 0.1441, 0.2005], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 08:13:32,814 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148854.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:13:41,288 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5867, 1.5725, 1.8439, 1.2647, 1.5626, 1.8218, 1.5922, 1.9695], + device='cuda:0'), covar=tensor([0.1292, 0.2052, 0.1250, 0.1522, 0.0974, 0.1380, 0.2645, 0.0789], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0191, 0.0175, 0.0214, 0.0217, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:13:41,376 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-27 08:13:42,331 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.500e+02 1.770e+02 2.150e+02 4.859e+02, threshold=3.539e+02, percent-clipped=1.0 +2023-03-27 08:13:47,752 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6759, 1.5745, 2.0957, 2.9600, 2.1557, 2.3272, 1.1642, 2.5469], + device='cuda:0'), covar=tensor([0.1540, 0.1199, 0.1005, 0.0545, 0.0714, 0.1302, 0.1571, 0.0507], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0114, 0.0131, 0.0162, 0.0100, 0.0134, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 08:13:51,272 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6494, 1.6803, 1.0270, 2.5731, 3.0152, 2.1798, 2.3327, 2.4687], + device='cuda:0'), covar=tensor([0.1229, 0.1904, 0.1868, 0.0927, 0.1322, 0.1565, 0.1250, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 08:13:55,305 INFO [finetune.py:976] (0/7) Epoch 26, batch 5700, loss[loss=0.1508, simple_loss=0.2149, pruned_loss=0.04336, over 4260.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2404, pruned_loss=0.04798, over 933544.49 frames. ], batch size: 18, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:14:12,137 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-26.pt +2023-03-27 08:14:24,190 INFO [finetune.py:976] (0/7) Epoch 27, batch 0, loss[loss=0.1489, simple_loss=0.2416, pruned_loss=0.02812, over 4792.00 frames. ], tot_loss[loss=0.1489, simple_loss=0.2416, pruned_loss=0.02812, over 4792.00 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:14:24,191 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 08:14:30,265 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6237, 1.2145, 0.8974, 1.5724, 2.1279, 1.1317, 1.5342, 1.5200], + device='cuda:0'), covar=tensor([0.1587, 0.2150, 0.1873, 0.1240, 0.1982, 0.2024, 0.1421, 0.2192], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 08:14:30,745 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4356, 1.3734, 1.2551, 1.3974, 1.6379, 1.6271, 1.3621, 1.2262], + device='cuda:0'), covar=tensor([0.0406, 0.0330, 0.0677, 0.0345, 0.0269, 0.0368, 0.0398, 0.0437], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0105, 0.0146, 0.0110, 0.0100, 0.0114, 0.0103, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7685e-05, 8.0337e-05, 1.1349e-04, 8.4398e-05, 7.7821e-05, 8.4339e-05, + 7.6199e-05, 8.4954e-05], device='cuda:0') +2023-03-27 08:14:43,013 INFO [finetune.py:1010] (0/7) Epoch 27, validation: loss=0.1593, simple_loss=0.2269, pruned_loss=0.04586, over 2265189.00 frames. +2023-03-27 08:14:43,014 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 08:14:57,049 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148943.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:15:27,448 INFO [finetune.py:976] (0/7) Epoch 27, batch 50, loss[loss=0.1467, simple_loss=0.2156, pruned_loss=0.03892, over 4723.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2448, pruned_loss=0.05094, over 215645.62 frames. ], batch size: 54, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:15:28,074 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.253e+01 1.427e+02 1.731e+02 2.058e+02 3.661e+02, threshold=3.462e+02, percent-clipped=4.0 +2023-03-27 08:15:30,790 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-27 08:15:44,092 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148991.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:15:53,996 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149007.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:03,651 INFO [finetune.py:976] (0/7) Epoch 27, batch 100, loss[loss=0.1811, simple_loss=0.2495, pruned_loss=0.05632, over 4785.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2416, pruned_loss=0.04979, over 380600.88 frames. ], batch size: 51, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:16:36,428 INFO [finetune.py:976] (0/7) Epoch 27, batch 150, loss[loss=0.1621, simple_loss=0.2324, pruned_loss=0.0459, over 4764.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2354, pruned_loss=0.04801, over 508541.44 frames. ], batch size: 28, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:16:37,488 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.570e+01 1.451e+02 1.770e+02 2.054e+02 3.397e+02, threshold=3.539e+02, percent-clipped=0.0 +2023-03-27 08:16:39,139 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:47,480 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149085.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:54,789 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4463, 1.5904, 1.7073, 0.8943, 1.6990, 1.8567, 1.9019, 1.5522], + device='cuda:0'), covar=tensor([0.0963, 0.0721, 0.0602, 0.0621, 0.0613, 0.0807, 0.0452, 0.0759], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0130, 0.0124, 0.0131, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.9437e-05, 1.0712e-04, 9.2253e-05, 8.6978e-05, 9.1978e-05, 9.2468e-05, + 1.0158e-04, 1.0762e-04], device='cuda:0') +2023-03-27 08:16:57,365 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 08:16:58,390 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4908, 1.4157, 1.3564, 1.4023, 0.8645, 2.2978, 0.7776, 1.2708], + device='cuda:0'), covar=tensor([0.3547, 0.2760, 0.2318, 0.2670, 0.2079, 0.0397, 0.2976, 0.1418], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:17:09,497 INFO [finetune.py:976] (0/7) Epoch 27, batch 200, loss[loss=0.1765, simple_loss=0.2447, pruned_loss=0.05416, over 3993.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2344, pruned_loss=0.04763, over 607190.63 frames. ], batch size: 65, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:17:19,405 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149133.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:17:39,178 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149154.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:17:52,873 INFO [finetune.py:976] (0/7) Epoch 27, batch 250, loss[loss=0.1486, simple_loss=0.2287, pruned_loss=0.03427, over 4763.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2389, pruned_loss=0.04886, over 683753.51 frames. ], batch size: 28, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:17:53,480 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.547e+02 1.763e+02 2.073e+02 3.560e+02, threshold=3.526e+02, percent-clipped=1.0 +2023-03-27 08:18:14,332 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-27 08:18:14,778 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149202.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:18:25,585 INFO [finetune.py:976] (0/7) Epoch 27, batch 300, loss[loss=0.1104, simple_loss=0.1842, pruned_loss=0.01833, over 4820.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2425, pruned_loss=0.04963, over 745256.22 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:18:45,459 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3127, 1.7426, 1.4160, 1.5120, 1.9058, 1.9201, 1.6298, 1.6660], + device='cuda:0'), covar=tensor([0.0663, 0.0322, 0.0526, 0.0359, 0.0290, 0.0469, 0.0391, 0.0384], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0107, 0.0147, 0.0112, 0.0102, 0.0116, 0.0105, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8546e-05, 8.1570e-05, 1.1494e-04, 8.5492e-05, 7.8742e-05, 8.5252e-05, + 7.7678e-05, 8.5762e-05], device='cuda:0') +2023-03-27 08:18:46,654 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.1297, 1.2422, 1.4567, 1.2198, 1.3561, 2.4416, 1.2335, 1.3741], + device='cuda:0'), covar=tensor([0.1021, 0.1893, 0.1053, 0.0961, 0.1697, 0.0368, 0.1583, 0.1808], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:18:55,222 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.3187, 4.5574, 4.8600, 5.1486, 4.9965, 4.7160, 5.3879, 1.5446], + device='cuda:0'), covar=tensor([0.0680, 0.0828, 0.0840, 0.0841, 0.1204, 0.1651, 0.0548, 0.6083], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0248, 0.0284, 0.0296, 0.0336, 0.0288, 0.0305, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:18:58,793 INFO [finetune.py:976] (0/7) Epoch 27, batch 350, loss[loss=0.169, simple_loss=0.2479, pruned_loss=0.04506, over 4810.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2453, pruned_loss=0.05032, over 792609.56 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:18:59,397 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.601e+02 1.876e+02 2.140e+02 5.128e+02, threshold=3.753e+02, percent-clipped=1.0 +2023-03-27 08:19:14,647 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0790, 2.0152, 1.6806, 1.7842, 1.8748, 1.8530, 1.9207, 2.5785], + device='cuda:0'), covar=tensor([0.3476, 0.3608, 0.3067, 0.3634, 0.3695, 0.2205, 0.3333, 0.1633], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0235, 0.0275, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:19:24,861 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:19:32,719 INFO [finetune.py:976] (0/7) Epoch 27, batch 400, loss[loss=0.1636, simple_loss=0.2305, pruned_loss=0.04837, over 4866.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2467, pruned_loss=0.05072, over 829453.90 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:19:34,675 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4816, 1.5532, 1.6748, 0.9576, 1.7594, 1.8747, 1.9615, 1.5072], + device='cuda:0'), covar=tensor([0.0820, 0.0641, 0.0530, 0.0502, 0.0447, 0.0632, 0.0348, 0.0664], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0128, 0.0122, 0.0130, 0.0129, 0.0141, 0.0149], + device='cuda:0'), out_proj_covar=tensor([8.7996e-05, 1.0552e-04, 9.1176e-05, 8.5589e-05, 9.0718e-05, 9.1437e-05, + 1.0023e-04, 1.0660e-04], device='cuda:0') +2023-03-27 08:19:35,439 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-27 08:20:07,340 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149355.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:20:10,495 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1863, 2.0385, 1.8969, 2.1376, 1.9844, 2.0042, 1.9548, 2.5933], + device='cuda:0'), covar=tensor([0.3123, 0.4250, 0.3052, 0.3743, 0.4137, 0.2272, 0.3800, 0.1538], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0275, 0.0259, 0.0229, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:20:16,469 INFO [finetune.py:976] (0/7) Epoch 27, batch 450, loss[loss=0.1746, simple_loss=0.2403, pruned_loss=0.05444, over 4814.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2455, pruned_loss=0.05, over 858138.53 frames. ], batch size: 40, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:20:17,064 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.956e+01 1.496e+02 1.736e+02 2.126e+02 4.914e+02, threshold=3.471e+02, percent-clipped=1.0 +2023-03-27 08:20:17,775 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149372.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:20:56,963 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 08:21:04,694 INFO [finetune.py:976] (0/7) Epoch 27, batch 500, loss[loss=0.1618, simple_loss=0.2285, pruned_loss=0.04756, over 4861.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2427, pruned_loss=0.04914, over 877786.78 frames. ], batch size: 44, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:21:04,758 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:21:38,463 INFO [finetune.py:976] (0/7) Epoch 27, batch 550, loss[loss=0.1607, simple_loss=0.227, pruned_loss=0.04721, over 4808.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2399, pruned_loss=0.04856, over 894003.81 frames. ], batch size: 51, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:21:39,062 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.918e+01 1.466e+02 1.717e+02 2.125e+02 3.295e+02, threshold=3.435e+02, percent-clipped=0.0 +2023-03-27 08:22:02,069 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 08:22:08,651 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149514.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:22:12,133 INFO [finetune.py:976] (0/7) Epoch 27, batch 600, loss[loss=0.1716, simple_loss=0.2508, pruned_loss=0.04623, over 4932.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.24, pruned_loss=0.04825, over 909298.17 frames. ], batch size: 33, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:22:20,074 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5435, 1.4777, 2.1510, 3.2072, 2.1749, 2.3205, 1.4893, 2.7163], + device='cuda:0'), covar=tensor([0.1649, 0.1379, 0.1152, 0.0445, 0.0769, 0.1699, 0.1296, 0.0425], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 08:22:27,602 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0038, 1.8948, 1.6890, 2.1577, 2.5606, 2.1499, 1.9003, 1.6248], + device='cuda:0'), covar=tensor([0.2124, 0.1885, 0.1884, 0.1645, 0.1486, 0.1130, 0.2022, 0.1927], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0247, 0.0193, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:22:45,156 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149565.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:22:48,083 INFO [finetune.py:976] (0/7) Epoch 27, batch 650, loss[loss=0.2164, simple_loss=0.2877, pruned_loss=0.07253, over 4873.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.244, pruned_loss=0.04958, over 921317.11 frames. ], batch size: 34, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:22:53,183 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.592e+02 1.975e+02 2.434e+02 4.045e+02, threshold=3.949e+02, percent-clipped=4.0 +2023-03-27 08:22:55,795 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:23:20,274 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-03-27 08:23:29,819 INFO [finetune.py:976] (0/7) Epoch 27, batch 700, loss[loss=0.1441, simple_loss=0.2222, pruned_loss=0.03296, over 4812.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2433, pruned_loss=0.04877, over 927481.99 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:23:33,651 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149626.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:23:57,741 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7634, 1.2135, 1.8610, 1.8133, 1.6075, 1.5540, 1.7249, 1.7764], + device='cuda:0'), covar=tensor([0.3388, 0.3274, 0.2548, 0.2896, 0.4009, 0.3362, 0.3441, 0.2413], + device='cuda:0'), in_proj_covar=tensor([0.0266, 0.0247, 0.0268, 0.0296, 0.0296, 0.0271, 0.0301, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:24:03,075 INFO [finetune.py:976] (0/7) Epoch 27, batch 750, loss[loss=0.1538, simple_loss=0.2185, pruned_loss=0.04455, over 4810.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2456, pruned_loss=0.04986, over 933415.87 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:24:03,182 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3611, 1.4154, 1.6060, 1.5527, 1.5290, 2.9548, 1.3548, 1.4834], + device='cuda:0'), covar=tensor([0.1117, 0.1919, 0.1189, 0.0973, 0.1733, 0.0309, 0.1571, 0.1973], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:24:03,219 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7659, 1.6821, 1.5142, 1.8946, 2.0571, 1.8314, 1.4807, 1.4771], + device='cuda:0'), covar=tensor([0.2084, 0.1898, 0.1903, 0.1533, 0.1532, 0.1154, 0.2306, 0.1957], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0198, 0.0245, 0.0192, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:24:03,698 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.520e+02 1.783e+02 2.094e+02 3.998e+02, threshold=3.567e+02, percent-clipped=1.0 +2023-03-27 08:24:36,876 INFO [finetune.py:976] (0/7) Epoch 27, batch 800, loss[loss=0.2, simple_loss=0.2675, pruned_loss=0.06626, over 4788.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2454, pruned_loss=0.04954, over 937869.08 frames. ], batch size: 26, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:24:53,185 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149746.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 08:25:07,191 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4420, 1.3336, 1.3447, 1.3225, 0.8143, 2.2351, 0.7031, 1.1618], + device='cuda:0'), covar=tensor([0.3088, 0.2431, 0.2095, 0.2347, 0.1932, 0.0343, 0.2755, 0.1299], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0112, 0.0095, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:25:08,426 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1551, 2.0826, 2.2982, 1.5647, 2.1769, 2.2278, 2.1995, 1.7946], + device='cuda:0'), covar=tensor([0.0556, 0.0627, 0.0574, 0.0806, 0.0653, 0.0651, 0.0595, 0.1058], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0136, 0.0141, 0.0119, 0.0129, 0.0138, 0.0141, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:25:15,279 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2794, 1.8685, 2.3320, 2.2932, 2.0076, 2.0044, 2.2427, 2.1358], + device='cuda:0'), covar=tensor([0.3869, 0.4117, 0.3229, 0.3868, 0.4903, 0.4116, 0.4625, 0.2991], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0248, 0.0268, 0.0296, 0.0296, 0.0272, 0.0301, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:25:20,610 INFO [finetune.py:976] (0/7) Epoch 27, batch 850, loss[loss=0.1693, simple_loss=0.2386, pruned_loss=0.04999, over 4846.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2436, pruned_loss=0.04876, over 942419.49 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:25:21,210 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.421e+02 1.714e+02 1.950e+02 4.580e+02, threshold=3.429e+02, percent-clipped=2.0 +2023-03-27 08:25:28,022 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4838, 2.6346, 2.5102, 1.8270, 2.4026, 2.7702, 2.6112, 2.1387], + device='cuda:0'), covar=tensor([0.0636, 0.0615, 0.0725, 0.0894, 0.0806, 0.0645, 0.0664, 0.1106], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0136, 0.0141, 0.0119, 0.0128, 0.0138, 0.0141, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:25:56,135 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4153, 3.8923, 4.1956, 4.1481, 3.9544, 3.7246, 4.5085, 1.3734], + device='cuda:0'), covar=tensor([0.1108, 0.1529, 0.1590, 0.1441, 0.2132, 0.2418, 0.1050, 0.7556], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0249, 0.0285, 0.0298, 0.0337, 0.0289, 0.0307, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:25:56,747 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149807.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 08:26:09,033 INFO [finetune.py:976] (0/7) Epoch 27, batch 900, loss[loss=0.1886, simple_loss=0.2467, pruned_loss=0.06524, over 4218.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2404, pruned_loss=0.04766, over 944337.68 frames. ], batch size: 65, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:26:42,227 INFO [finetune.py:976] (0/7) Epoch 27, batch 950, loss[loss=0.1881, simple_loss=0.2646, pruned_loss=0.05583, over 4864.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2392, pruned_loss=0.04799, over 946520.95 frames. ], batch size: 44, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:26:42,288 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:26:42,815 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.619e+01 1.503e+02 1.866e+02 2.296e+02 3.689e+02, threshold=3.732e+02, percent-clipped=3.0 +2023-03-27 08:26:52,764 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-27 08:27:12,840 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-27 08:27:15,535 INFO [finetune.py:976] (0/7) Epoch 27, batch 1000, loss[loss=0.1846, simple_loss=0.2622, pruned_loss=0.05351, over 4819.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2409, pruned_loss=0.0483, over 950709.25 frames. ], batch size: 38, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:27:16,175 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149921.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 08:27:26,059 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 08:27:48,832 INFO [finetune.py:976] (0/7) Epoch 27, batch 1050, loss[loss=0.2549, simple_loss=0.3099, pruned_loss=0.09995, over 4826.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2447, pruned_loss=0.04932, over 953681.16 frames. ], batch size: 33, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:27:49,418 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.561e+02 1.767e+02 2.240e+02 3.870e+02, threshold=3.534e+02, percent-clipped=1.0 +2023-03-27 08:28:08,887 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 08:28:18,394 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-150000.pt +2023-03-27 08:28:29,304 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 08:28:33,249 INFO [finetune.py:976] (0/7) Epoch 27, batch 1100, loss[loss=0.1695, simple_loss=0.2462, pruned_loss=0.04634, over 4882.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2461, pruned_loss=0.04964, over 954132.65 frames. ], batch size: 43, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:28:44,736 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-27 08:28:51,574 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 08:29:06,474 INFO [finetune.py:976] (0/7) Epoch 27, batch 1150, loss[loss=0.1693, simple_loss=0.2468, pruned_loss=0.04587, over 4756.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.247, pruned_loss=0.05016, over 956393.00 frames. ], batch size: 28, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:29:07,079 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.913e+01 1.470e+02 1.766e+02 2.217e+02 3.439e+02, threshold=3.531e+02, percent-clipped=0.0 +2023-03-27 08:29:13,044 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150079.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:29:26,976 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150102.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 08:29:39,272 INFO [finetune.py:976] (0/7) Epoch 27, batch 1200, loss[loss=0.2218, simple_loss=0.2874, pruned_loss=0.07808, over 4853.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2462, pruned_loss=0.04997, over 956241.14 frames. ], batch size: 44, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:29:40,449 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0986, 1.2643, 1.3981, 1.3009, 1.3938, 2.4080, 1.1499, 1.3646], + device='cuda:0'), covar=tensor([0.1173, 0.2066, 0.1192, 0.1045, 0.1810, 0.0402, 0.1809, 0.2057], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0090, 0.0079, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:29:52,949 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150140.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:14,458 INFO [finetune.py:976] (0/7) Epoch 27, batch 1250, loss[loss=0.1629, simple_loss=0.2332, pruned_loss=0.04626, over 4876.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2437, pruned_loss=0.04909, over 958331.53 frames. ], batch size: 34, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:30:14,545 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150170.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:15,534 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.554e+02 1.886e+02 2.235e+02 6.588e+02, threshold=3.772e+02, percent-clipped=2.0 +2023-03-27 08:30:54,025 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1722, 2.0350, 1.6851, 1.9409, 1.8944, 1.9538, 1.9560, 2.6701], + device='cuda:0'), covar=tensor([0.3287, 0.3630, 0.3126, 0.3556, 0.3829, 0.2334, 0.3448, 0.1531], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0237, 0.0277, 0.0259, 0.0229, 0.0259, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:30:56,374 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:57,558 INFO [finetune.py:976] (0/7) Epoch 27, batch 1300, loss[loss=0.1738, simple_loss=0.2391, pruned_loss=0.05422, over 4675.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2405, pruned_loss=0.0477, over 958270.71 frames. ], batch size: 59, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:31:02,854 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:05,650 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8118, 1.6401, 2.2017, 3.2967, 2.3311, 2.5445, 1.3081, 2.8385], + device='cuda:0'), covar=tensor([0.1638, 0.1338, 0.1252, 0.0548, 0.0737, 0.1163, 0.1651, 0.0483], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0131, 0.0162, 0.0100, 0.0134, 0.0123, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 08:31:42,207 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:42,763 INFO [finetune.py:976] (0/7) Epoch 27, batch 1350, loss[loss=0.1426, simple_loss=0.2058, pruned_loss=0.03975, over 4807.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2408, pruned_loss=0.04777, over 957942.34 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:31:43,342 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.453e+02 1.768e+02 2.125e+02 3.830e+02, threshold=3.537e+02, percent-clipped=1.0 +2023-03-27 08:31:45,772 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150274.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:54,347 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3819, 1.5707, 1.2191, 1.4273, 1.8262, 1.7166, 1.5137, 1.4168], + device='cuda:0'), covar=tensor([0.0430, 0.0274, 0.0588, 0.0309, 0.0224, 0.0501, 0.0370, 0.0402], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0105, 0.0145, 0.0111, 0.0100, 0.0115, 0.0103, 0.0112], + device='cuda:0'), out_proj_covar=tensor([7.7592e-05, 8.0418e-05, 1.1339e-04, 8.4490e-05, 7.7771e-05, 8.4722e-05, + 7.6600e-05, 8.5367e-05], device='cuda:0') +2023-03-27 08:32:16,593 INFO [finetune.py:976] (0/7) Epoch 27, batch 1400, loss[loss=0.2048, simple_loss=0.2831, pruned_loss=0.06326, over 4758.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2426, pruned_loss=0.04792, over 957394.11 frames. ], batch size: 54, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:32:28,280 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:32:49,841 INFO [finetune.py:976] (0/7) Epoch 27, batch 1450, loss[loss=0.1578, simple_loss=0.2308, pruned_loss=0.04235, over 4782.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2451, pruned_loss=0.04898, over 953518.82 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:32:50,437 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.587e+02 1.925e+02 2.309e+02 4.827e+02, threshold=3.851e+02, percent-clipped=3.0 +2023-03-27 08:33:11,834 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150402.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 08:33:17,285 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9181, 1.4784, 1.9475, 1.9675, 1.7302, 1.6735, 1.9098, 1.8396], + device='cuda:0'), covar=tensor([0.4118, 0.4063, 0.3233, 0.3586, 0.4821, 0.3785, 0.4493, 0.2942], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0248, 0.0268, 0.0296, 0.0295, 0.0271, 0.0300, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:33:24,245 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7266, 1.0765, 1.7465, 1.7856, 1.5910, 1.5365, 1.6803, 1.6955], + device='cuda:0'), covar=tensor([0.3572, 0.3832, 0.3152, 0.3320, 0.4508, 0.3433, 0.3938, 0.2790], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0247, 0.0268, 0.0296, 0.0295, 0.0271, 0.0300, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:33:29,002 INFO [finetune.py:976] (0/7) Epoch 27, batch 1500, loss[loss=0.1355, simple_loss=0.2181, pruned_loss=0.02642, over 4004.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2475, pruned_loss=0.05046, over 951522.19 frames. ], batch size: 17, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:33:42,494 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150435.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:33:53,594 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150450.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 08:34:05,551 INFO [finetune.py:976] (0/7) Epoch 27, batch 1550, loss[loss=0.1805, simple_loss=0.2606, pruned_loss=0.05017, over 4805.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2478, pruned_loss=0.05045, over 953191.92 frames. ], batch size: 40, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:34:06,129 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.290e+01 1.580e+02 1.863e+02 2.206e+02 4.598e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-27 08:34:18,286 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 08:34:20,417 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5216, 1.4528, 1.9728, 1.9401, 1.5001, 3.3928, 1.3519, 1.5036], + device='cuda:0'), covar=tensor([0.1211, 0.2471, 0.1286, 0.1022, 0.1953, 0.0282, 0.1986, 0.2463], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:34:38,724 INFO [finetune.py:976] (0/7) Epoch 27, batch 1600, loss[loss=0.1796, simple_loss=0.2413, pruned_loss=0.059, over 4911.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2461, pruned_loss=0.05028, over 953230.87 frames. ], batch size: 43, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:34:46,294 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 08:34:50,074 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150537.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:34:59,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0674, 1.7357, 2.1863, 1.4906, 1.9962, 2.1719, 1.6667, 2.3416], + device='cuda:0'), covar=tensor([0.1273, 0.2107, 0.1329, 0.1922, 0.0950, 0.1359, 0.2719, 0.0855], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0193, 0.0190, 0.0174, 0.0213, 0.0217, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:35:08,683 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 08:35:09,314 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 08:35:11,517 INFO [finetune.py:976] (0/7) Epoch 27, batch 1650, loss[loss=0.119, simple_loss=0.1966, pruned_loss=0.02067, over 4813.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2432, pruned_loss=0.04944, over 953969.71 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:35:12,131 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.535e+02 1.741e+02 2.182e+02 5.670e+02, threshold=3.482e+02, percent-clipped=1.0 +2023-03-27 08:35:37,646 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150598.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:35:54,938 INFO [finetune.py:976] (0/7) Epoch 27, batch 1700, loss[loss=0.1587, simple_loss=0.2307, pruned_loss=0.04338, over 4824.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2412, pruned_loss=0.04916, over 953367.92 frames. ], batch size: 39, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:36:01,033 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150630.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:36:19,762 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-27 08:36:27,951 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9431, 1.8802, 1.7457, 2.1632, 2.3986, 2.1306, 1.8350, 1.6253], + device='cuda:0'), covar=tensor([0.2045, 0.1820, 0.1787, 0.1487, 0.1539, 0.1123, 0.2047, 0.1846], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0211, 0.0214, 0.0199, 0.0246, 0.0192, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:36:41,945 INFO [finetune.py:976] (0/7) Epoch 27, batch 1750, loss[loss=0.2015, simple_loss=0.2895, pruned_loss=0.05671, over 4798.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2434, pruned_loss=0.05009, over 955565.40 frames. ], batch size: 45, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:36:42,540 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.512e+01 1.530e+02 1.821e+02 2.198e+02 3.521e+02, threshold=3.642e+02, percent-clipped=1.0 +2023-03-27 08:36:51,148 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4105, 1.3706, 1.4924, 1.6337, 1.5559, 2.9899, 1.3220, 1.4421], + device='cuda:0'), covar=tensor([0.1037, 0.1855, 0.1111, 0.0934, 0.1617, 0.0258, 0.1552, 0.1942], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:36:54,005 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7894, 1.3451, 0.8925, 1.7118, 2.2237, 1.5600, 1.7256, 1.6980], + device='cuda:0'), covar=tensor([0.1495, 0.2166, 0.1854, 0.1129, 0.1814, 0.1876, 0.1384, 0.1983], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 08:37:05,442 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9889, 1.9521, 1.6316, 1.7469, 1.8486, 1.7906, 1.8959, 2.5491], + device='cuda:0'), covar=tensor([0.3869, 0.4073, 0.3244, 0.3861, 0.3921, 0.2560, 0.3624, 0.1647], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0276, 0.0259, 0.0229, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:37:15,430 INFO [finetune.py:976] (0/7) Epoch 27, batch 1800, loss[loss=0.1538, simple_loss=0.2124, pruned_loss=0.04757, over 4720.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2461, pruned_loss=0.05076, over 956005.96 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:37:24,764 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3984, 2.2189, 1.8343, 0.9068, 2.0512, 1.8043, 1.5809, 2.0978], + device='cuda:0'), covar=tensor([0.0972, 0.0864, 0.1666, 0.2245, 0.1476, 0.2494, 0.2590, 0.1028], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0192, 0.0202, 0.0183, 0.0210, 0.0212, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:37:27,759 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:37:33,297 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:37:57,192 INFO [finetune.py:976] (0/7) Epoch 27, batch 1850, loss[loss=0.1648, simple_loss=0.2451, pruned_loss=0.04226, over 4793.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.248, pruned_loss=0.05148, over 956088.61 frames. ], batch size: 51, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:37:57,786 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.537e+02 1.800e+02 2.248e+02 4.542e+02, threshold=3.600e+02, percent-clipped=6.0 +2023-03-27 08:38:05,039 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:38:11,138 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150793.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:38:20,972 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8992, 1.3304, 1.8923, 1.8738, 1.6753, 1.6427, 1.8185, 1.7962], + device='cuda:0'), covar=tensor([0.3989, 0.3758, 0.3240, 0.3526, 0.4568, 0.3761, 0.4060, 0.2973], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0247, 0.0267, 0.0295, 0.0295, 0.0271, 0.0301, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:38:23,471 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 08:38:30,231 INFO [finetune.py:976] (0/7) Epoch 27, batch 1900, loss[loss=0.1928, simple_loss=0.2619, pruned_loss=0.0618, over 4935.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2487, pruned_loss=0.05207, over 953561.01 frames. ], batch size: 42, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:39:14,080 INFO [finetune.py:976] (0/7) Epoch 27, batch 1950, loss[loss=0.1539, simple_loss=0.2272, pruned_loss=0.04026, over 4759.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2467, pruned_loss=0.05083, over 955682.87 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 8.0 +2023-03-27 08:39:14,655 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.460e+02 1.651e+02 1.933e+02 3.642e+02, threshold=3.302e+02, percent-clipped=1.0 +2023-03-27 08:39:28,774 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150893.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:39:30,070 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2989, 2.3258, 1.9033, 2.3179, 2.2139, 2.2241, 2.2256, 3.1084], + device='cuda:0'), covar=tensor([0.3543, 0.4681, 0.3335, 0.4042, 0.4329, 0.2361, 0.4123, 0.1508], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0236, 0.0275, 0.0259, 0.0229, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:39:30,640 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2642, 1.9369, 2.3625, 4.3006, 2.9643, 2.9277, 1.0057, 3.6927], + device='cuda:0'), covar=tensor([0.1685, 0.1380, 0.1488, 0.0450, 0.0743, 0.1214, 0.2005, 0.0317], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0163, 0.0101, 0.0136, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 08:39:33,120 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150900.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:39:35,409 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.24 vs. limit=5.0 +2023-03-27 08:39:47,875 INFO [finetune.py:976] (0/7) Epoch 27, batch 2000, loss[loss=0.1749, simple_loss=0.2315, pruned_loss=0.05913, over 4661.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2439, pruned_loss=0.05014, over 952007.73 frames. ], batch size: 23, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:39:54,535 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150930.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:12,127 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3041, 1.3410, 1.5229, 1.0518, 1.3842, 1.4364, 1.3118, 1.6388], + device='cuda:0'), covar=tensor([0.1158, 0.2166, 0.1261, 0.1529, 0.0877, 0.1150, 0.3085, 0.0849], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0193, 0.0190, 0.0174, 0.0212, 0.0217, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:40:15,356 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150961.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:19,466 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1598, 1.2499, 0.7756, 1.9011, 2.4591, 1.8353, 1.5574, 1.8682], + device='cuda:0'), covar=tensor([0.1270, 0.2002, 0.1828, 0.1039, 0.1772, 0.1845, 0.1359, 0.1790], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 08:40:21,604 INFO [finetune.py:976] (0/7) Epoch 27, batch 2050, loss[loss=0.1672, simple_loss=0.2346, pruned_loss=0.04994, over 4802.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2406, pruned_loss=0.04902, over 952472.36 frames. ], batch size: 51, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:40:22,190 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.304e+01 1.432e+02 1.658e+02 2.071e+02 3.830e+02, threshold=3.317e+02, percent-clipped=1.0 +2023-03-27 08:40:23,650 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-03-27 08:40:27,055 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:56,345 INFO [finetune.py:976] (0/7) Epoch 27, batch 2100, loss[loss=0.1336, simple_loss=0.2034, pruned_loss=0.03194, over 4896.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2401, pruned_loss=0.04874, over 953739.93 frames. ], batch size: 32, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:40:58,417 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 08:41:47,146 INFO [finetune.py:976] (0/7) Epoch 27, batch 2150, loss[loss=0.1621, simple_loss=0.2463, pruned_loss=0.03897, over 4748.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2428, pruned_loss=0.0491, over 954395.36 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:41:48,291 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.525e+02 1.813e+02 2.166e+02 3.448e+02, threshold=3.626e+02, percent-clipped=1.0 +2023-03-27 08:42:02,494 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151088.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 08:42:04,349 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 08:42:23,618 INFO [finetune.py:976] (0/7) Epoch 27, batch 2200, loss[loss=0.1882, simple_loss=0.2682, pruned_loss=0.05413, over 4194.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2452, pruned_loss=0.04959, over 954068.32 frames. ], batch size: 65, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:04,288 INFO [finetune.py:976] (0/7) Epoch 27, batch 2250, loss[loss=0.148, simple_loss=0.2297, pruned_loss=0.03312, over 4761.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2453, pruned_loss=0.04953, over 952308.41 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:04,889 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.120e+01 1.457e+02 1.754e+02 2.221e+02 3.820e+02, threshold=3.509e+02, percent-clipped=1.0 +2023-03-27 08:43:08,906 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8877, 1.7611, 2.3464, 2.0177, 2.0272, 4.4098, 1.8619, 1.9201], + device='cuda:0'), covar=tensor([0.0882, 0.1744, 0.1095, 0.0912, 0.1412, 0.0174, 0.1346, 0.1756], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:43:14,149 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6944, 3.6885, 3.4409, 2.0676, 3.8247, 2.9020, 1.3207, 2.7874], + device='cuda:0'), covar=tensor([0.2510, 0.2105, 0.1924, 0.3252, 0.1257, 0.1108, 0.4277, 0.1408], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0180, 0.0161, 0.0130, 0.0161, 0.0125, 0.0150, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 08:43:14,777 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151184.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:20,808 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151193.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:37,565 INFO [finetune.py:976] (0/7) Epoch 27, batch 2300, loss[loss=0.1576, simple_loss=0.2188, pruned_loss=0.04825, over 4767.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2449, pruned_loss=0.04925, over 951605.74 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:42,925 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6656, 2.4772, 2.0916, 1.0616, 2.2261, 1.9991, 1.9293, 2.3707], + device='cuda:0'), covar=tensor([0.0828, 0.0802, 0.1672, 0.2036, 0.1342, 0.2200, 0.2066, 0.0873], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0192, 0.0202, 0.0183, 0.0210, 0.0212, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:43:51,733 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:51,792 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2971, 1.2995, 1.5915, 1.0690, 1.3221, 1.4350, 1.3275, 1.5907], + device='cuda:0'), covar=tensor([0.1131, 0.1840, 0.1069, 0.1270, 0.0813, 0.1141, 0.2539, 0.0753], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0193, 0.0190, 0.0174, 0.0213, 0.0217, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:43:54,670 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151245.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:06,971 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151256.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:17,227 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4767, 1.4424, 1.5556, 0.8117, 1.5616, 1.5554, 1.5176, 1.4096], + device='cuda:0'), covar=tensor([0.0593, 0.0875, 0.0720, 0.0954, 0.0916, 0.0656, 0.0675, 0.1262], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0138, 0.0142, 0.0121, 0.0129, 0.0139, 0.0142, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:44:18,928 INFO [finetune.py:976] (0/7) Epoch 27, batch 2350, loss[loss=0.1378, simple_loss=0.2128, pruned_loss=0.03138, over 4896.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.243, pruned_loss=0.04845, over 952792.08 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:44:19,967 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.208e+01 1.503e+02 1.827e+02 2.189e+02 3.264e+02, threshold=3.653e+02, percent-clipped=0.0 +2023-03-27 08:44:21,300 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7615, 1.7651, 1.6757, 1.8761, 1.5202, 4.0139, 1.6063, 1.9561], + device='cuda:0'), covar=tensor([0.3094, 0.2316, 0.1893, 0.2090, 0.1430, 0.0195, 0.2362, 0.1131], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 08:44:28,755 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:32,923 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9659, 1.3526, 1.9880, 1.9774, 1.7472, 1.7159, 1.9339, 1.9044], + device='cuda:0'), covar=tensor([0.3851, 0.3811, 0.2988, 0.3536, 0.4217, 0.3793, 0.4028, 0.2801], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0249, 0.0270, 0.0298, 0.0297, 0.0274, 0.0303, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:44:39,407 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151298.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:52,608 INFO [finetune.py:976] (0/7) Epoch 27, batch 2400, loss[loss=0.1807, simple_loss=0.2487, pruned_loss=0.05633, over 4831.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2403, pruned_loss=0.04798, over 952642.35 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:44:53,924 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4156, 1.4189, 1.3063, 1.4975, 1.7050, 1.6783, 1.5192, 1.2947], + device='cuda:0'), covar=tensor([0.0410, 0.0301, 0.0619, 0.0290, 0.0227, 0.0374, 0.0301, 0.0391], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0112, 0.0102, 0.0116, 0.0104, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.8299e-05, 8.1359e-05, 1.1492e-04, 8.5488e-05, 7.8680e-05, 8.5264e-05, + 7.6967e-05, 8.6406e-05], device='cuda:0') +2023-03-27 08:45:09,245 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151343.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:19,403 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151359.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:26,012 INFO [finetune.py:976] (0/7) Epoch 27, batch 2450, loss[loss=0.1822, simple_loss=0.2495, pruned_loss=0.05741, over 4923.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2376, pruned_loss=0.04745, over 953033.81 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:45:26,601 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.415e+02 1.689e+02 1.968e+02 4.441e+02, threshold=3.378e+02, percent-clipped=1.0 +2023-03-27 08:45:38,464 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151388.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:58,928 INFO [finetune.py:976] (0/7) Epoch 27, batch 2500, loss[loss=0.2075, simple_loss=0.2676, pruned_loss=0.07367, over 4859.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2405, pruned_loss=0.04886, over 952778.31 frames. ], batch size: 31, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:46:02,017 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8825, 2.5997, 2.4674, 1.2677, 2.6713, 2.0869, 2.0878, 2.4233], + device='cuda:0'), covar=tensor([0.0956, 0.0743, 0.1455, 0.2097, 0.1473, 0.2116, 0.1969, 0.1099], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0182, 0.0209, 0.0211, 0.0225, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:46:12,815 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151436.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:46:48,695 INFO [finetune.py:976] (0/7) Epoch 27, batch 2550, loss[loss=0.1267, simple_loss=0.2069, pruned_loss=0.02329, over 4799.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.244, pruned_loss=0.04992, over 953210.01 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:46:49,279 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.209e+01 1.472e+02 1.881e+02 2.470e+02 3.912e+02, threshold=3.762e+02, percent-clipped=2.0 +2023-03-27 08:47:13,637 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-03-27 08:47:24,842 INFO [finetune.py:976] (0/7) Epoch 27, batch 2600, loss[loss=0.1896, simple_loss=0.2579, pruned_loss=0.06062, over 4865.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2462, pruned_loss=0.05084, over 954900.42 frames. ], batch size: 31, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:47:42,179 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151540.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:03,459 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151556.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:16,179 INFO [finetune.py:976] (0/7) Epoch 27, batch 2650, loss[loss=0.1584, simple_loss=0.243, pruned_loss=0.03694, over 4753.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.248, pruned_loss=0.05104, over 956218.23 frames. ], batch size: 27, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:48:16,787 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.649e+02 1.887e+02 2.270e+02 4.456e+02, threshold=3.774e+02, percent-clipped=3.0 +2023-03-27 08:48:39,915 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151604.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:49,946 INFO [finetune.py:976] (0/7) Epoch 27, batch 2700, loss[loss=0.1786, simple_loss=0.2479, pruned_loss=0.05465, over 4818.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2454, pruned_loss=0.05012, over 955477.00 frames. ], batch size: 39, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:49:01,305 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:49:02,009 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9108, 1.4441, 2.0380, 1.9864, 1.7924, 1.7010, 1.9091, 1.9191], + device='cuda:0'), covar=tensor([0.4063, 0.4091, 0.3092, 0.3460, 0.4446, 0.3674, 0.4304, 0.3005], + device='cuda:0'), in_proj_covar=tensor([0.0268, 0.0249, 0.0270, 0.0298, 0.0297, 0.0273, 0.0303, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:49:06,003 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5140, 1.5180, 2.2315, 1.7112, 1.7506, 3.8930, 1.4880, 1.7462], + device='cuda:0'), covar=tensor([0.1009, 0.1729, 0.1447, 0.0962, 0.1571, 0.0252, 0.1432, 0.1674], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0074, 0.0077, 0.0092, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 08:49:12,819 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:49:29,776 INFO [finetune.py:976] (0/7) Epoch 27, batch 2750, loss[loss=0.1598, simple_loss=0.2286, pruned_loss=0.04548, over 4772.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2422, pruned_loss=0.04922, over 956270.07 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:49:30,372 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.940e+01 1.418e+02 1.693e+02 2.178e+02 3.976e+02, threshold=3.385e+02, percent-clipped=1.0 +2023-03-27 08:49:44,693 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151688.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:50:06,331 INFO [finetune.py:976] (0/7) Epoch 27, batch 2800, loss[loss=0.1296, simple_loss=0.2035, pruned_loss=0.02787, over 4832.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2393, pruned_loss=0.04805, over 957112.56 frames. ], batch size: 40, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:50:24,992 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151749.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:50:39,488 INFO [finetune.py:976] (0/7) Epoch 27, batch 2850, loss[loss=0.1583, simple_loss=0.2386, pruned_loss=0.03903, over 4035.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2381, pruned_loss=0.04766, over 955247.90 frames. ], batch size: 65, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:50:40,096 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.485e+02 1.795e+02 2.169e+02 3.375e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-27 08:50:59,467 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151801.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:51:12,912 INFO [finetune.py:976] (0/7) Epoch 27, batch 2900, loss[loss=0.1449, simple_loss=0.218, pruned_loss=0.0359, over 4706.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2413, pruned_loss=0.04908, over 951887.99 frames. ], batch size: 23, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:51:25,549 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151840.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:51:32,893 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-27 08:51:53,916 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151862.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:52:04,134 INFO [finetune.py:976] (0/7) Epoch 27, batch 2950, loss[loss=0.1196, simple_loss=0.1986, pruned_loss=0.02031, over 4795.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2448, pruned_loss=0.04938, over 952565.35 frames. ], batch size: 29, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:52:04,749 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.531e+02 1.876e+02 2.281e+02 4.815e+02, threshold=3.752e+02, percent-clipped=2.0 +2023-03-27 08:52:12,022 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.3085, 3.7182, 3.9545, 4.1369, 4.0702, 3.8214, 4.3772, 1.4742], + device='cuda:0'), covar=tensor([0.0695, 0.0805, 0.0777, 0.0795, 0.1131, 0.1627, 0.0680, 0.5657], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0249, 0.0282, 0.0297, 0.0335, 0.0288, 0.0306, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:52:15,638 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151888.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:52:37,406 INFO [finetune.py:976] (0/7) Epoch 27, batch 3000, loss[loss=0.1426, simple_loss=0.2023, pruned_loss=0.04141, over 4717.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2452, pruned_loss=0.04912, over 952322.26 frames. ], batch size: 23, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:52:37,408 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 08:52:50,759 INFO [finetune.py:1010] (0/7) Epoch 27, validation: loss=0.1572, simple_loss=0.2248, pruned_loss=0.04486, over 2265189.00 frames. +2023-03-27 08:52:50,760 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 08:53:01,831 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:14,186 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151954.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:32,162 INFO [finetune.py:976] (0/7) Epoch 27, batch 3050, loss[loss=0.193, simple_loss=0.2699, pruned_loss=0.05808, over 4821.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2453, pruned_loss=0.04904, over 950451.77 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:53:32,747 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.546e+02 1.837e+02 2.199e+02 4.500e+02, threshold=3.674e+02, percent-clipped=2.0 +2023-03-27 08:53:44,356 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151986.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:50,363 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6827, 1.6331, 1.3914, 1.7777, 2.0287, 1.8039, 1.4349, 1.4170], + device='cuda:0'), covar=tensor([0.2177, 0.1923, 0.2005, 0.1650, 0.1608, 0.1190, 0.2414, 0.1916], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0212, 0.0216, 0.0201, 0.0248, 0.0192, 0.0220, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:53:53,470 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-152000.pt +2023-03-27 08:53:55,882 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:54:06,689 INFO [finetune.py:976] (0/7) Epoch 27, batch 3100, loss[loss=0.1782, simple_loss=0.2386, pruned_loss=0.05893, over 4871.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2436, pruned_loss=0.04876, over 952442.70 frames. ], batch size: 34, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:54:18,470 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-27 08:54:18,918 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2314, 1.2605, 1.3332, 0.7863, 1.3000, 1.4230, 1.5757, 1.2653], + device='cuda:0'), covar=tensor([0.0855, 0.0641, 0.0581, 0.0407, 0.0521, 0.0625, 0.0355, 0.0667], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0147, 0.0129, 0.0123, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.8814e-05, 1.0554e-04, 9.1981e-05, 8.6245e-05, 9.1774e-05, 9.1957e-05, + 1.0062e-04, 1.0737e-04], device='cuda:0') +2023-03-27 08:54:23,683 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152044.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:54:38,251 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-27 08:54:40,562 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4408, 1.4843, 1.9855, 1.7073, 1.5768, 3.4395, 1.3510, 1.6018], + device='cuda:0'), covar=tensor([0.1096, 0.1832, 0.1303, 0.0989, 0.1617, 0.0230, 0.1542, 0.1832], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0074, 0.0077, 0.0092, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 08:54:41,692 INFO [finetune.py:976] (0/7) Epoch 27, batch 3150, loss[loss=0.1393, simple_loss=0.2197, pruned_loss=0.02944, over 4765.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2425, pruned_loss=0.04936, over 954729.37 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:54:42,282 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.945e+01 1.491e+02 1.827e+02 2.202e+02 3.039e+02, threshold=3.654e+02, percent-clipped=0.0 +2023-03-27 08:54:50,862 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-03-27 08:55:15,177 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6552, 3.3861, 3.3214, 1.5297, 3.5384, 2.7056, 0.9819, 2.4075], + device='cuda:0'), covar=tensor([0.2073, 0.2355, 0.1683, 0.3422, 0.1135, 0.1068, 0.4169, 0.1731], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0179, 0.0159, 0.0129, 0.0160, 0.0124, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 08:55:21,810 INFO [finetune.py:976] (0/7) Epoch 27, batch 3200, loss[loss=0.1729, simple_loss=0.2333, pruned_loss=0.05622, over 4911.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2398, pruned_loss=0.04845, over 954306.95 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:55:46,784 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152157.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:55:54,656 INFO [finetune.py:976] (0/7) Epoch 27, batch 3250, loss[loss=0.1497, simple_loss=0.2204, pruned_loss=0.03951, over 4869.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.239, pruned_loss=0.04818, over 954158.24 frames. ], batch size: 34, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:55:55,264 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.795e+01 1.453e+02 1.756e+02 2.073e+02 3.538e+02, threshold=3.512e+02, percent-clipped=0.0 +2023-03-27 08:55:56,048 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-27 08:56:10,657 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3739, 2.2343, 2.0009, 2.3828, 2.3647, 2.1645, 2.6407, 2.4287], + device='cuda:0'), covar=tensor([0.1349, 0.1937, 0.2825, 0.2093, 0.2387, 0.1708, 0.2345, 0.1728], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0252, 0.0248, 0.0206, 0.0214, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:56:32,279 INFO [finetune.py:976] (0/7) Epoch 27, batch 3300, loss[loss=0.1634, simple_loss=0.2271, pruned_loss=0.04984, over 4664.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2419, pruned_loss=0.04855, over 955457.17 frames. ], batch size: 23, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:56:35,416 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:57:13,841 INFO [finetune.py:976] (0/7) Epoch 27, batch 3350, loss[loss=0.2113, simple_loss=0.2771, pruned_loss=0.07274, over 4893.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2438, pruned_loss=0.04919, over 952529.31 frames. ], batch size: 43, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:57:14,395 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.581e+01 1.606e+02 1.884e+02 2.337e+02 3.345e+02, threshold=3.768e+02, percent-clipped=0.0 +2023-03-27 08:57:20,866 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:57:33,567 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:57:42,730 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.33 vs. limit=5.0 +2023-03-27 08:57:49,803 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-27 08:58:01,023 INFO [finetune.py:976] (0/7) Epoch 27, batch 3400, loss[loss=0.1912, simple_loss=0.2778, pruned_loss=0.05232, over 4242.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2449, pruned_loss=0.04959, over 953766.04 frames. ], batch size: 65, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:58:09,662 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152334.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:58:16,178 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152344.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:58:23,113 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8428, 1.9518, 1.7136, 1.7661, 2.3159, 2.4066, 2.0908, 1.9932], + device='cuda:0'), covar=tensor([0.0375, 0.0371, 0.0625, 0.0341, 0.0281, 0.0516, 0.0332, 0.0380], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0107, 0.0149, 0.0112, 0.0102, 0.0116, 0.0104, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.8814e-05, 8.1887e-05, 1.1574e-04, 8.5855e-05, 7.9044e-05, 8.5787e-05, + 7.7166e-05, 8.6914e-05], device='cuda:0') +2023-03-27 08:58:36,170 INFO [finetune.py:976] (0/7) Epoch 27, batch 3450, loss[loss=0.1576, simple_loss=0.2226, pruned_loss=0.04625, over 4736.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2436, pruned_loss=0.0486, over 955388.93 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:58:36,743 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.467e+02 1.787e+02 2.252e+02 4.149e+02, threshold=3.573e+02, percent-clipped=3.0 +2023-03-27 08:58:58,950 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152392.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:18,830 INFO [finetune.py:976] (0/7) Epoch 27, batch 3500, loss[loss=0.1906, simple_loss=0.2641, pruned_loss=0.05858, over 4769.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2416, pruned_loss=0.04827, over 956785.62 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:59:34,569 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:43,765 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152457.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:45,651 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1073, 1.9956, 1.7486, 1.9610, 1.8548, 1.9050, 1.9672, 2.5895], + device='cuda:0'), covar=tensor([0.3713, 0.4009, 0.3269, 0.3769, 0.4277, 0.2414, 0.3658, 0.1709], + device='cuda:0'), in_proj_covar=tensor([0.0292, 0.0265, 0.0239, 0.0278, 0.0262, 0.0231, 0.0260, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:59:48,865 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 08:59:52,114 INFO [finetune.py:976] (0/7) Epoch 27, batch 3550, loss[loss=0.1762, simple_loss=0.2433, pruned_loss=0.05454, over 4879.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2394, pruned_loss=0.04766, over 957667.98 frames. ], batch size: 31, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:59:52,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9927, 1.8014, 2.2270, 1.5081, 2.0398, 2.2362, 1.6532, 2.3866], + device='cuda:0'), covar=tensor([0.1141, 0.2003, 0.1368, 0.1789, 0.0933, 0.1274, 0.2900, 0.0823], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0208, 0.0194, 0.0190, 0.0175, 0.0214, 0.0218, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 08:59:52,706 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.528e+01 1.381e+02 1.664e+02 2.040e+02 3.997e+02, threshold=3.328e+02, percent-clipped=1.0 +2023-03-27 08:59:56,454 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2457, 3.6852, 3.8669, 4.0807, 4.0096, 3.7094, 4.3305, 1.4520], + device='cuda:0'), covar=tensor([0.0767, 0.0791, 0.0804, 0.0931, 0.1136, 0.1434, 0.0625, 0.5441], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0244, 0.0279, 0.0292, 0.0329, 0.0284, 0.0301, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:00:03,457 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-03-27 09:00:21,708 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.77 vs. limit=5.0 +2023-03-27 09:00:25,119 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:00:26,309 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:00:36,269 INFO [finetune.py:976] (0/7) Epoch 27, batch 3600, loss[loss=0.2103, simple_loss=0.2722, pruned_loss=0.07426, over 4212.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2381, pruned_loss=0.04805, over 954857.48 frames. ], batch size: 65, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:01:10,219 INFO [finetune.py:976] (0/7) Epoch 27, batch 3650, loss[loss=0.1608, simple_loss=0.2425, pruned_loss=0.03956, over 4857.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2395, pruned_loss=0.04832, over 955289.41 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:01:10,827 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.982e+01 1.594e+02 1.907e+02 2.265e+02 4.160e+02, threshold=3.814e+02, percent-clipped=3.0 +2023-03-27 09:01:17,559 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152581.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:18,188 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152582.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:22,466 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:44,756 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7495, 1.5472, 1.5281, 0.7865, 1.6617, 1.8157, 1.7736, 1.4192], + device='cuda:0'), covar=tensor([0.0896, 0.0593, 0.0518, 0.0576, 0.0492, 0.0560, 0.0331, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0147, 0.0130, 0.0123, 0.0132, 0.0131, 0.0142, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.8968e-05, 1.0576e-04, 9.2812e-05, 8.6143e-05, 9.2234e-05, 9.2483e-05, + 1.0092e-04, 1.0778e-04], device='cuda:0') +2023-03-27 09:01:46,436 INFO [finetune.py:976] (0/7) Epoch 27, batch 3700, loss[loss=0.1617, simple_loss=0.2422, pruned_loss=0.04062, over 4925.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2425, pruned_loss=0.04903, over 954633.56 frames. ], batch size: 29, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:01:50,262 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8388, 1.7532, 1.5437, 1.8690, 2.2270, 1.9249, 1.7420, 1.5302], + device='cuda:0'), covar=tensor([0.1862, 0.1714, 0.1672, 0.1462, 0.1510, 0.1062, 0.2171, 0.1661], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0212, 0.0216, 0.0201, 0.0247, 0.0192, 0.0220, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:01:52,533 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:01,199 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:05,482 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152650.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:22,199 INFO [finetune.py:976] (0/7) Epoch 27, batch 3750, loss[loss=0.1888, simple_loss=0.2561, pruned_loss=0.06073, over 4059.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2444, pruned_loss=0.04929, over 954389.91 frames. ], batch size: 65, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:02:22,799 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.491e+02 1.751e+02 2.166e+02 4.226e+02, threshold=3.502e+02, percent-clipped=3.0 +2023-03-27 09:02:33,496 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5495, 1.4924, 1.4116, 1.5240, 0.9839, 2.9391, 1.0413, 1.4010], + device='cuda:0'), covar=tensor([0.3340, 0.2449, 0.2177, 0.2462, 0.1786, 0.0269, 0.2717, 0.1288], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 09:03:12,605 INFO [finetune.py:976] (0/7) Epoch 27, batch 3800, loss[loss=0.177, simple_loss=0.2507, pruned_loss=0.0516, over 4890.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2454, pruned_loss=0.04957, over 954326.41 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:03:16,897 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:03:26,558 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0526, 1.9968, 1.7628, 1.9117, 1.8564, 1.8901, 1.9407, 2.5679], + device='cuda:0'), covar=tensor([0.3735, 0.4074, 0.3219, 0.3978, 0.4094, 0.2410, 0.3493, 0.1811], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0262, 0.0236, 0.0275, 0.0260, 0.0229, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:03:28,536 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-27 09:03:45,579 INFO [finetune.py:976] (0/7) Epoch 27, batch 3850, loss[loss=0.1547, simple_loss=0.2337, pruned_loss=0.03791, over 4270.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.244, pruned_loss=0.04851, over 955613.98 frames. ], batch size: 66, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:03:46,653 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.206e+01 1.335e+02 1.631e+02 2.144e+02 3.589e+02, threshold=3.262e+02, percent-clipped=1.0 +2023-03-27 09:03:52,609 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8302, 1.6976, 1.5031, 1.4781, 1.8564, 1.6170, 1.7925, 1.8142], + device='cuda:0'), covar=tensor([0.1430, 0.1952, 0.2935, 0.2429, 0.2617, 0.1718, 0.2945, 0.1727], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0253, 0.0249, 0.0207, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:03:59,992 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152787.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:04:12,778 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152801.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:04:18,093 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1572, 2.0303, 1.7026, 2.1127, 2.5110, 2.1914, 2.0465, 1.6381], + device='cuda:0'), covar=tensor([0.2057, 0.1846, 0.1885, 0.1586, 0.1765, 0.1124, 0.2047, 0.1815], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0210, 0.0214, 0.0199, 0.0245, 0.0191, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:04:28,201 INFO [finetune.py:976] (0/7) Epoch 27, batch 3900, loss[loss=0.1624, simple_loss=0.2245, pruned_loss=0.05016, over 4915.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2418, pruned_loss=0.04849, over 954416.53 frames. ], batch size: 46, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:05:01,433 INFO [finetune.py:976] (0/7) Epoch 27, batch 3950, loss[loss=0.1506, simple_loss=0.2148, pruned_loss=0.04321, over 4926.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2394, pruned_loss=0.04802, over 955814.80 frames. ], batch size: 33, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:05:02,041 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.440e+02 1.686e+02 2.039e+02 3.105e+02, threshold=3.372e+02, percent-clipped=0.0 +2023-03-27 09:05:06,294 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.9150, 3.4742, 3.6032, 3.7881, 3.6695, 3.4035, 3.9653, 1.3194], + device='cuda:0'), covar=tensor([0.0816, 0.0832, 0.0911, 0.0910, 0.1208, 0.1758, 0.0841, 0.5283], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0246, 0.0281, 0.0293, 0.0331, 0.0285, 0.0302, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:05:09,729 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:16,462 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4574, 1.5864, 1.6781, 0.9815, 1.7332, 2.0073, 1.9833, 1.5070], + device='cuda:0'), covar=tensor([0.0854, 0.0644, 0.0554, 0.0542, 0.0535, 0.0602, 0.0321, 0.0780], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0131, 0.0124, 0.0133, 0.0132, 0.0144, 0.0152], + device='cuda:0'), out_proj_covar=tensor([8.9919e-05, 1.0687e-04, 9.3484e-05, 8.6950e-05, 9.3020e-05, 9.3582e-05, + 1.0209e-04, 1.0868e-04], device='cuda:0') +2023-03-27 09:05:43,211 INFO [finetune.py:976] (0/7) Epoch 27, batch 4000, loss[loss=0.1144, simple_loss=0.1864, pruned_loss=0.02116, over 4676.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2391, pruned_loss=0.0482, over 955550.42 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:05:49,732 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:49,759 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:56,139 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:00,330 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:16,490 INFO [finetune.py:976] (0/7) Epoch 27, batch 4050, loss[loss=0.1915, simple_loss=0.2851, pruned_loss=0.04897, over 4871.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2414, pruned_loss=0.04918, over 954665.87 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:06:17,094 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.469e+02 1.767e+02 2.180e+02 3.425e+02, threshold=3.534e+02, percent-clipped=1.0 +2023-03-27 09:06:20,836 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152977.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:49,253 INFO [finetune.py:976] (0/7) Epoch 27, batch 4100, loss[loss=0.1496, simple_loss=0.2195, pruned_loss=0.03986, over 4726.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.244, pruned_loss=0.04978, over 954352.14 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:06:52,866 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3094, 2.1949, 1.7776, 2.2549, 2.1865, 1.9632, 2.5430, 2.3411], + device='cuda:0'), covar=tensor([0.1257, 0.2099, 0.2914, 0.2487, 0.2529, 0.1650, 0.2931, 0.1581], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0253, 0.0249, 0.0206, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:07:06,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2871, 2.9145, 3.0497, 3.2253, 3.0768, 2.8683, 3.3297, 0.9417], + device='cuda:0'), covar=tensor([0.1013, 0.1034, 0.1086, 0.1068, 0.1449, 0.1980, 0.1115, 0.5687], + device='cuda:0'), in_proj_covar=tensor([0.0356, 0.0250, 0.0286, 0.0297, 0.0336, 0.0290, 0.0307, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:07:08,549 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4661, 2.3849, 2.0243, 0.9732, 2.1013, 1.8587, 1.8043, 2.1814], + device='cuda:0'), covar=tensor([0.0827, 0.0810, 0.1571, 0.2084, 0.1505, 0.2485, 0.2182, 0.0974], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0210, 0.0211, 0.0225, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:07:22,843 INFO [finetune.py:976] (0/7) Epoch 27, batch 4150, loss[loss=0.1987, simple_loss=0.2718, pruned_loss=0.06279, over 4904.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2449, pruned_loss=0.04976, over 954656.24 frames. ], batch size: 36, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:07:23,441 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.644e+02 1.926e+02 2.373e+02 3.999e+02, threshold=3.851e+02, percent-clipped=3.0 +2023-03-27 09:07:31,222 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153082.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:07:51,129 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153101.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:07,054 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:13,347 INFO [finetune.py:976] (0/7) Epoch 27, batch 4200, loss[loss=0.1413, simple_loss=0.2216, pruned_loss=0.03048, over 4765.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2449, pruned_loss=0.04938, over 956153.74 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:08:36,713 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153149.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:39,705 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1356, 1.4337, 0.5030, 2.0067, 2.4855, 1.7558, 1.8326, 1.9730], + device='cuda:0'), covar=tensor([0.1403, 0.2093, 0.2187, 0.1109, 0.1784, 0.1830, 0.1353, 0.1941], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 09:08:49,826 INFO [finetune.py:976] (0/7) Epoch 27, batch 4250, loss[loss=0.1793, simple_loss=0.2495, pruned_loss=0.05456, over 4898.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.242, pruned_loss=0.04824, over 957454.06 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:08:50,415 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.682e+01 1.570e+02 1.909e+02 2.227e+02 3.978e+02, threshold=3.818e+02, percent-clipped=1.0 +2023-03-27 09:08:54,811 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:09:04,059 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6766, 1.5332, 2.1526, 3.2819, 2.2423, 2.3612, 0.9239, 2.8123], + device='cuda:0'), covar=tensor([0.1590, 0.1352, 0.1240, 0.0519, 0.0751, 0.1463, 0.1823, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0163, 0.0101, 0.0135, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:09:33,254 INFO [finetune.py:976] (0/7) Epoch 27, batch 4300, loss[loss=0.1467, simple_loss=0.2219, pruned_loss=0.03574, over 4815.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2398, pruned_loss=0.04807, over 955897.48 frames. ], batch size: 51, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:09:44,711 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:09:49,326 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153245.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:06,729 INFO [finetune.py:976] (0/7) Epoch 27, batch 4350, loss[loss=0.1067, simple_loss=0.1791, pruned_loss=0.01716, over 4785.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2363, pruned_loss=0.04673, over 955034.69 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:10:07,328 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.429e+01 1.453e+02 1.753e+02 2.112e+02 4.699e+02, threshold=3.507e+02, percent-clipped=1.0 +2023-03-27 09:10:11,676 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153278.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:16,463 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:21,194 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153293.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:26,515 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2018, 2.2883, 1.8461, 1.9316, 2.6907, 2.8440, 2.1939, 2.1693], + device='cuda:0'), covar=tensor([0.0327, 0.0332, 0.0573, 0.0339, 0.0216, 0.0386, 0.0345, 0.0394], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0112, 0.0101, 0.0116, 0.0104, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8218e-05, 8.1144e-05, 1.1514e-04, 8.5624e-05, 7.8567e-05, 8.5320e-05, + 7.7063e-05, 8.6013e-05], device='cuda:0') +2023-03-27 09:10:41,647 INFO [finetune.py:976] (0/7) Epoch 27, batch 4400, loss[loss=0.1485, simple_loss=0.2175, pruned_loss=0.0398, over 4752.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2373, pruned_loss=0.04695, over 957212.24 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:11:01,195 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153339.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:11:04,367 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 09:11:20,226 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-03-27 09:11:23,046 INFO [finetune.py:976] (0/7) Epoch 27, batch 4450, loss[loss=0.142, simple_loss=0.2092, pruned_loss=0.03742, over 4038.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2393, pruned_loss=0.04718, over 957490.11 frames. ], batch size: 17, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:11:23,634 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.494e+02 1.793e+02 2.132e+02 3.020e+02, threshold=3.586e+02, percent-clipped=0.0 +2023-03-27 09:11:30,936 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153382.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:11:36,522 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 09:11:52,109 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9555, 1.7454, 1.9618, 1.2026, 1.8724, 1.9571, 1.9505, 1.5758], + device='cuda:0'), covar=tensor([0.0561, 0.0788, 0.0652, 0.0946, 0.0832, 0.0685, 0.0654, 0.1224], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0138, 0.0142, 0.0121, 0.0130, 0.0140, 0.0141, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:11:56,790 INFO [finetune.py:976] (0/7) Epoch 27, batch 4500, loss[loss=0.1702, simple_loss=0.2505, pruned_loss=0.04493, over 4898.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.241, pruned_loss=0.04776, over 953697.60 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:11:57,450 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:00,490 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 09:12:03,376 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153430.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:29,935 INFO [finetune.py:976] (0/7) Epoch 27, batch 4550, loss[loss=0.1812, simple_loss=0.2608, pruned_loss=0.05076, over 4904.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2423, pruned_loss=0.04764, over 955238.77 frames. ], batch size: 37, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:12:30,508 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.518e+01 1.590e+02 1.867e+02 2.229e+02 3.919e+02, threshold=3.734e+02, percent-clipped=1.0 +2023-03-27 09:12:30,649 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7522, 1.8886, 1.5847, 2.1476, 2.3279, 1.9711, 1.6937, 1.4346], + device='cuda:0'), covar=tensor([0.2360, 0.1922, 0.2113, 0.1554, 0.1838, 0.1250, 0.2487, 0.2150], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0212, 0.0216, 0.0201, 0.0246, 0.0192, 0.0219, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:12:31,756 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153473.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:37,665 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153482.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:13:14,340 INFO [finetune.py:976] (0/7) Epoch 27, batch 4600, loss[loss=0.1935, simple_loss=0.2552, pruned_loss=0.06591, over 4718.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2423, pruned_loss=0.04755, over 954791.24 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:13:56,975 INFO [finetune.py:976] (0/7) Epoch 27, batch 4650, loss[loss=0.1131, simple_loss=0.1909, pruned_loss=0.01764, over 4804.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2402, pruned_loss=0.04724, over 955120.20 frames. ], batch size: 25, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:13:57,586 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.461e+02 1.737e+02 2.165e+02 3.643e+02, threshold=3.475e+02, percent-clipped=0.0 +2023-03-27 09:14:04,935 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-03-27 09:14:31,815 INFO [finetune.py:976] (0/7) Epoch 27, batch 4700, loss[loss=0.1604, simple_loss=0.222, pruned_loss=0.04943, over 4817.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.238, pruned_loss=0.04688, over 954729.97 frames. ], batch size: 25, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:14:34,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2375, 2.1200, 1.8579, 2.1007, 2.0002, 2.0177, 1.9950, 2.8140], + device='cuda:0'), covar=tensor([0.3768, 0.4013, 0.3139, 0.3752, 0.3970, 0.2398, 0.3696, 0.1533], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0264, 0.0237, 0.0276, 0.0261, 0.0230, 0.0259, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:14:44,692 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153634.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:15:12,023 INFO [finetune.py:976] (0/7) Epoch 27, batch 4750, loss[loss=0.1738, simple_loss=0.2416, pruned_loss=0.05295, over 4766.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2366, pruned_loss=0.04677, over 953714.39 frames. ], batch size: 27, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:15:12,097 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3672, 1.2633, 1.8029, 2.6970, 1.7872, 2.2559, 1.0196, 2.4167], + device='cuda:0'), covar=tensor([0.1855, 0.1809, 0.1406, 0.1027, 0.1005, 0.1755, 0.1911, 0.0659], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0163, 0.0100, 0.0134, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:15:12,112 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3087, 1.3964, 1.5991, 1.5272, 1.5044, 2.8039, 1.2174, 1.4262], + device='cuda:0'), covar=tensor([0.0997, 0.1697, 0.1236, 0.0937, 0.1525, 0.0328, 0.1526, 0.1678], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 09:15:13,088 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.473e+02 1.795e+02 2.173e+02 4.465e+02, threshold=3.590e+02, percent-clipped=3.0 +2023-03-27 09:15:18,668 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-03-27 09:15:45,873 INFO [finetune.py:976] (0/7) Epoch 27, batch 4800, loss[loss=0.1577, simple_loss=0.2225, pruned_loss=0.04643, over 4413.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2403, pruned_loss=0.04811, over 953967.08 frames. ], batch size: 18, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:15:48,405 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.13 vs. limit=5.0 +2023-03-27 09:16:28,415 INFO [finetune.py:976] (0/7) Epoch 27, batch 4850, loss[loss=0.2169, simple_loss=0.2712, pruned_loss=0.08137, over 4823.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2428, pruned_loss=0.04878, over 955089.94 frames. ], batch size: 40, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:16:28,977 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.544e+02 1.777e+02 2.223e+02 4.381e+02, threshold=3.554e+02, percent-clipped=2.0 +2023-03-27 09:16:30,737 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:16:33,635 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:16:36,624 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5800, 3.4197, 3.2322, 1.4464, 3.5561, 2.6448, 0.7241, 2.2829], + device='cuda:0'), covar=tensor([0.2188, 0.2242, 0.1720, 0.3737, 0.1303, 0.1158, 0.4446, 0.1742], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0178, 0.0159, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 09:17:00,916 INFO [finetune.py:976] (0/7) Epoch 27, batch 4900, loss[loss=0.2012, simple_loss=0.2722, pruned_loss=0.06507, over 4892.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2442, pruned_loss=0.04917, over 954955.31 frames. ], batch size: 35, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:17:01,612 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153821.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:17:05,056 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4859, 1.4495, 1.3890, 0.8911, 1.5983, 1.7471, 1.7767, 1.3290], + device='cuda:0'), covar=tensor([0.0860, 0.0584, 0.0656, 0.0504, 0.0480, 0.0607, 0.0289, 0.0691], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0149, 0.0131, 0.0124, 0.0133, 0.0132, 0.0144, 0.0152], + device='cuda:0'), out_proj_covar=tensor([8.9586e-05, 1.0675e-04, 9.3043e-05, 8.6828e-05, 9.2909e-05, 9.3263e-05, + 1.0212e-04, 1.0849e-04], device='cuda:0') +2023-03-27 09:17:34,599 INFO [finetune.py:976] (0/7) Epoch 27, batch 4950, loss[loss=0.1518, simple_loss=0.2087, pruned_loss=0.04745, over 4090.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2452, pruned_loss=0.04912, over 955028.68 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:17:35,192 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.496e+02 1.750e+02 2.158e+02 3.393e+02, threshold=3.501e+02, percent-clipped=0.0 +2023-03-27 09:17:57,018 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8709, 2.6130, 2.1519, 1.1868, 2.2137, 2.2593, 2.1704, 2.4041], + device='cuda:0'), covar=tensor([0.0676, 0.0718, 0.1353, 0.1796, 0.1390, 0.2093, 0.1796, 0.0802], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0192, 0.0202, 0.0183, 0.0211, 0.0212, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:18:10,031 INFO [finetune.py:976] (0/7) Epoch 27, batch 5000, loss[loss=0.2015, simple_loss=0.2608, pruned_loss=0.07112, over 4751.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2438, pruned_loss=0.04932, over 953520.91 frames. ], batch size: 59, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:18:12,446 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153923.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:18:28,736 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153934.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:18:39,041 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4658, 1.5796, 1.6280, 1.0046, 1.6741, 1.9985, 1.9595, 1.4674], + device='cuda:0'), covar=tensor([0.0944, 0.0628, 0.0610, 0.0521, 0.0551, 0.0617, 0.0333, 0.0703], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0130, 0.0122, 0.0132, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.8611e-05, 1.0574e-04, 9.2165e-05, 8.6000e-05, 9.2130e-05, 9.2512e-05, + 1.0139e-04, 1.0748e-04], device='cuda:0') +2023-03-27 09:19:01,713 INFO [finetune.py:976] (0/7) Epoch 27, batch 5050, loss[loss=0.1383, simple_loss=0.2155, pruned_loss=0.03057, over 4764.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2405, pruned_loss=0.04855, over 956210.90 frames. ], batch size: 27, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:19:02,311 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.210e+01 1.435e+02 1.808e+02 2.168e+02 4.775e+02, threshold=3.617e+02, percent-clipped=1.0 +2023-03-27 09:19:08,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9242, 1.3064, 1.9755, 1.9446, 1.7549, 1.7081, 1.8059, 1.9208], + device='cuda:0'), covar=tensor([0.3934, 0.4031, 0.3109, 0.3728, 0.4949, 0.4025, 0.4836, 0.2966], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0249, 0.0270, 0.0299, 0.0297, 0.0275, 0.0304, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:19:10,558 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:19:11,834 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153984.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:19:22,917 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-154000.pt +2023-03-27 09:19:33,701 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6770, 1.6019, 1.2565, 0.3898, 1.3690, 1.5211, 1.4670, 1.5589], + device='cuda:0'), covar=tensor([0.0873, 0.0734, 0.1070, 0.1793, 0.1198, 0.2446, 0.2069, 0.0710], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0192, 0.0202, 0.0182, 0.0211, 0.0212, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:19:36,570 INFO [finetune.py:976] (0/7) Epoch 27, batch 5100, loss[loss=0.1629, simple_loss=0.221, pruned_loss=0.0524, over 4726.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2376, pruned_loss=0.04758, over 956440.06 frames. ], batch size: 23, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:19:37,892 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8590, 1.6344, 1.9882, 1.4036, 1.8503, 1.9961, 1.5357, 2.1302], + device='cuda:0'), covar=tensor([0.1223, 0.2211, 0.1335, 0.1655, 0.0861, 0.1245, 0.2860, 0.0794], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0209, 0.0195, 0.0191, 0.0176, 0.0215, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:20:06,422 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154049.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:20:07,179 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 09:20:19,663 INFO [finetune.py:976] (0/7) Epoch 27, batch 5150, loss[loss=0.1791, simple_loss=0.2525, pruned_loss=0.05288, over 4806.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2381, pruned_loss=0.04771, over 957919.17 frames. ], batch size: 45, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:20:20,252 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.883e+01 1.500e+02 1.789e+02 2.109e+02 4.792e+02, threshold=3.578e+02, percent-clipped=3.0 +2023-03-27 09:20:23,491 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154076.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:24,056 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:28,816 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:46,991 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154110.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:20:53,422 INFO [finetune.py:976] (0/7) Epoch 27, batch 5200, loss[loss=0.1693, simple_loss=0.2587, pruned_loss=0.03992, over 4816.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2417, pruned_loss=0.04853, over 955912.97 frames. ], batch size: 51, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:20:56,441 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:21:04,335 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154137.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:21:12,326 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:21:34,724 INFO [finetune.py:976] (0/7) Epoch 27, batch 5250, loss[loss=0.1737, simple_loss=0.2537, pruned_loss=0.04685, over 4862.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2442, pruned_loss=0.04902, over 954934.68 frames. ], batch size: 31, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:21:35,336 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.556e+02 1.889e+02 2.346e+02 3.556e+02, threshold=3.778e+02, percent-clipped=0.0 +2023-03-27 09:22:08,469 INFO [finetune.py:976] (0/7) Epoch 27, batch 5300, loss[loss=0.204, simple_loss=0.2862, pruned_loss=0.06087, over 4830.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2466, pruned_loss=0.0499, over 956587.36 frames. ], batch size: 47, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:22:09,179 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:41,917 INFO [finetune.py:976] (0/7) Epoch 27, batch 5350, loss[loss=0.1669, simple_loss=0.2332, pruned_loss=0.05033, over 4890.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2462, pruned_loss=0.0498, over 955045.79 frames. ], batch size: 32, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:22:42,509 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.999e+01 1.513e+02 1.798e+02 2.139e+02 3.270e+02, threshold=3.596e+02, percent-clipped=0.0 +2023-03-27 09:22:47,876 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154279.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:49,737 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154282.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:22:52,176 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:55,606 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154291.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:23:15,340 INFO [finetune.py:976] (0/7) Epoch 27, batch 5400, loss[loss=0.1443, simple_loss=0.2163, pruned_loss=0.03614, over 4771.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2436, pruned_loss=0.04937, over 955967.48 frames. ], batch size: 28, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:23:17,207 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8277, 3.9565, 3.8658, 1.8763, 4.1202, 3.2370, 1.0823, 2.9986], + device='cuda:0'), covar=tensor([0.2127, 0.1895, 0.1319, 0.3502, 0.1010, 0.0918, 0.4323, 0.1420], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0178, 0.0159, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 09:23:27,670 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6217, 1.6435, 2.2075, 3.3804, 2.2160, 2.4390, 0.9533, 2.7801], + device='cuda:0'), covar=tensor([0.1716, 0.1245, 0.1227, 0.0545, 0.0850, 0.1302, 0.1896, 0.0485], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0164, 0.0100, 0.0135, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:23:27,707 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5616, 2.4507, 2.1063, 0.9840, 2.1006, 2.0308, 1.9918, 2.2414], + device='cuda:0'), covar=tensor([0.0971, 0.0705, 0.1523, 0.2013, 0.1318, 0.2260, 0.1921, 0.0951], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0182, 0.0211, 0.0210, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:23:33,024 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154337.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:23:42,778 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:23:44,616 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 09:23:46,800 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154352.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:24:08,868 INFO [finetune.py:976] (0/7) Epoch 27, batch 5450, loss[loss=0.1497, simple_loss=0.2266, pruned_loss=0.03643, over 4929.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2401, pruned_loss=0.0481, over 955954.11 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:24:09,460 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.752e+01 1.439e+02 1.730e+02 2.063e+02 4.741e+02, threshold=3.460e+02, percent-clipped=1.0 +2023-03-27 09:24:15,605 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1171, 1.9685, 2.6450, 4.0497, 2.8078, 2.8013, 0.9311, 3.2921], + device='cuda:0'), covar=tensor([0.1593, 0.1337, 0.1263, 0.0630, 0.0734, 0.1391, 0.2029, 0.0509], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0164, 0.0100, 0.0135, 0.0124, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:24:26,423 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154398.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:24:31,735 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154405.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:24:42,541 INFO [finetune.py:976] (0/7) Epoch 27, batch 5500, loss[loss=0.1655, simple_loss=0.2334, pruned_loss=0.04878, over 4809.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2372, pruned_loss=0.04718, over 954455.19 frames. ], batch size: 25, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:24:49,883 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154432.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:24:54,769 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154440.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:25:27,010 INFO [finetune.py:976] (0/7) Epoch 27, batch 5550, loss[loss=0.1547, simple_loss=0.223, pruned_loss=0.0432, over 4767.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2386, pruned_loss=0.04728, over 955499.96 frames. ], batch size: 27, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:25:27,600 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.282e+01 1.501e+02 1.802e+02 2.038e+02 5.335e+02, threshold=3.603e+02, percent-clipped=2.0 +2023-03-27 09:25:29,549 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:25:57,467 INFO [finetune.py:976] (0/7) Epoch 27, batch 5600, loss[loss=0.1882, simple_loss=0.2805, pruned_loss=0.04799, over 4811.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.242, pruned_loss=0.04832, over 954696.83 frames. ], batch size: 41, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:26:07,307 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154535.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:26:28,586 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9496, 4.4399, 4.1136, 2.2617, 4.5777, 3.4880, 0.7706, 3.1582], + device='cuda:0'), covar=tensor([0.2382, 0.1659, 0.1429, 0.3129, 0.0751, 0.0779, 0.4411, 0.1391], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0179, 0.0159, 0.0130, 0.0162, 0.0123, 0.0148, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 09:26:30,306 INFO [finetune.py:976] (0/7) Epoch 27, batch 5650, loss[loss=0.2005, simple_loss=0.2798, pruned_loss=0.06062, over 4814.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2446, pruned_loss=0.04834, over 956855.34 frames. ], batch size: 39, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:26:30,864 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.398e+02 1.740e+02 2.119e+02 4.723e+02, threshold=3.480e+02, percent-clipped=2.0 +2023-03-27 09:26:39,111 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154577.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:26:40,304 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154579.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:07,857 INFO [finetune.py:976] (0/7) Epoch 27, batch 5700, loss[loss=0.12, simple_loss=0.1897, pruned_loss=0.02516, over 3866.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.241, pruned_loss=0.048, over 934380.07 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:27:12,037 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154627.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:12,687 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154628.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:20,867 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154642.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:25,042 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-27.pt +2023-03-27 09:27:34,186 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154647.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:34,737 INFO [finetune.py:976] (0/7) Epoch 28, batch 0, loss[loss=0.1644, simple_loss=0.2396, pruned_loss=0.04459, over 4878.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2396, pruned_loss=0.04459, over 4878.00 frames. ], batch size: 32, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:27:34,738 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 09:27:44,389 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1711, 1.4093, 1.3918, 0.7900, 1.3913, 1.6060, 1.6689, 1.3257], + device='cuda:0'), covar=tensor([0.1066, 0.0608, 0.0619, 0.0539, 0.0634, 0.0653, 0.0349, 0.0695], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0129, 0.0121, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.8337e-05, 1.0504e-04, 9.1824e-05, 8.5222e-05, 9.1862e-05, 9.1862e-05, + 1.0083e-04, 1.0708e-04], device='cuda:0') +2023-03-27 09:27:54,287 INFO [finetune.py:1010] (0/7) Epoch 28, validation: loss=0.1583, simple_loss=0.2265, pruned_loss=0.04511, over 2265189.00 frames. +2023-03-27 09:27:54,288 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 09:27:54,989 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9777, 1.6006, 0.8598, 1.8340, 2.2287, 1.5533, 1.8856, 1.7371], + device='cuda:0'), covar=tensor([0.1266, 0.1821, 0.1824, 0.1041, 0.1818, 0.1723, 0.1211, 0.1783], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 09:27:59,648 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7748, 1.6175, 2.2316, 3.2933, 2.2375, 2.6380, 1.4389, 2.8789], + device='cuda:0'), covar=tensor([0.1807, 0.1752, 0.1466, 0.0922, 0.0933, 0.1259, 0.1812, 0.0636], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:28:08,714 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.488e+02 1.773e+02 2.221e+02 3.199e+02, threshold=3.546e+02, percent-clipped=0.0 +2023-03-27 09:28:19,538 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0360, 1.9663, 2.1899, 1.3642, 2.0803, 2.1221, 2.1581, 1.7079], + device='cuda:0'), covar=tensor([0.0583, 0.0666, 0.0579, 0.0934, 0.0648, 0.0659, 0.0527, 0.1142], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0120, 0.0129, 0.0140, 0.0140, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:28:21,216 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154689.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:24,051 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154693.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:28:27,016 INFO [finetune.py:976] (0/7) Epoch 28, batch 50, loss[loss=0.16, simple_loss=0.236, pruned_loss=0.042, over 4922.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2457, pruned_loss=0.05193, over 216085.52 frames. ], batch size: 33, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:28:32,656 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154705.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:28:35,797 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1669, 2.1000, 1.7883, 2.0310, 1.9425, 1.9814, 2.0335, 2.7201], + device='cuda:0'), covar=tensor([0.3425, 0.3993, 0.3171, 0.3627, 0.3727, 0.2309, 0.3525, 0.1572], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0274, 0.0260, 0.0228, 0.0258, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:28:41,926 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154720.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:52,101 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:57,937 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:03,133 INFO [finetune.py:976] (0/7) Epoch 28, batch 100, loss[loss=0.1287, simple_loss=0.2016, pruned_loss=0.02793, over 4898.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2389, pruned_loss=0.04746, over 382071.25 frames. ], batch size: 43, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:29:10,673 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1880, 2.9038, 2.9765, 2.9827, 2.8296, 2.8086, 3.2797, 1.0759], + device='cuda:0'), covar=tensor([0.1915, 0.1834, 0.2008, 0.2248, 0.3033, 0.3241, 0.2039, 0.7665], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0247, 0.0283, 0.0296, 0.0335, 0.0287, 0.0305, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:29:11,906 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154753.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:29:13,666 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1615, 2.1375, 2.3088, 1.4490, 2.2303, 2.2643, 2.2597, 1.8249], + device='cuda:0'), covar=tensor([0.0595, 0.0609, 0.0579, 0.0845, 0.0644, 0.0675, 0.0565, 0.1030], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0138, 0.0141, 0.0120, 0.0128, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:29:26,907 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.413e+02 1.713e+02 2.093e+02 4.180e+02, threshold=3.426e+02, percent-clipped=2.0 +2023-03-27 09:29:32,443 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154780.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:33,106 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:37,833 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:44,899 INFO [finetune.py:976] (0/7) Epoch 28, batch 150, loss[loss=0.1854, simple_loss=0.2533, pruned_loss=0.05877, over 4895.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2353, pruned_loss=0.04614, over 510635.74 frames. ], batch size: 35, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:29:50,543 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 09:29:56,984 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8638, 1.8149, 1.5692, 1.9411, 2.4283, 2.0680, 1.7991, 1.5430], + device='cuda:0'), covar=tensor([0.2179, 0.1871, 0.1794, 0.1602, 0.1560, 0.1109, 0.2193, 0.1852], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0216, 0.0200, 0.0246, 0.0192, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:29:57,695 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 09:29:59,451 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 09:30:06,029 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:30:18,388 INFO [finetune.py:976] (0/7) Epoch 28, batch 200, loss[loss=0.1558, simple_loss=0.2426, pruned_loss=0.03452, over 4772.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2325, pruned_loss=0.04534, over 610701.49 frames. ], batch size: 28, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:30:40,612 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.190e+01 1.565e+02 1.831e+02 2.234e+02 3.641e+02, threshold=3.662e+02, percent-clipped=1.0 +2023-03-27 09:30:48,053 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154877.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:31:02,678 INFO [finetune.py:976] (0/7) Epoch 28, batch 250, loss[loss=0.1441, simple_loss=0.2053, pruned_loss=0.04146, over 4701.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2384, pruned_loss=0.04741, over 687574.17 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:31:14,735 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-27 09:31:20,578 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:31,415 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154942.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:34,492 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:35,502 INFO [finetune.py:976] (0/7) Epoch 28, batch 300, loss[loss=0.163, simple_loss=0.241, pruned_loss=0.0425, over 4840.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2422, pruned_loss=0.04848, over 748411.60 frames. ], batch size: 49, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:31:42,228 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.7564, 1.4932, 1.3899, 0.8852, 1.6381, 1.6916, 1.7927, 1.3410], + device='cuda:0'), covar=tensor([0.0916, 0.0691, 0.0568, 0.0511, 0.0485, 0.0834, 0.0364, 0.0729], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0129, 0.0122, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.8628e-05, 1.0545e-04, 9.1969e-05, 8.5533e-05, 9.1986e-05, 9.1993e-05, + 1.0103e-04, 1.0706e-04], device='cuda:0') +2023-03-27 09:31:51,475 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.523e+02 1.869e+02 2.212e+02 3.864e+02, threshold=3.739e+02, percent-clipped=1.0 +2023-03-27 09:31:59,664 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:07,903 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154984.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:11,537 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154990.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:13,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154993.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:32:15,018 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154995.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:16,795 INFO [finetune.py:976] (0/7) Epoch 28, batch 350, loss[loss=0.1547, simple_loss=0.2397, pruned_loss=0.03482, over 4868.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2439, pruned_loss=0.0489, over 795013.73 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:32:28,694 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5018, 2.4843, 1.9061, 2.6795, 2.3852, 2.0123, 3.0374, 2.5239], + device='cuda:0'), covar=tensor([0.1330, 0.2216, 0.3157, 0.2671, 0.2631, 0.1756, 0.2403, 0.1732], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0254, 0.0250, 0.0207, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:32:31,989 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0921, 3.5344, 3.7937, 3.8567, 3.9003, 3.6879, 4.1543, 1.7651], + device='cuda:0'), covar=tensor([0.0834, 0.0873, 0.0725, 0.1028, 0.1105, 0.1265, 0.0712, 0.5210], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0248, 0.0285, 0.0297, 0.0336, 0.0288, 0.0306, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:32:43,042 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155037.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:45,378 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155041.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:32:49,883 INFO [finetune.py:976] (0/7) Epoch 28, batch 400, loss[loss=0.1635, simple_loss=0.2322, pruned_loss=0.04736, over 4755.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2451, pruned_loss=0.04925, over 829777.34 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:33:13,473 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:15,062 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.906e+01 1.559e+02 1.879e+02 2.352e+02 4.263e+02, threshold=3.758e+02, percent-clipped=3.0 +2023-03-27 09:33:18,274 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155076.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:19,637 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 09:33:31,495 INFO [finetune.py:976] (0/7) Epoch 28, batch 450, loss[loss=0.1401, simple_loss=0.2229, pruned_loss=0.02863, over 4818.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2433, pruned_loss=0.04862, over 855657.59 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:33:54,283 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:54,330 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:34:03,587 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-03-27 09:34:05,149 INFO [finetune.py:976] (0/7) Epoch 28, batch 500, loss[loss=0.1355, simple_loss=0.2121, pruned_loss=0.02952, over 4780.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2408, pruned_loss=0.0478, over 877053.10 frames. ], batch size: 26, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:34:28,557 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.687e+01 1.475e+02 1.683e+02 2.204e+02 4.497e+02, threshold=3.366e+02, percent-clipped=1.0 +2023-03-27 09:34:29,934 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2734, 1.9282, 1.5452, 0.5571, 1.7030, 1.9147, 1.7405, 1.8707], + device='cuda:0'), covar=tensor([0.1010, 0.0910, 0.1521, 0.2108, 0.1468, 0.2428, 0.2325, 0.0885], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0191, 0.0200, 0.0180, 0.0209, 0.0209, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:34:33,437 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:34:38,427 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4145, 1.7207, 2.3374, 1.8006, 2.0009, 4.5024, 1.8144, 1.8829], + device='cuda:0'), covar=tensor([0.0999, 0.1666, 0.0980, 0.0939, 0.1361, 0.0152, 0.1350, 0.1634], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 09:34:49,245 INFO [finetune.py:976] (0/7) Epoch 28, batch 550, loss[loss=0.1789, simple_loss=0.2354, pruned_loss=0.06123, over 4935.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2374, pruned_loss=0.04672, over 893893.26 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:35:03,623 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6910, 1.6057, 1.4706, 1.6513, 2.0502, 1.9039, 1.7339, 1.4710], + device='cuda:0'), covar=tensor([0.0386, 0.0363, 0.0652, 0.0339, 0.0206, 0.0513, 0.0319, 0.0463], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0111, 0.0102, 0.0116, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.8264e-05, 8.1370e-05, 1.1529e-04, 8.4988e-05, 7.8724e-05, 8.5108e-05, + 7.6631e-05, 8.6072e-05], device='cuda:0') +2023-03-27 09:35:23,082 INFO [finetune.py:976] (0/7) Epoch 28, batch 600, loss[loss=0.2294, simple_loss=0.3074, pruned_loss=0.07565, over 4815.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2388, pruned_loss=0.04771, over 907335.06 frames. ], batch size: 51, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:35:38,977 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.640e+01 1.454e+02 1.702e+02 1.998e+02 4.828e+02, threshold=3.403e+02, percent-clipped=2.0 +2023-03-27 09:35:53,262 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155284.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:05,097 INFO [finetune.py:976] (0/7) Epoch 28, batch 650, loss[loss=0.1728, simple_loss=0.2488, pruned_loss=0.04841, over 4915.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2408, pruned_loss=0.04807, over 914723.64 frames. ], batch size: 36, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:36:07,082 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3931, 2.2186, 2.3603, 1.6946, 2.2654, 2.4814, 2.4085, 1.7395], + device='cuda:0'), covar=tensor([0.0539, 0.0698, 0.0688, 0.0876, 0.0763, 0.0681, 0.0613, 0.1316], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0138, 0.0141, 0.0119, 0.0128, 0.0139, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:36:29,094 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:29,103 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:38,729 INFO [finetune.py:976] (0/7) Epoch 28, batch 700, loss[loss=0.184, simple_loss=0.2585, pruned_loss=0.05479, over 4894.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2421, pruned_loss=0.04851, over 923459.63 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:36:43,902 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-27 09:36:54,659 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.563e+02 1.812e+02 2.261e+02 4.160e+02, threshold=3.625e+02, percent-clipped=3.0 +2023-03-27 09:36:57,810 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:19,506 INFO [finetune.py:976] (0/7) Epoch 28, batch 750, loss[loss=0.2055, simple_loss=0.2825, pruned_loss=0.06425, over 4902.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2444, pruned_loss=0.04946, over 930311.62 frames. ], batch size: 36, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:37:40,154 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155424.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:41,312 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:56,711 INFO [finetune.py:976] (0/7) Epoch 28, batch 800, loss[loss=0.1663, simple_loss=0.2157, pruned_loss=0.05847, over 4053.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2445, pruned_loss=0.04949, over 933942.55 frames. ], batch size: 17, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:38:02,334 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155457.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:38:11,933 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.495e+02 1.724e+02 1.967e+02 3.002e+02, threshold=3.447e+02, percent-clipped=0.0 +2023-03-27 09:38:39,880 INFO [finetune.py:976] (0/7) Epoch 28, batch 850, loss[loss=0.1687, simple_loss=0.2341, pruned_loss=0.05165, over 4873.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2427, pruned_loss=0.04898, over 938001.18 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:38:52,637 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155518.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:39:09,171 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2811, 1.3762, 1.5129, 1.0870, 1.2441, 1.4194, 1.3477, 1.6545], + device='cuda:0'), covar=tensor([0.1345, 0.2392, 0.1455, 0.1580, 0.1093, 0.1358, 0.3192, 0.0972], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0209, 0.0195, 0.0191, 0.0175, 0.0215, 0.0221, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:39:13,737 INFO [finetune.py:976] (0/7) Epoch 28, batch 900, loss[loss=0.1518, simple_loss=0.2152, pruned_loss=0.04415, over 4058.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2398, pruned_loss=0.04823, over 941940.76 frames. ], batch size: 17, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:39:28,255 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.783e+01 1.413e+02 1.787e+02 2.288e+02 4.282e+02, threshold=3.575e+02, percent-clipped=3.0 +2023-03-27 09:39:32,996 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8846, 1.8882, 1.6179, 1.9899, 2.3002, 2.0455, 1.7568, 1.5522], + device='cuda:0'), covar=tensor([0.2371, 0.2043, 0.2078, 0.1743, 0.1770, 0.1253, 0.2517, 0.2112], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0213, 0.0216, 0.0200, 0.0247, 0.0191, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:39:37,765 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9074, 1.3938, 1.9885, 1.9767, 1.7855, 1.7489, 1.9041, 1.9055], + device='cuda:0'), covar=tensor([0.4132, 0.3933, 0.3139, 0.3781, 0.4579, 0.3639, 0.4279, 0.2925], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0248, 0.0269, 0.0298, 0.0296, 0.0273, 0.0302, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:39:54,475 INFO [finetune.py:976] (0/7) Epoch 28, batch 950, loss[loss=0.1474, simple_loss=0.2156, pruned_loss=0.03954, over 4818.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2373, pruned_loss=0.04728, over 945009.39 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:40:02,641 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:20,529 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:31,719 INFO [finetune.py:976] (0/7) Epoch 28, batch 1000, loss[loss=0.2086, simple_loss=0.2832, pruned_loss=0.06701, over 4810.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2405, pruned_loss=0.04859, over 948067.80 frames. ], batch size: 45, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:40:37,297 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155657.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:41,577 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0900, 2.0527, 1.7124, 1.9594, 1.9264, 1.9574, 1.9622, 2.6403], + device='cuda:0'), covar=tensor([0.3622, 0.3996, 0.3183, 0.3614, 0.3887, 0.2314, 0.3571, 0.1589], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0265, 0.0239, 0.0277, 0.0262, 0.0231, 0.0261, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:40:42,784 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:45,718 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.619e+01 1.519e+02 1.814e+02 2.190e+02 3.109e+02, threshold=3.628e+02, percent-clipped=0.0 +2023-03-27 09:40:54,174 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155680.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:13,979 INFO [finetune.py:976] (0/7) Epoch 28, batch 1050, loss[loss=0.1534, simple_loss=0.2112, pruned_loss=0.0478, over 4560.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2439, pruned_loss=0.04932, over 951321.03 frames. ], batch size: 20, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:41:23,709 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6001, 1.4599, 1.3008, 1.6356, 1.7003, 1.6466, 1.1274, 1.3603], + device='cuda:0'), covar=tensor([0.2105, 0.1990, 0.1865, 0.1657, 0.1455, 0.1167, 0.2384, 0.1807], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0213, 0.0217, 0.0200, 0.0248, 0.0192, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:41:26,783 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:30,998 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:43,213 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6288, 1.5089, 2.2385, 3.2259, 2.0962, 2.3660, 1.1092, 2.7375], + device='cuda:0'), covar=tensor([0.1694, 0.1414, 0.1193, 0.0562, 0.0837, 0.1662, 0.1698, 0.0442], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0165, 0.0101, 0.0136, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:41:46,713 INFO [finetune.py:976] (0/7) Epoch 28, batch 1100, loss[loss=0.1599, simple_loss=0.2446, pruned_loss=0.0376, over 4785.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2449, pruned_loss=0.04915, over 952269.75 frames. ], batch size: 29, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:42:01,622 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.629e+02 1.915e+02 2.256e+02 9.973e+02, threshold=3.830e+02, percent-clipped=2.0 +2023-03-27 09:42:02,893 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:42:10,610 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:42:21,324 INFO [finetune.py:976] (0/7) Epoch 28, batch 1150, loss[loss=0.1848, simple_loss=0.2509, pruned_loss=0.05932, over 4738.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2453, pruned_loss=0.04859, over 954201.17 frames. ], batch size: 54, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:42:39,752 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155813.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:43:01,209 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155845.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:43:02,854 INFO [finetune.py:976] (0/7) Epoch 28, batch 1200, loss[loss=0.164, simple_loss=0.2338, pruned_loss=0.04707, over 4819.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2435, pruned_loss=0.0482, over 953662.52 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:43:18,194 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.526e+02 1.746e+02 2.167e+02 3.236e+02, threshold=3.492e+02, percent-clipped=0.0 +2023-03-27 09:43:45,613 INFO [finetune.py:976] (0/7) Epoch 28, batch 1250, loss[loss=0.1486, simple_loss=0.22, pruned_loss=0.03859, over 4849.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2407, pruned_loss=0.04753, over 955143.88 frames. ], batch size: 49, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:44:01,999 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-27 09:44:15,234 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.61 vs. limit=5.0 +2023-03-27 09:44:19,202 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7846, 1.8490, 1.6410, 2.1026, 2.2607, 2.1835, 1.8106, 1.5187], + device='cuda:0'), covar=tensor([0.2352, 0.1902, 0.1890, 0.1558, 0.2000, 0.1112, 0.2333, 0.1968], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0212, 0.0215, 0.0199, 0.0246, 0.0190, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:44:22,108 INFO [finetune.py:976] (0/7) Epoch 28, batch 1300, loss[loss=0.1272, simple_loss=0.2024, pruned_loss=0.02598, over 4745.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2375, pruned_loss=0.04682, over 955602.47 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:44:32,014 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155961.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:44:38,004 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.818e+01 1.476e+02 1.729e+02 2.197e+02 4.050e+02, threshold=3.458e+02, percent-clipped=1.0 +2023-03-27 09:44:55,319 INFO [finetune.py:976] (0/7) Epoch 28, batch 1350, loss[loss=0.198, simple_loss=0.2645, pruned_loss=0.06578, over 4688.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.239, pruned_loss=0.04785, over 956310.41 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:44:56,674 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-156000.pt +2023-03-27 09:45:10,170 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156013.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:45:32,752 INFO [finetune.py:976] (0/7) Epoch 28, batch 1400, loss[loss=0.1285, simple_loss=0.2, pruned_loss=0.02855, over 4725.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2426, pruned_loss=0.04928, over 955009.20 frames. ], batch size: 23, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:45:48,231 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156070.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:45:48,708 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.532e+02 1.806e+02 2.221e+02 4.474e+02, threshold=3.612e+02, percent-clipped=3.0 +2023-03-27 09:46:02,171 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 09:46:06,151 INFO [finetune.py:976] (0/7) Epoch 28, batch 1450, loss[loss=0.167, simple_loss=0.2409, pruned_loss=0.04656, over 4744.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2448, pruned_loss=0.04955, over 955670.08 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:46:23,163 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156113.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:46:35,162 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156131.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:46:40,537 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156140.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:46:45,776 INFO [finetune.py:976] (0/7) Epoch 28, batch 1500, loss[loss=0.1808, simple_loss=0.2502, pruned_loss=0.0557, over 4822.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2446, pruned_loss=0.0492, over 955229.60 frames. ], batch size: 30, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:46:54,174 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:46:58,358 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2037, 2.1762, 2.2890, 1.6168, 2.1636, 2.3377, 2.3766, 1.8696], + device='cuda:0'), covar=tensor([0.0575, 0.0650, 0.0664, 0.0849, 0.0711, 0.0702, 0.0543, 0.1069], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0139, 0.0141, 0.0120, 0.0129, 0.0139, 0.0140, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:47:02,076 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.585e+02 1.899e+02 2.331e+02 3.577e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-03-27 09:47:18,929 INFO [finetune.py:976] (0/7) Epoch 28, batch 1550, loss[loss=0.1896, simple_loss=0.2605, pruned_loss=0.05932, over 4933.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2447, pruned_loss=0.04896, over 956040.47 frames. ], batch size: 33, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:47:20,422 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.2219, 2.0035, 2.1230, 0.9559, 2.4225, 2.6919, 2.2537, 1.8761], + device='cuda:0'), covar=tensor([0.1162, 0.0805, 0.0579, 0.0731, 0.0548, 0.0683, 0.0527, 0.0853], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0130, 0.0123, 0.0132, 0.0131, 0.0143, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.8875e-05, 1.0592e-04, 9.2763e-05, 8.6086e-05, 9.2364e-05, 9.2622e-05, + 1.0149e-04, 1.0737e-04], device='cuda:0') +2023-03-27 09:47:30,552 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:47:59,289 INFO [finetune.py:976] (0/7) Epoch 28, batch 1600, loss[loss=0.1786, simple_loss=0.2407, pruned_loss=0.05824, over 4826.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2419, pruned_loss=0.04806, over 955051.05 frames. ], batch size: 39, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:48:08,661 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156261.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:15,839 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.435e+02 1.767e+02 2.111e+02 3.704e+02, threshold=3.535e+02, percent-clipped=0.0 +2023-03-27 09:48:19,968 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:20,760 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 09:48:27,876 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156290.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:32,611 INFO [finetune.py:976] (0/7) Epoch 28, batch 1650, loss[loss=0.1295, simple_loss=0.2066, pruned_loss=0.02622, over 4817.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2393, pruned_loss=0.04707, over 954538.63 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:48:40,816 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156309.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:43,252 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156313.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:04,232 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9913, 1.8235, 1.5987, 1.6440, 1.7220, 1.7451, 1.7974, 2.5042], + device='cuda:0'), covar=tensor([0.3436, 0.3600, 0.3061, 0.3466, 0.4033, 0.2159, 0.3437, 0.1459], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0264, 0.0237, 0.0276, 0.0261, 0.0230, 0.0259, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:49:18,367 INFO [finetune.py:976] (0/7) Epoch 28, batch 1700, loss[loss=0.2059, simple_loss=0.265, pruned_loss=0.07338, over 4820.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2372, pruned_loss=0.0464, over 952679.78 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:49:20,356 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156351.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:26,952 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156361.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:29,217 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4461, 1.3870, 1.2568, 1.5868, 1.6219, 1.5788, 1.0931, 1.2854], + device='cuda:0'), covar=tensor([0.2298, 0.2072, 0.1975, 0.1613, 0.1759, 0.1287, 0.2565, 0.1916], + device='cuda:0'), in_proj_covar=tensor([0.0245, 0.0212, 0.0215, 0.0199, 0.0246, 0.0190, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:49:34,436 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.260e+01 1.525e+02 1.796e+02 2.204e+02 4.546e+02, threshold=3.593e+02, percent-clipped=3.0 +2023-03-27 09:49:51,286 INFO [finetune.py:976] (0/7) Epoch 28, batch 1750, loss[loss=0.1935, simple_loss=0.2616, pruned_loss=0.06267, over 4747.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2404, pruned_loss=0.04827, over 953657.91 frames. ], batch size: 54, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:49:57,119 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5663, 2.3904, 2.1862, 1.0826, 2.3509, 1.9063, 1.8194, 2.2606], + device='cuda:0'), covar=tensor([0.0936, 0.0731, 0.1561, 0.1899, 0.1216, 0.2211, 0.2251, 0.0940], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0191, 0.0201, 0.0181, 0.0210, 0.0210, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:50:09,513 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156426.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:50:19,439 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156440.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:50:24,159 INFO [finetune.py:976] (0/7) Epoch 28, batch 1800, loss[loss=0.2439, simple_loss=0.3065, pruned_loss=0.09069, over 4906.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2426, pruned_loss=0.04862, over 954400.24 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:50:27,456 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-03-27 09:50:39,992 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.522e+02 1.851e+02 2.291e+02 4.651e+02, threshold=3.702e+02, percent-clipped=5.0 +2023-03-27 09:50:51,501 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156488.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:50:57,061 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2197, 2.2216, 2.3393, 1.5291, 2.2649, 2.4225, 2.3566, 1.9483], + device='cuda:0'), covar=tensor([0.0602, 0.0631, 0.0625, 0.0887, 0.0717, 0.0643, 0.0565, 0.1064], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0118, 0.0128, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:50:57,177 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.73 vs. limit=5.0 +2023-03-27 09:50:57,542 INFO [finetune.py:976] (0/7) Epoch 28, batch 1850, loss[loss=0.123, simple_loss=0.1941, pruned_loss=0.02596, over 4770.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2439, pruned_loss=0.04901, over 954110.93 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:51:19,564 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 09:51:40,568 INFO [finetune.py:976] (0/7) Epoch 28, batch 1900, loss[loss=0.1837, simple_loss=0.2433, pruned_loss=0.06204, over 4115.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2446, pruned_loss=0.04935, over 953711.11 frames. ], batch size: 65, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:51:52,196 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8058, 1.7316, 2.2825, 3.2835, 2.2594, 2.5182, 1.3732, 2.8141], + device='cuda:0'), covar=tensor([0.1591, 0.1300, 0.1173, 0.0680, 0.0760, 0.1273, 0.1532, 0.0577], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0166, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 09:51:56,034 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.831e+01 1.540e+02 1.871e+02 2.242e+02 4.934e+02, threshold=3.741e+02, percent-clipped=1.0 +2023-03-27 09:51:56,113 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:52:13,684 INFO [finetune.py:976] (0/7) Epoch 28, batch 1950, loss[loss=0.1625, simple_loss=0.2269, pruned_loss=0.04904, over 4753.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2423, pruned_loss=0.04829, over 953423.79 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:52:16,391 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0201, 1.3377, 1.3825, 1.2913, 1.4515, 2.4155, 1.2826, 1.4877], + device='cuda:0'), covar=tensor([0.1039, 0.1718, 0.1063, 0.0894, 0.1486, 0.0379, 0.1425, 0.1696], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 09:52:24,926 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4604, 1.4623, 1.3614, 1.4753, 1.7355, 1.6556, 1.5348, 1.3164], + device='cuda:0'), covar=tensor([0.0373, 0.0342, 0.0674, 0.0312, 0.0282, 0.0553, 0.0320, 0.0440], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0106, 0.0149, 0.0112, 0.0102, 0.0116, 0.0104, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.8570e-05, 8.1298e-05, 1.1589e-04, 8.5337e-05, 7.8713e-05, 8.5471e-05, + 7.7108e-05, 8.6230e-05], device='cuda:0') +2023-03-27 09:52:24,940 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8427, 1.7917, 1.5250, 1.9792, 2.1771, 1.9926, 1.5364, 1.5367], + device='cuda:0'), covar=tensor([0.2120, 0.1923, 0.1958, 0.1559, 0.1619, 0.1177, 0.2359, 0.1881], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0210, 0.0214, 0.0198, 0.0244, 0.0189, 0.0215, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:52:46,381 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:52:47,506 INFO [finetune.py:976] (0/7) Epoch 28, batch 2000, loss[loss=0.1733, simple_loss=0.2488, pruned_loss=0.04889, over 4802.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2401, pruned_loss=0.04761, over 954245.97 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:53:04,320 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.919e+01 1.529e+02 1.760e+02 2.169e+02 4.761e+02, threshold=3.520e+02, percent-clipped=1.0 +2023-03-27 09:53:20,830 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4921, 2.4890, 2.6819, 1.9955, 2.5333, 2.7897, 2.7786, 2.2127], + device='cuda:0'), covar=tensor([0.0565, 0.0600, 0.0589, 0.0794, 0.0784, 0.0567, 0.0485, 0.1037], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0140, 0.0118, 0.0128, 0.0138, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:53:21,453 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5590, 2.4662, 1.9259, 2.7941, 2.4910, 2.1686, 2.9759, 2.6648], + device='cuda:0'), covar=tensor([0.1208, 0.2066, 0.2855, 0.2254, 0.2398, 0.1663, 0.2731, 0.1519], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0255, 0.0251, 0.0208, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:53:29,981 INFO [finetune.py:976] (0/7) Epoch 28, batch 2050, loss[loss=0.152, simple_loss=0.2289, pruned_loss=0.03759, over 4700.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2374, pruned_loss=0.04684, over 953725.64 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:53:41,429 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4430, 3.8576, 4.0320, 4.2349, 4.1936, 3.9387, 4.4979, 1.3484], + device='cuda:0'), covar=tensor([0.0749, 0.0869, 0.0856, 0.0974, 0.1221, 0.1723, 0.0710, 0.6277], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0244, 0.0282, 0.0293, 0.0332, 0.0286, 0.0303, 0.0302], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:53:47,457 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156726.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:53:56,866 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 09:54:08,961 INFO [finetune.py:976] (0/7) Epoch 28, batch 2100, loss[loss=0.1872, simple_loss=0.2665, pruned_loss=0.05398, over 4852.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2389, pruned_loss=0.04739, over 954505.37 frames. ], batch size: 44, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:54:09,684 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156749.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:54:37,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.521e+02 1.844e+02 2.179e+02 3.224e+02, threshold=3.687e+02, percent-clipped=0.0 +2023-03-27 09:54:38,469 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156774.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:54:49,239 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 09:54:54,965 INFO [finetune.py:976] (0/7) Epoch 28, batch 2150, loss[loss=0.1838, simple_loss=0.2444, pruned_loss=0.06161, over 4801.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2415, pruned_loss=0.04839, over 954368.75 frames. ], batch size: 25, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:55:03,474 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156810.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:55:27,792 INFO [finetune.py:976] (0/7) Epoch 28, batch 2200, loss[loss=0.178, simple_loss=0.2609, pruned_loss=0.04751, over 4926.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2431, pruned_loss=0.04868, over 951889.46 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:55:44,138 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:55:44,654 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.543e+02 1.740e+02 2.127e+02 4.555e+02, threshold=3.480e+02, percent-clipped=1.0 +2023-03-27 09:55:52,565 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156885.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:00,343 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 09:56:01,289 INFO [finetune.py:976] (0/7) Epoch 28, batch 2250, loss[loss=0.1371, simple_loss=0.2137, pruned_loss=0.03018, over 4739.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2432, pruned_loss=0.0485, over 954464.32 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:56:16,207 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156920.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:32,862 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:32,881 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:33,999 INFO [finetune.py:976] (0/7) Epoch 28, batch 2300, loss[loss=0.2055, simple_loss=0.2764, pruned_loss=0.06734, over 4812.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2439, pruned_loss=0.04897, over 952747.57 frames. ], batch size: 39, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:57:00,119 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.437e+02 1.655e+02 2.039e+02 3.893e+02, threshold=3.311e+02, percent-clipped=1.0 +2023-03-27 09:57:17,059 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:57:20,396 INFO [finetune.py:976] (0/7) Epoch 28, batch 2350, loss[loss=0.1728, simple_loss=0.2486, pruned_loss=0.04851, over 4901.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2411, pruned_loss=0.04827, over 951908.18 frames. ], batch size: 37, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:57:28,259 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157010.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:57:52,971 INFO [finetune.py:976] (0/7) Epoch 28, batch 2400, loss[loss=0.1396, simple_loss=0.2163, pruned_loss=0.03143, over 4723.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2403, pruned_loss=0.04883, over 953144.30 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:57:54,814 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2409, 2.0980, 1.7220, 1.8067, 2.1427, 1.8715, 2.2727, 2.1894], + device='cuda:0'), covar=tensor([0.1273, 0.1750, 0.2996, 0.2480, 0.2557, 0.1715, 0.2907, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0257, 0.0252, 0.0209, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:58:08,943 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157071.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:58:12,820 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.490e+02 1.799e+02 2.218e+02 3.254e+02, threshold=3.597e+02, percent-clipped=0.0 +2023-03-27 09:58:28,593 INFO [finetune.py:976] (0/7) Epoch 28, batch 2450, loss[loss=0.2312, simple_loss=0.2939, pruned_loss=0.08429, over 4917.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2372, pruned_loss=0.04766, over 951868.51 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:58:33,538 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:58:57,725 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:58:58,343 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7016, 2.5713, 2.2077, 2.7982, 2.5693, 2.3129, 2.8172, 2.6655], + device='cuda:0'), covar=tensor([0.1197, 0.1837, 0.2667, 0.2125, 0.2354, 0.1641, 0.2703, 0.1578], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0256, 0.0252, 0.0209, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:59:01,794 INFO [finetune.py:976] (0/7) Epoch 28, batch 2500, loss[loss=0.2193, simple_loss=0.3075, pruned_loss=0.06557, over 4799.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2401, pruned_loss=0.04869, over 950691.02 frames. ], batch size: 51, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:59:15,523 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0307, 1.8975, 1.6257, 1.8116, 1.8079, 1.7760, 1.8806, 2.5186], + device='cuda:0'), covar=tensor([0.3648, 0.4072, 0.3237, 0.3635, 0.3907, 0.2408, 0.3515, 0.1765], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0238, 0.0276, 0.0262, 0.0231, 0.0259, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 09:59:23,966 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.586e+02 1.871e+02 2.257e+02 5.817e+02, threshold=3.742e+02, percent-clipped=3.0 +2023-03-27 09:59:51,556 INFO [finetune.py:976] (0/7) Epoch 28, batch 2550, loss[loss=0.1702, simple_loss=0.2628, pruned_loss=0.03877, over 4901.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2446, pruned_loss=0.04985, over 950118.64 frames. ], batch size: 43, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:59:55,383 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157203.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:00:20,145 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.5038, 1.4184, 1.4221, 0.7428, 1.5489, 1.7334, 1.7483, 1.3180], + device='cuda:0'), covar=tensor([0.0977, 0.0770, 0.0531, 0.0651, 0.0472, 0.0700, 0.0372, 0.0794], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0130, 0.0123, 0.0132, 0.0131, 0.0143, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.9057e-05, 1.0652e-04, 9.2497e-05, 8.6411e-05, 9.2720e-05, 9.2504e-05, + 1.0179e-04, 1.0806e-04], device='cuda:0') +2023-03-27 10:00:20,725 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:00:24,904 INFO [finetune.py:976] (0/7) Epoch 28, batch 2600, loss[loss=0.1903, simple_loss=0.276, pruned_loss=0.05228, over 4819.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2469, pruned_loss=0.05099, over 951208.10 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:00:41,899 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.561e+02 1.846e+02 2.212e+02 4.271e+02, threshold=3.692e+02, percent-clipped=1.0 +2023-03-27 10:00:57,758 INFO [finetune.py:976] (0/7) Epoch 28, batch 2650, loss[loss=0.1362, simple_loss=0.202, pruned_loss=0.03518, over 4715.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.248, pruned_loss=0.05126, over 952665.39 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:01:30,662 INFO [finetune.py:976] (0/7) Epoch 28, batch 2700, loss[loss=0.163, simple_loss=0.2436, pruned_loss=0.04123, over 4783.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2457, pruned_loss=0.05008, over 951195.74 frames. ], batch size: 51, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:01:35,623 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:01:42,590 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157366.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:01:47,170 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.455e+02 1.749e+02 2.146e+02 4.370e+02, threshold=3.498e+02, percent-clipped=1.0 +2023-03-27 10:01:53,271 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157382.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:01:53,288 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6452, 1.5489, 1.3925, 1.7222, 1.6504, 1.7054, 1.0658, 1.4367], + device='cuda:0'), covar=tensor([0.2124, 0.2021, 0.1875, 0.1669, 0.1574, 0.1225, 0.2366, 0.1872], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0213, 0.0217, 0.0201, 0.0247, 0.0192, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:02:12,808 INFO [finetune.py:976] (0/7) Epoch 28, batch 2750, loss[loss=0.1842, simple_loss=0.258, pruned_loss=0.05517, over 4920.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2437, pruned_loss=0.04984, over 952664.98 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:02:20,342 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157404.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:02:20,945 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157405.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:28,669 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157417.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:40,445 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8081, 1.7904, 1.7234, 1.7282, 1.5069, 3.8568, 1.6186, 2.0968], + device='cuda:0'), covar=tensor([0.3320, 0.2542, 0.2026, 0.2346, 0.1589, 0.0217, 0.2371, 0.1157], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0114, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 10:02:46,843 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157443.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:49,206 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8827, 1.2581, 1.9438, 1.9410, 1.7320, 1.7037, 1.8133, 1.8696], + device='cuda:0'), covar=tensor([0.3848, 0.3779, 0.3035, 0.3455, 0.4592, 0.3543, 0.4257, 0.2845], + device='cuda:0'), in_proj_covar=tensor([0.0269, 0.0250, 0.0270, 0.0299, 0.0299, 0.0275, 0.0304, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:02:50,279 INFO [finetune.py:976] (0/7) Epoch 28, batch 2800, loss[loss=0.1343, simple_loss=0.2126, pruned_loss=0.02799, over 4769.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2398, pruned_loss=0.04866, over 953080.33 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:02:53,300 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157453.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:03:01,083 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157465.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:03:05,884 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.67 vs. limit=5.0 +2023-03-27 10:03:06,316 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.288e+01 1.520e+02 1.688e+02 2.071e+02 7.416e+02, threshold=3.376e+02, percent-clipped=4.0 +2023-03-27 10:03:12,302 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5413, 1.4839, 2.2200, 3.3482, 2.1938, 2.3627, 0.9026, 2.7776], + device='cuda:0'), covar=tensor([0.1814, 0.1454, 0.1231, 0.0549, 0.0868, 0.1523, 0.1979, 0.0473], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0166, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 10:03:14,245 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.61 vs. limit=2.0 +2023-03-27 10:03:21,672 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([5.3799, 4.6531, 4.9303, 5.2002, 5.0893, 4.8004, 5.5105, 1.6747], + device='cuda:0'), covar=tensor([0.0646, 0.0935, 0.0776, 0.0838, 0.1118, 0.1659, 0.0485, 0.5752], + device='cuda:0'), in_proj_covar=tensor([0.0350, 0.0245, 0.0284, 0.0294, 0.0334, 0.0285, 0.0305, 0.0303], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:03:22,348 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4674, 1.4717, 1.4719, 0.7471, 1.5959, 1.7839, 1.7630, 1.4046], + device='cuda:0'), covar=tensor([0.1072, 0.0715, 0.0555, 0.0640, 0.0496, 0.0649, 0.0349, 0.0770], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0149, 0.0130, 0.0123, 0.0133, 0.0131, 0.0143, 0.0152], + device='cuda:0'), out_proj_covar=tensor([8.9032e-05, 1.0673e-04, 9.2538e-05, 8.6573e-05, 9.2875e-05, 9.2852e-05, + 1.0159e-04, 1.0835e-04], device='cuda:0') +2023-03-27 10:03:23,413 INFO [finetune.py:976] (0/7) Epoch 28, batch 2850, loss[loss=0.1694, simple_loss=0.2391, pruned_loss=0.04981, over 4169.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2389, pruned_loss=0.0483, over 953143.62 frames. ], batch size: 65, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:03:23,480 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:03:28,870 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157506.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:03:43,584 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6929, 4.1693, 3.8278, 1.9759, 4.3045, 3.2958, 0.7027, 2.9920], + device='cuda:0'), covar=tensor([0.2348, 0.1998, 0.1663, 0.3258, 0.1060, 0.0878, 0.4672, 0.1469], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0180, 0.0160, 0.0130, 0.0164, 0.0124, 0.0148, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 10:03:45,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3343, 1.2704, 1.7690, 2.3937, 1.5471, 2.1747, 0.8777, 2.0817], + device='cuda:0'), covar=tensor([0.1834, 0.1471, 0.1075, 0.0789, 0.1025, 0.1268, 0.1619, 0.0599], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0166, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 10:03:52,490 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157541.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:03:57,079 INFO [finetune.py:976] (0/7) Epoch 28, batch 2900, loss[loss=0.193, simple_loss=0.2412, pruned_loss=0.07246, over 4698.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2417, pruned_loss=0.04877, over 952969.64 frames. ], batch size: 23, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:04:05,001 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7347, 1.6127, 1.4337, 1.8778, 1.9344, 1.8082, 1.4008, 1.4382], + device='cuda:0'), covar=tensor([0.2219, 0.1895, 0.1947, 0.1603, 0.1655, 0.1138, 0.2355, 0.1934], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0200, 0.0246, 0.0191, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:04:09,770 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157567.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:04:10,388 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157568.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:04:13,242 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.581e+01 1.484e+02 1.768e+02 2.098e+02 4.175e+02, threshold=3.535e+02, percent-clipped=1.0 +2023-03-27 10:04:18,693 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-03-27 10:04:24,445 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:04:32,418 INFO [finetune.py:976] (0/7) Epoch 28, batch 2950, loss[loss=0.1575, simple_loss=0.2302, pruned_loss=0.04238, over 4832.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2449, pruned_loss=0.04977, over 953087.37 frames. ], batch size: 30, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:04:52,133 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 10:05:06,831 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:05:23,717 INFO [finetune.py:976] (0/7) Epoch 28, batch 3000, loss[loss=0.1836, simple_loss=0.2565, pruned_loss=0.05532, over 4842.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2461, pruned_loss=0.04997, over 953445.83 frames. ], batch size: 44, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:05:23,719 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 10:05:27,278 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8650, 1.1482, 1.9626, 1.8930, 1.7140, 1.6596, 1.7319, 1.8651], + device='cuda:0'), covar=tensor([0.4125, 0.3996, 0.3478, 0.3870, 0.5249, 0.3998, 0.4645, 0.3080], + device='cuda:0'), in_proj_covar=tensor([0.0268, 0.0249, 0.0269, 0.0297, 0.0297, 0.0274, 0.0303, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:05:34,509 INFO [finetune.py:1010] (0/7) Epoch 28, validation: loss=0.1567, simple_loss=0.2243, pruned_loss=0.04455, over 2265189.00 frames. +2023-03-27 10:05:34,510 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 10:05:46,459 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:05:50,618 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.524e+02 1.841e+02 2.248e+02 4.082e+02, threshold=3.682e+02, percent-clipped=3.0 +2023-03-27 10:06:07,189 INFO [finetune.py:976] (0/7) Epoch 28, batch 3050, loss[loss=0.1796, simple_loss=0.2539, pruned_loss=0.05269, over 4777.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2462, pruned_loss=0.04919, over 954926.21 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:06:16,639 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157712.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:17,819 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157714.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:27,378 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9271, 1.9702, 1.6158, 2.1734, 2.4823, 2.1371, 1.9429, 1.5966], + device='cuda:0'), covar=tensor([0.2147, 0.1748, 0.1787, 0.1487, 0.1553, 0.1085, 0.1927, 0.1784], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0245, 0.0190, 0.0217, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:06:33,273 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157738.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:40,233 INFO [finetune.py:976] (0/7) Epoch 28, batch 3100, loss[loss=0.1604, simple_loss=0.2317, pruned_loss=0.04458, over 4719.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2434, pruned_loss=0.04814, over 956774.24 frames. ], batch size: 59, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:06:48,824 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157760.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:06:57,051 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.389e+02 1.744e+02 2.105e+02 3.209e+02, threshold=3.488e+02, percent-clipped=0.0 +2023-03-27 10:07:19,510 INFO [finetune.py:976] (0/7) Epoch 28, batch 3150, loss[loss=0.1986, simple_loss=0.2575, pruned_loss=0.06982, over 4753.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2407, pruned_loss=0.04787, over 956657.24 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:07:19,629 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157798.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:07:27,436 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-27 10:07:54,538 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-27 10:08:03,987 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157846.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:05,677 INFO [finetune.py:976] (0/7) Epoch 28, batch 3200, loss[loss=0.2008, simple_loss=0.2783, pruned_loss=0.06169, over 4920.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2369, pruned_loss=0.04641, over 955583.63 frames. ], batch size: 43, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:08:09,369 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157853.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:15,273 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157862.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:08:22,848 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.817e+01 1.555e+02 1.831e+02 2.254e+02 7.078e+02, threshold=3.662e+02, percent-clipped=7.0 +2023-03-27 10:08:38,478 INFO [finetune.py:976] (0/7) Epoch 28, batch 3250, loss[loss=0.1309, simple_loss=0.2043, pruned_loss=0.02876, over 4826.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2369, pruned_loss=0.04668, over 953242.70 frames. ], batch size: 25, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:08:42,798 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1432, 2.1276, 1.7211, 2.0605, 1.9816, 1.9318, 2.0091, 2.6868], + device='cuda:0'), covar=tensor([0.3504, 0.3874, 0.3110, 0.3358, 0.3576, 0.2302, 0.3253, 0.1544], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0264, 0.0237, 0.0275, 0.0260, 0.0230, 0.0259, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:08:49,841 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157914.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:56,871 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157924.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:09:11,705 INFO [finetune.py:976] (0/7) Epoch 28, batch 3300, loss[loss=0.1726, simple_loss=0.2508, pruned_loss=0.04721, over 4778.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.242, pruned_loss=0.04857, over 951879.92 frames. ], batch size: 54, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:09:20,545 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5929, 1.4881, 1.3538, 1.6943, 1.6493, 1.6522, 1.1024, 1.3689], + device='cuda:0'), covar=tensor([0.2306, 0.2107, 0.2099, 0.1779, 0.1707, 0.1308, 0.2592, 0.2038], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0245, 0.0190, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:09:29,137 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.521e+02 1.758e+02 2.140e+02 4.599e+02, threshold=3.515e+02, percent-clipped=1.0 +2023-03-27 10:09:44,716 INFO [finetune.py:976] (0/7) Epoch 28, batch 3350, loss[loss=0.1688, simple_loss=0.2426, pruned_loss=0.04755, over 4789.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2437, pruned_loss=0.04882, over 953370.42 frames. ], batch size: 29, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:09:46,064 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-158000.pt +2023-03-27 10:09:54,557 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3344, 2.2013, 1.9810, 2.2837, 2.1367, 2.0611, 2.1457, 2.8574], + device='cuda:0'), covar=tensor([0.3793, 0.4369, 0.3360, 0.3626, 0.4194, 0.2388, 0.3934, 0.1612], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0267, 0.0240, 0.0278, 0.0262, 0.0232, 0.0261, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:09:58,153 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158012.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:18,315 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7040, 1.6114, 1.3897, 1.6759, 2.1056, 2.1181, 1.7826, 1.5168], + device='cuda:0'), covar=tensor([0.0405, 0.0409, 0.0725, 0.0366, 0.0229, 0.0506, 0.0369, 0.0464], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0105, 0.0147, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:0'), out_proj_covar=tensor([7.7985e-05, 8.0566e-05, 1.1455e-04, 8.4496e-05, 7.8350e-05, 8.4861e-05, + 7.6498e-05, 8.5642e-05], device='cuda:0') +2023-03-27 10:10:29,909 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158038.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:38,995 INFO [finetune.py:976] (0/7) Epoch 28, batch 3400, loss[loss=0.1679, simple_loss=0.2523, pruned_loss=0.04172, over 4821.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2451, pruned_loss=0.04934, over 952446.39 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:10:46,888 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:46,913 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158060.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:10:55,546 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.550e+02 1.897e+02 2.350e+02 3.360e+02, threshold=3.793e+02, percent-clipped=0.0 +2023-03-27 10:11:01,964 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2186, 2.1905, 1.7574, 2.1593, 2.1300, 1.8865, 2.4664, 2.2653], + device='cuda:0'), covar=tensor([0.1210, 0.1915, 0.2631, 0.2388, 0.2323, 0.1564, 0.2889, 0.1468], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0189, 0.0236, 0.0253, 0.0250, 0.0206, 0.0213, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:11:04,972 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158086.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:11:12,585 INFO [finetune.py:976] (0/7) Epoch 28, batch 3450, loss[loss=0.1432, simple_loss=0.2106, pruned_loss=0.03792, over 4712.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2453, pruned_loss=0.04934, over 952670.06 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:11:18,606 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158108.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:11:26,804 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-27 10:11:28,432 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158122.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:11:30,867 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5763, 2.4251, 1.9851, 2.6452, 2.5384, 2.1791, 2.9624, 2.6178], + device='cuda:0'), covar=tensor([0.1251, 0.2181, 0.2894, 0.2499, 0.2415, 0.1669, 0.2749, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0189, 0.0236, 0.0253, 0.0250, 0.0207, 0.0214, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:11:46,008 INFO [finetune.py:976] (0/7) Epoch 28, batch 3500, loss[loss=0.1281, simple_loss=0.2037, pruned_loss=0.02624, over 4734.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2429, pruned_loss=0.04879, over 952990.67 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:11:48,559 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2191, 2.0366, 2.1902, 1.3783, 2.1041, 2.1992, 2.2649, 1.7929], + device='cuda:0'), covar=tensor([0.0566, 0.0698, 0.0633, 0.0882, 0.0785, 0.0705, 0.0581, 0.1202], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0119, 0.0130, 0.0140, 0.0140, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:11:54,572 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158162.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:11:58,011 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.0698, 3.5261, 3.6929, 3.7896, 3.8515, 3.6207, 4.1044, 1.7024], + device='cuda:0'), covar=tensor([0.0757, 0.0857, 0.0875, 0.0947, 0.1071, 0.1442, 0.0759, 0.4888], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0247, 0.0286, 0.0295, 0.0337, 0.0287, 0.0307, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:12:02,510 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.432e+02 1.800e+02 2.074e+02 3.411e+02, threshold=3.600e+02, percent-clipped=0.0 +2023-03-27 10:12:09,603 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158183.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:12:19,504 INFO [finetune.py:976] (0/7) Epoch 28, batch 3550, loss[loss=0.1587, simple_loss=0.2343, pruned_loss=0.0415, over 4794.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2403, pruned_loss=0.04804, over 954277.48 frames. ], batch size: 29, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:12:28,583 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158209.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:12:29,190 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158210.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:12:38,711 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:13:02,523 INFO [finetune.py:976] (0/7) Epoch 28, batch 3600, loss[loss=0.1647, simple_loss=0.2358, pruned_loss=0.04677, over 4831.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2378, pruned_loss=0.04756, over 955181.51 frames. ], batch size: 39, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:13:18,161 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:13:18,718 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.425e+02 1.679e+02 2.016e+02 3.584e+02, threshold=3.358e+02, percent-clipped=0.0 +2023-03-27 10:13:36,304 INFO [finetune.py:976] (0/7) Epoch 28, batch 3650, loss[loss=0.1816, simple_loss=0.2531, pruned_loss=0.05505, over 4927.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2392, pruned_loss=0.04763, over 955836.98 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:13:47,303 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0808, 2.0254, 1.6371, 1.7918, 2.0219, 1.7676, 2.2106, 2.0730], + device='cuda:0'), covar=tensor([0.1321, 0.1767, 0.2886, 0.2423, 0.2486, 0.1730, 0.2743, 0.1678], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0189, 0.0236, 0.0253, 0.0249, 0.0206, 0.0213, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:14:10,066 INFO [finetune.py:976] (0/7) Epoch 28, batch 3700, loss[loss=0.1947, simple_loss=0.2728, pruned_loss=0.05833, over 4887.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2425, pruned_loss=0.04826, over 956123.91 frames. ], batch size: 32, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:14:26,136 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.573e+02 1.954e+02 2.338e+02 5.991e+02, threshold=3.909e+02, percent-clipped=5.0 +2023-03-27 10:14:33,429 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:14:43,252 INFO [finetune.py:976] (0/7) Epoch 28, batch 3750, loss[loss=0.157, simple_loss=0.2292, pruned_loss=0.04241, over 4912.00 frames. ], tot_loss[loss=0.171, simple_loss=0.244, pruned_loss=0.04904, over 953695.26 frames. ], batch size: 42, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:14:48,359 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.0389, 1.2834, 1.3724, 1.3214, 1.4008, 2.4522, 1.2274, 1.4234], + device='cuda:0'), covar=tensor([0.1085, 0.1953, 0.1135, 0.0940, 0.1716, 0.0384, 0.1530, 0.1864], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 10:15:19,904 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:15:22,040 INFO [finetune.py:976] (0/7) Epoch 28, batch 3800, loss[loss=0.1859, simple_loss=0.2569, pruned_loss=0.0575, over 4920.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2455, pruned_loss=0.04984, over 954092.64 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:15:24,416 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 10:15:51,768 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.580e+02 1.909e+02 2.258e+02 3.504e+02, threshold=3.818e+02, percent-clipped=0.0 +2023-03-27 10:15:53,162 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1142, 1.9425, 1.7843, 2.0302, 1.8446, 1.9589, 1.8724, 2.6065], + device='cuda:0'), covar=tensor([0.3400, 0.4091, 0.3134, 0.3658, 0.4177, 0.2263, 0.3779, 0.1586], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0267, 0.0240, 0.0277, 0.0263, 0.0232, 0.0261, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:15:55,431 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158478.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:08,881 INFO [finetune.py:976] (0/7) Epoch 28, batch 3850, loss[loss=0.1573, simple_loss=0.2326, pruned_loss=0.04095, over 4757.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2443, pruned_loss=0.04921, over 955642.59 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:16:09,035 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2738, 2.2227, 1.9269, 2.2405, 2.0597, 2.1117, 2.1510, 2.7862], + device='cuda:0'), covar=tensor([0.3440, 0.4167, 0.3079, 0.3440, 0.3526, 0.2358, 0.3516, 0.1673], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0267, 0.0240, 0.0278, 0.0263, 0.0233, 0.0262, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:16:16,602 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:42,035 INFO [finetune.py:976] (0/7) Epoch 28, batch 3900, loss[loss=0.1719, simple_loss=0.2407, pruned_loss=0.05152, over 4817.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2411, pruned_loss=0.04811, over 955459.79 frames. ], batch size: 30, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:16:48,971 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158557.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:58,921 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.423e+02 1.682e+02 2.011e+02 3.434e+02, threshold=3.365e+02, percent-clipped=0.0 +2023-03-27 10:17:15,468 INFO [finetune.py:976] (0/7) Epoch 28, batch 3950, loss[loss=0.2007, simple_loss=0.2688, pruned_loss=0.06628, over 4820.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.238, pruned_loss=0.04719, over 956253.81 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:17:25,180 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.83 vs. limit=2.0 +2023-03-27 10:17:48,877 INFO [finetune.py:976] (0/7) Epoch 28, batch 4000, loss[loss=0.1476, simple_loss=0.2152, pruned_loss=0.04001, over 4728.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2382, pruned_loss=0.04761, over 957020.66 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:18:15,605 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.131e+01 1.444e+02 1.852e+02 2.094e+02 7.470e+02, threshold=3.703e+02, percent-clipped=2.0 +2023-03-27 10:18:32,166 INFO [finetune.py:976] (0/7) Epoch 28, batch 4050, loss[loss=0.1767, simple_loss=0.2565, pruned_loss=0.04842, over 4862.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2416, pruned_loss=0.04834, over 955509.00 frames. ], batch size: 31, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:18:47,514 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7117, 1.0667, 1.8095, 1.7894, 1.5713, 1.5203, 1.6619, 1.7307], + device='cuda:0'), covar=tensor([0.3311, 0.3592, 0.2774, 0.3204, 0.4107, 0.3387, 0.3723, 0.2587], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0247, 0.0267, 0.0296, 0.0295, 0.0272, 0.0301, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:18:59,957 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:05,203 INFO [finetune.py:976] (0/7) Epoch 28, batch 4100, loss[loss=0.1794, simple_loss=0.2459, pruned_loss=0.05643, over 4903.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2448, pruned_loss=0.04958, over 956055.67 frames. ], batch size: 37, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:19:10,081 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4563, 0.9913, 0.7813, 1.3221, 1.9206, 0.7846, 1.1656, 1.3346], + device='cuda:0'), covar=tensor([0.1627, 0.2410, 0.1793, 0.1305, 0.2010, 0.2043, 0.1660, 0.2084], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 10:19:15,943 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 10:19:22,719 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.534e+02 1.832e+02 2.269e+02 3.411e+02, threshold=3.665e+02, percent-clipped=0.0 +2023-03-27 10:19:25,289 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:25,877 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158778.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:29,377 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:38,838 INFO [finetune.py:976] (0/7) Epoch 28, batch 4150, loss[loss=0.1812, simple_loss=0.2567, pruned_loss=0.05287, over 4748.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2445, pruned_loss=0.04905, over 954832.78 frames. ], batch size: 26, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:19:58,145 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158826.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:03,637 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158835.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:20:05,953 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158838.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:09,569 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158844.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:11,899 INFO [finetune.py:976] (0/7) Epoch 28, batch 4200, loss[loss=0.1519, simple_loss=0.2208, pruned_loss=0.04156, over 4734.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2453, pruned_loss=0.04897, over 955143.02 frames. ], batch size: 54, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:20:35,368 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.763e+01 1.457e+02 1.627e+02 2.050e+02 3.601e+02, threshold=3.253e+02, percent-clipped=0.0 +2023-03-27 10:20:48,292 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 10:20:58,713 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158896.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:21:04,568 INFO [finetune.py:976] (0/7) Epoch 28, batch 4250, loss[loss=0.1922, simple_loss=0.2573, pruned_loss=0.06362, over 4819.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2428, pruned_loss=0.04873, over 954042.50 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:21:25,370 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 10:21:34,101 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 10:21:46,469 INFO [finetune.py:976] (0/7) Epoch 28, batch 4300, loss[loss=0.157, simple_loss=0.2293, pruned_loss=0.04232, over 4926.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2395, pruned_loss=0.0476, over 956102.29 frames. ], batch size: 37, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:22:03,614 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.473e+02 1.789e+02 2.045e+02 3.501e+02, threshold=3.577e+02, percent-clipped=2.0 +2023-03-27 10:22:20,212 INFO [finetune.py:976] (0/7) Epoch 28, batch 4350, loss[loss=0.1451, simple_loss=0.219, pruned_loss=0.0356, over 4817.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2367, pruned_loss=0.04677, over 956911.27 frames. ], batch size: 40, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:22:48,325 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:22:49,540 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4825, 1.3662, 1.3236, 1.3774, 1.0552, 2.8818, 1.0413, 1.4678], + device='cuda:0'), covar=tensor([0.3175, 0.2455, 0.2231, 0.2527, 0.1870, 0.0250, 0.2861, 0.1284], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0114, 0.0095, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 10:22:53,108 INFO [finetune.py:976] (0/7) Epoch 28, batch 4400, loss[loss=0.1764, simple_loss=0.2487, pruned_loss=0.05202, over 4831.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2386, pruned_loss=0.04782, over 957726.47 frames. ], batch size: 47, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:23:06,926 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:23:09,714 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.529e+02 1.726e+02 2.218e+02 4.795e+02, threshold=3.452e+02, percent-clipped=2.0 +2023-03-27 10:23:24,708 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:23:25,994 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.36 vs. limit=5.0 +2023-03-27 10:23:35,366 INFO [finetune.py:976] (0/7) Epoch 28, batch 4450, loss[loss=0.1404, simple_loss=0.2194, pruned_loss=0.0307, over 4737.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2426, pruned_loss=0.04863, over 958308.18 frames. ], batch size: 27, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:24:00,771 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:02,993 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159133.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:07,605 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159139.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:12,909 INFO [finetune.py:976] (0/7) Epoch 28, batch 4500, loss[loss=0.1949, simple_loss=0.2685, pruned_loss=0.0607, over 4921.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.244, pruned_loss=0.04943, over 957334.96 frames. ], batch size: 36, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:24:16,651 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.6332, 3.1705, 3.3780, 3.4990, 3.4622, 3.2306, 3.6975, 1.4662], + device='cuda:0'), covar=tensor([0.0971, 0.0929, 0.1061, 0.1106, 0.1226, 0.1446, 0.0906, 0.5397], + device='cuda:0'), in_proj_covar=tensor([0.0356, 0.0249, 0.0288, 0.0299, 0.0342, 0.0289, 0.0309, 0.0307], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:24:20,181 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159159.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:28,960 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.553e+02 1.848e+02 2.152e+02 3.966e+02, threshold=3.696e+02, percent-clipped=2.0 +2023-03-27 10:24:41,936 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159191.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:24:46,582 INFO [finetune.py:976] (0/7) Epoch 28, batch 4550, loss[loss=0.1722, simple_loss=0.2483, pruned_loss=0.048, over 4744.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2455, pruned_loss=0.04982, over 956654.78 frames. ], batch size: 59, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:24:46,652 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.9689, 4.6324, 4.4401, 2.5395, 4.8090, 3.5110, 0.8667, 3.2967], + device='cuda:0'), covar=tensor([0.2300, 0.1430, 0.1272, 0.2786, 0.0736, 0.0851, 0.4462, 0.1193], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0183, 0.0162, 0.0132, 0.0165, 0.0126, 0.0151, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 10:25:00,556 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159220.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:25:20,091 INFO [finetune.py:976] (0/7) Epoch 28, batch 4600, loss[loss=0.2228, simple_loss=0.2862, pruned_loss=0.07977, over 4256.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2451, pruned_loss=0.0497, over 953185.81 frames. ], batch size: 66, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:25:25,692 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7513, 1.2418, 0.8233, 1.5901, 2.2037, 1.0908, 1.5965, 1.5801], + device='cuda:0'), covar=tensor([0.1467, 0.2069, 0.1958, 0.1275, 0.1882, 0.2025, 0.1371, 0.2024], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0121, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 10:25:31,001 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8966, 1.1211, 1.9708, 1.9241, 1.7562, 1.6792, 1.7688, 1.9133], + device='cuda:0'), covar=tensor([0.3472, 0.3798, 0.3085, 0.3155, 0.4337, 0.3617, 0.4231, 0.3040], + device='cuda:0'), in_proj_covar=tensor([0.0268, 0.0248, 0.0269, 0.0298, 0.0297, 0.0274, 0.0303, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:25:35,682 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.454e+02 1.668e+02 2.062e+02 3.965e+02, threshold=3.336e+02, percent-clipped=3.0 +2023-03-27 10:25:54,150 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159290.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:26:05,013 INFO [finetune.py:976] (0/7) Epoch 28, batch 4650, loss[loss=0.1764, simple_loss=0.2363, pruned_loss=0.05828, over 4906.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2432, pruned_loss=0.04941, over 954665.11 frames. ], batch size: 36, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:26:55,598 INFO [finetune.py:976] (0/7) Epoch 28, batch 4700, loss[loss=0.1411, simple_loss=0.2175, pruned_loss=0.03233, over 4182.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2411, pruned_loss=0.04887, over 956014.18 frames. ], batch size: 18, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:26:58,035 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159351.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:11,661 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.356e+01 1.407e+02 1.592e+02 1.978e+02 3.098e+02, threshold=3.183e+02, percent-clipped=0.0 +2023-03-27 10:27:28,644 INFO [finetune.py:976] (0/7) Epoch 28, batch 4750, loss[loss=0.1837, simple_loss=0.2549, pruned_loss=0.05622, over 4925.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2394, pruned_loss=0.0486, over 956137.87 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:27:46,452 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:51,860 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159433.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:55,944 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159439.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:02,248 INFO [finetune.py:976] (0/7) Epoch 28, batch 4800, loss[loss=0.2145, simple_loss=0.301, pruned_loss=0.06402, over 4808.00 frames. ], tot_loss[loss=0.169, simple_loss=0.241, pruned_loss=0.04854, over 957241.76 frames. ], batch size: 45, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:28:18,791 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.565e+02 1.779e+02 2.195e+02 3.956e+02, threshold=3.558e+02, percent-clipped=1.0 +2023-03-27 10:28:23,700 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159481.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:26,061 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3657, 2.5082, 2.4447, 1.9202, 2.3217, 2.7316, 2.7528, 2.2557], + device='cuda:0'), covar=tensor([0.0655, 0.0649, 0.0741, 0.0771, 0.0799, 0.0694, 0.0582, 0.1093], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0139, 0.0142, 0.0120, 0.0130, 0.0141, 0.0142, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:28:27,808 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159487.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:30,761 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159491.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:28:36,823 INFO [finetune.py:976] (0/7) Epoch 28, batch 4850, loss[loss=0.1664, simple_loss=0.2647, pruned_loss=0.03403, over 4847.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04859, over 956144.89 frames. ], batch size: 44, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:28:57,878 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159515.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:29:00,947 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6588, 1.5555, 1.6298, 0.9760, 1.8032, 1.9582, 1.8663, 1.4632], + device='cuda:0'), covar=tensor([0.0997, 0.0757, 0.0557, 0.0581, 0.0575, 0.0719, 0.0403, 0.0830], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0130, 0.0122, 0.0131, 0.0129, 0.0142, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.8467e-05, 1.0569e-04, 9.2115e-05, 8.5596e-05, 9.2103e-05, 9.1563e-05, + 1.0098e-04, 1.0798e-04], device='cuda:0') +2023-03-27 10:29:17,405 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159539.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:29:20,468 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2153, 1.8447, 2.7171, 1.7334, 2.2370, 2.4469, 1.6676, 2.5140], + device='cuda:0'), covar=tensor([0.1464, 0.2375, 0.1587, 0.2040, 0.1030, 0.1511, 0.3059, 0.0991], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0205, 0.0192, 0.0188, 0.0174, 0.0213, 0.0217, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:29:23,228 INFO [finetune.py:976] (0/7) Epoch 28, batch 4900, loss[loss=0.1782, simple_loss=0.2704, pruned_loss=0.04294, over 4708.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2449, pruned_loss=0.04931, over 955236.36 frames. ], batch size: 54, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:29:25,676 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4328, 1.5342, 1.6057, 0.8880, 1.6823, 1.8649, 1.8757, 1.4615], + device='cuda:0'), covar=tensor([0.1081, 0.0707, 0.0558, 0.0610, 0.0518, 0.0632, 0.0340, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0148, 0.0130, 0.0123, 0.0132, 0.0130, 0.0143, 0.0152], + device='cuda:0'), out_proj_covar=tensor([8.8947e-05, 1.0621e-04, 9.2628e-05, 8.6059e-05, 9.2706e-05, 9.2173e-05, + 1.0161e-04, 1.0862e-04], device='cuda:0') +2023-03-27 10:29:40,329 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.553e+02 1.825e+02 2.271e+02 5.584e+02, threshold=3.651e+02, percent-clipped=3.0 +2023-03-27 10:29:56,962 INFO [finetune.py:976] (0/7) Epoch 28, batch 4950, loss[loss=0.1598, simple_loss=0.2332, pruned_loss=0.04317, over 4918.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2455, pruned_loss=0.04958, over 955566.77 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:30:18,778 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159631.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:30:28,807 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:30:29,930 INFO [finetune.py:976] (0/7) Epoch 28, batch 5000, loss[loss=0.1894, simple_loss=0.2469, pruned_loss=0.06595, over 4819.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2442, pruned_loss=0.04944, over 953740.63 frames. ], batch size: 25, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:30:47,438 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.490e+02 1.734e+02 1.957e+02 3.436e+02, threshold=3.469e+02, percent-clipped=0.0 +2023-03-27 10:30:48,753 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7747, 1.2662, 0.9306, 1.7442, 2.2351, 1.5613, 1.5339, 1.6491], + device='cuda:0'), covar=tensor([0.1451, 0.2085, 0.1912, 0.1161, 0.1872, 0.1856, 0.1502, 0.2010], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0093, 0.0109, 0.0093, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 10:30:59,460 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:31:05,619 INFO [finetune.py:976] (0/7) Epoch 28, batch 5050, loss[loss=0.1559, simple_loss=0.2193, pruned_loss=0.0463, over 4930.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2419, pruned_loss=0.04883, over 954323.25 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:31:32,248 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:31:55,163 INFO [finetune.py:976] (0/7) Epoch 28, batch 5100, loss[loss=0.1314, simple_loss=0.1951, pruned_loss=0.03385, over 4690.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.238, pruned_loss=0.04728, over 954231.51 frames. ], batch size: 23, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:32:18,633 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 10:32:21,761 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.553e+02 1.866e+02 2.180e+02 3.771e+02, threshold=3.731e+02, percent-clipped=1.0 +2023-03-27 10:32:21,835 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:32:37,623 INFO [finetune.py:976] (0/7) Epoch 28, batch 5150, loss[loss=0.1513, simple_loss=0.2365, pruned_loss=0.03309, over 4927.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2376, pruned_loss=0.04725, over 954504.31 frames. ], batch size: 38, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:32:45,146 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1217, 2.7004, 2.5485, 1.2272, 2.6866, 2.0865, 2.0820, 2.4417], + device='cuda:0'), covar=tensor([0.0992, 0.0927, 0.1788, 0.2298, 0.1826, 0.2467, 0.2349, 0.1347], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0190, 0.0201, 0.0181, 0.0210, 0.0210, 0.0224, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:32:49,242 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:11,651 INFO [finetune.py:976] (0/7) Epoch 28, batch 5200, loss[loss=0.1566, simple_loss=0.239, pruned_loss=0.03712, over 4829.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2425, pruned_loss=0.04876, over 955731.34 frames. ], batch size: 30, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:33:16,452 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159855.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:21,703 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159863.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:28,748 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.499e+02 1.871e+02 2.174e+02 4.313e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-27 10:33:44,908 INFO [finetune.py:976] (0/7) Epoch 28, batch 5250, loss[loss=0.1371, simple_loss=0.2107, pruned_loss=0.03171, over 4783.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.244, pruned_loss=0.04891, over 956614.67 frames. ], batch size: 29, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:33:53,295 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159910.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:59,361 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159916.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:06,924 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4037, 2.2943, 1.7546, 2.4059, 2.2443, 1.9478, 2.5851, 2.3894], + device='cuda:0'), covar=tensor([0.1291, 0.1938, 0.3060, 0.2437, 0.2558, 0.1663, 0.3093, 0.1588], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0253, 0.0250, 0.0208, 0.0215, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:34:26,200 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:27,321 INFO [finetune.py:976] (0/7) Epoch 28, batch 5300, loss[loss=0.1764, simple_loss=0.2516, pruned_loss=0.05057, over 4817.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2439, pruned_loss=0.04849, over 955185.94 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:34:50,614 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159971.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:52,304 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.498e+02 1.765e+02 2.107e+02 3.764e+02, threshold=3.530e+02, percent-clipped=1.0 +2023-03-27 10:35:01,756 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:35:06,012 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:35:08,412 INFO [finetune.py:976] (0/7) Epoch 28, batch 5350, loss[loss=0.1606, simple_loss=0.2203, pruned_loss=0.05039, over 4826.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2447, pruned_loss=0.04873, over 954596.29 frames. ], batch size: 30, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:35:09,764 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-160000.pt +2023-03-27 10:35:19,539 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 10:35:25,222 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6753, 1.6391, 1.9238, 1.2977, 1.7589, 1.9155, 1.5916, 2.1151], + device='cuda:0'), covar=tensor([0.1262, 0.2039, 0.1408, 0.1704, 0.0957, 0.1345, 0.2743, 0.0863], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0204, 0.0191, 0.0187, 0.0173, 0.0212, 0.0216, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:35:43,144 INFO [finetune.py:976] (0/7) Epoch 28, batch 5400, loss[loss=0.1365, simple_loss=0.2141, pruned_loss=0.02942, over 4758.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2421, pruned_loss=0.04823, over 955068.83 frames. ], batch size: 28, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:36:00,196 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.280e+01 1.494e+02 1.877e+02 2.215e+02 3.734e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-27 10:36:13,770 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0373, 1.9390, 1.7385, 2.0481, 2.3623, 2.0415, 1.9586, 1.7332], + device='cuda:0'), covar=tensor([0.1658, 0.1624, 0.1640, 0.1384, 0.1558, 0.1070, 0.2112, 0.1581], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0246, 0.0191, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:36:16,671 INFO [finetune.py:976] (0/7) Epoch 28, batch 5450, loss[loss=0.1447, simple_loss=0.2112, pruned_loss=0.03912, over 4742.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2388, pruned_loss=0.04702, over 952997.80 frames. ], batch size: 27, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:36:30,621 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4863, 2.7393, 2.6141, 2.0354, 2.7045, 2.9822, 3.1617, 2.4333], + device='cuda:0'), covar=tensor([0.0674, 0.0648, 0.0799, 0.0853, 0.0674, 0.0717, 0.0568, 0.1082], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0119, 0.0129, 0.0140, 0.0141, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:37:07,793 INFO [finetune.py:976] (0/7) Epoch 28, batch 5500, loss[loss=0.1522, simple_loss=0.2256, pruned_loss=0.03938, over 4751.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2354, pruned_loss=0.04578, over 951521.44 frames. ], batch size: 54, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:37:38,240 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.497e+02 1.808e+02 2.088e+02 3.346e+02, threshold=3.616e+02, percent-clipped=0.0 +2023-03-27 10:37:39,506 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-27 10:37:43,651 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1840, 2.0001, 1.7352, 1.8278, 1.9137, 1.8736, 1.9812, 2.6182], + device='cuda:0'), covar=tensor([0.3417, 0.3969, 0.3133, 0.3585, 0.3828, 0.2426, 0.3374, 0.1684], + device='cuda:0'), in_proj_covar=tensor([0.0287, 0.0264, 0.0237, 0.0275, 0.0260, 0.0231, 0.0258, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:37:55,667 INFO [finetune.py:976] (0/7) Epoch 28, batch 5550, loss[loss=0.1362, simple_loss=0.213, pruned_loss=0.02965, over 4795.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2383, pruned_loss=0.04714, over 950121.89 frames. ], batch size: 25, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:38:03,851 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160211.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:27,206 INFO [finetune.py:976] (0/7) Epoch 28, batch 5600, loss[loss=0.2062, simple_loss=0.2902, pruned_loss=0.06104, over 4905.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2418, pruned_loss=0.04806, over 950086.39 frames. ], batch size: 36, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:38:34,750 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.4893, 3.9282, 4.1464, 4.3415, 4.2540, 3.9495, 4.5869, 1.4314], + device='cuda:0'), covar=tensor([0.0808, 0.0966, 0.0936, 0.1107, 0.1328, 0.1861, 0.0678, 0.6207], + device='cuda:0'), in_proj_covar=tensor([0.0359, 0.0253, 0.0291, 0.0303, 0.0345, 0.0293, 0.0311, 0.0310], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:38:37,606 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160266.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:42,223 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.448e+02 1.891e+02 2.366e+02 4.690e+02, threshold=3.782e+02, percent-clipped=2.0 +2023-03-27 10:38:48,786 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 10:38:49,854 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:50,459 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2276, 2.0400, 1.6130, 0.6740, 1.8026, 1.8276, 1.7026, 1.8803], + device='cuda:0'), covar=tensor([0.0955, 0.0814, 0.1472, 0.2028, 0.1296, 0.2738, 0.2545, 0.0829], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0190, 0.0202, 0.0181, 0.0211, 0.0211, 0.0225, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:38:56,601 INFO [finetune.py:976] (0/7) Epoch 28, batch 5650, loss[loss=0.1815, simple_loss=0.2659, pruned_loss=0.04854, over 4903.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2446, pruned_loss=0.04817, over 951472.82 frames. ], batch size: 43, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:39:00,837 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2791, 1.3305, 1.3946, 0.8017, 1.5116, 1.6725, 1.6990, 1.2806], + device='cuda:0'), covar=tensor([0.1234, 0.0974, 0.0646, 0.0660, 0.0567, 0.0719, 0.0430, 0.1029], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0147, 0.0130, 0.0121, 0.0130, 0.0129, 0.0142, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.7462e-05, 1.0543e-04, 9.2133e-05, 8.5161e-05, 9.1295e-05, 9.1571e-05, + 1.0078e-04, 1.0729e-04], device='cuda:0') +2023-03-27 10:39:01,985 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160307.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:39:10,570 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 10:39:24,562 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:39:35,222 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2349, 1.7822, 2.2847, 2.2980, 2.0838, 2.0104, 2.1845, 2.1967], + device='cuda:0'), covar=tensor([0.3645, 0.3791, 0.3281, 0.3305, 0.4567, 0.3762, 0.4249, 0.2815], + device='cuda:0'), in_proj_covar=tensor([0.0269, 0.0248, 0.0269, 0.0298, 0.0298, 0.0275, 0.0303, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:39:36,255 INFO [finetune.py:976] (0/7) Epoch 28, batch 5700, loss[loss=0.1361, simple_loss=0.2035, pruned_loss=0.03434, over 4270.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2407, pruned_loss=0.0476, over 931363.03 frames. ], batch size: 18, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:39:48,023 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160368.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:39:53,237 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.932e+01 1.353e+02 1.695e+02 2.046e+02 5.031e+02, threshold=3.390e+02, percent-clipped=3.0 +2023-03-27 10:39:54,340 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-28.pt +2023-03-27 10:40:10,868 INFO [finetune.py:976] (0/7) Epoch 29, batch 0, loss[loss=0.1772, simple_loss=0.2476, pruned_loss=0.05345, over 4795.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2476, pruned_loss=0.05345, over 4795.00 frames. ], batch size: 51, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:40:10,869 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 10:40:21,879 INFO [finetune.py:1010] (0/7) Epoch 29, validation: loss=0.1588, simple_loss=0.2262, pruned_loss=0.04569, over 2265189.00 frames. +2023-03-27 10:40:21,880 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 10:40:24,835 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8371, 1.5794, 2.1515, 3.3831, 2.3481, 2.3817, 0.9556, 2.9256], + device='cuda:0'), covar=tensor([0.1661, 0.1445, 0.1404, 0.0485, 0.0792, 0.1426, 0.1998, 0.0471], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0136, 0.0125, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 10:40:57,961 INFO [finetune.py:976] (0/7) Epoch 29, batch 50, loss[loss=0.1857, simple_loss=0.2508, pruned_loss=0.06029, over 4820.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2437, pruned_loss=0.04948, over 216712.58 frames. ], batch size: 30, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:41:07,177 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 10:41:17,648 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.29 vs. limit=5.0 +2023-03-27 10:41:23,840 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-27 10:41:38,855 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.505e+02 1.852e+02 2.145e+02 8.823e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-27 10:41:39,935 INFO [finetune.py:976] (0/7) Epoch 29, batch 100, loss[loss=0.1783, simple_loss=0.2526, pruned_loss=0.05196, over 4928.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2397, pruned_loss=0.04832, over 380914.46 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:42:14,756 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:42:34,272 INFO [finetune.py:976] (0/7) Epoch 29, batch 150, loss[loss=0.1282, simple_loss=0.2033, pruned_loss=0.02654, over 4818.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2346, pruned_loss=0.0459, over 505846.02 frames. ], batch size: 38, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:42:43,896 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4675, 2.3312, 1.9159, 2.4874, 2.3848, 2.0923, 2.7402, 2.3835], + device='cuda:0'), covar=tensor([0.1339, 0.2085, 0.3013, 0.2380, 0.2571, 0.1783, 0.2649, 0.1834], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0191, 0.0238, 0.0254, 0.0251, 0.0209, 0.0217, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:42:55,942 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160559.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:43:00,754 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:43:06,534 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.446e+02 1.774e+02 2.135e+02 3.412e+02, threshold=3.547e+02, percent-clipped=0.0 +2023-03-27 10:43:07,133 INFO [finetune.py:976] (0/7) Epoch 29, batch 200, loss[loss=0.1668, simple_loss=0.2262, pruned_loss=0.0537, over 4038.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2327, pruned_loss=0.04554, over 605150.26 frames. ], batch size: 65, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:43:12,404 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1331, 1.8916, 2.5703, 4.1433, 2.9239, 2.7590, 0.6692, 3.5376], + device='cuda:0'), covar=tensor([0.1702, 0.1372, 0.1373, 0.0658, 0.0688, 0.1504, 0.2148, 0.0433], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0165, 0.0100, 0.0135, 0.0125, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 10:43:32,496 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160614.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:43:40,981 INFO [finetune.py:976] (0/7) Epoch 29, batch 250, loss[loss=0.148, simple_loss=0.2334, pruned_loss=0.03132, over 4863.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2355, pruned_loss=0.0452, over 682772.53 frames. ], batch size: 31, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:44:05,538 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160663.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:44:12,920 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.455e+02 1.868e+02 2.134e+02 3.558e+02, threshold=3.736e+02, percent-clipped=1.0 +2023-03-27 10:44:13,967 INFO [finetune.py:976] (0/7) Epoch 29, batch 300, loss[loss=0.1399, simple_loss=0.2209, pruned_loss=0.0294, over 4771.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2387, pruned_loss=0.04571, over 744201.79 frames. ], batch size: 28, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:44:26,383 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-27 10:44:36,465 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 10:44:57,148 INFO [finetune.py:976] (0/7) Epoch 29, batch 350, loss[loss=0.1706, simple_loss=0.2227, pruned_loss=0.0592, over 3465.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2414, pruned_loss=0.04694, over 789946.56 frames. ], batch size: 15, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:45:20,172 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9082, 1.8964, 2.0031, 1.3650, 1.9464, 2.0265, 2.0715, 1.6679], + device='cuda:0'), covar=tensor([0.0608, 0.0682, 0.0669, 0.0863, 0.0717, 0.0752, 0.0622, 0.1121], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0119, 0.0128, 0.0140, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:45:32,461 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 10:45:37,468 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.928e+01 1.581e+02 1.864e+02 2.210e+02 3.251e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-27 10:45:38,107 INFO [finetune.py:976] (0/7) Epoch 29, batch 400, loss[loss=0.1713, simple_loss=0.2495, pruned_loss=0.04657, over 4924.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2455, pruned_loss=0.04821, over 827670.96 frames. ], batch size: 38, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:45:59,569 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3306, 1.4361, 1.5565, 0.8217, 1.5453, 1.7819, 1.8267, 1.4266], + device='cuda:0'), covar=tensor([0.0978, 0.0643, 0.0510, 0.0560, 0.0437, 0.0573, 0.0329, 0.0708], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0148, 0.0130, 0.0122, 0.0131, 0.0130, 0.0142, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.7830e-05, 1.0579e-04, 9.2256e-05, 8.5358e-05, 9.1659e-05, 9.1751e-05, + 1.0126e-04, 1.0756e-04], device='cuda:0') +2023-03-27 10:46:05,503 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8644, 1.7409, 1.6941, 1.8999, 1.5215, 3.8908, 1.5727, 2.1650], + device='cuda:0'), covar=tensor([0.3130, 0.2342, 0.2051, 0.2237, 0.1591, 0.0175, 0.2479, 0.1134], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0093, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 10:46:11,857 INFO [finetune.py:976] (0/7) Epoch 29, batch 450, loss[loss=0.1449, simple_loss=0.2294, pruned_loss=0.03024, over 4774.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2451, pruned_loss=0.04867, over 855276.82 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:46:39,400 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.6658, 3.9760, 3.7767, 1.8725, 4.1602, 3.1269, 0.8764, 2.7482], + device='cuda:0'), covar=tensor([0.2308, 0.1505, 0.1481, 0.3095, 0.0840, 0.0853, 0.4060, 0.1388], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0180, 0.0160, 0.0130, 0.0163, 0.0124, 0.0150, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 10:46:55,074 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.473e+02 1.674e+02 2.082e+02 4.783e+02, threshold=3.348e+02, percent-clipped=2.0 +2023-03-27 10:46:55,687 INFO [finetune.py:976] (0/7) Epoch 29, batch 500, loss[loss=0.1802, simple_loss=0.2449, pruned_loss=0.05771, over 4911.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2432, pruned_loss=0.04824, over 879277.37 frames. ], batch size: 46, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:47:02,992 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 10:47:15,117 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160901.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:47:36,924 INFO [finetune.py:976] (0/7) Epoch 29, batch 550, loss[loss=0.1556, simple_loss=0.2259, pruned_loss=0.04264, over 4902.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2392, pruned_loss=0.04738, over 897815.09 frames. ], batch size: 43, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:48:08,206 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160962.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:48:08,820 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160963.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:48:10,048 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160965.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:48:11,242 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4103, 1.3316, 1.2205, 1.3917, 1.6874, 1.5630, 1.4316, 1.2340], + device='cuda:0'), covar=tensor([0.0369, 0.0339, 0.0673, 0.0315, 0.0229, 0.0537, 0.0363, 0.0478], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0106, 0.0148, 0.0111, 0.0102, 0.0117, 0.0105, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.8707e-05, 8.1178e-05, 1.1546e-04, 8.4904e-05, 7.8574e-05, 8.5927e-05, + 7.7588e-05, 8.6940e-05], device='cuda:0') +2023-03-27 10:48:15,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.000e+01 1.387e+02 1.769e+02 2.082e+02 3.377e+02, threshold=3.538e+02, percent-clipped=1.0 +2023-03-27 10:48:16,011 INFO [finetune.py:976] (0/7) Epoch 29, batch 600, loss[loss=0.1707, simple_loss=0.2554, pruned_loss=0.04296, over 4835.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2391, pruned_loss=0.04764, over 909821.11 frames. ], batch size: 47, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:48:24,053 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 10:48:41,019 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161011.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:48:41,638 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8029, 1.5997, 2.1151, 1.4161, 1.8216, 2.0082, 1.4886, 2.1881], + device='cuda:0'), covar=tensor([0.1293, 0.2125, 0.1288, 0.1830, 0.0994, 0.1359, 0.3074, 0.0889], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0206, 0.0193, 0.0190, 0.0174, 0.0214, 0.0219, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:48:44,046 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7762, 1.6862, 1.8400, 1.3382, 1.7253, 1.9400, 1.9330, 1.4706], + device='cuda:0'), covar=tensor([0.0555, 0.0662, 0.0619, 0.0852, 0.1053, 0.0599, 0.0519, 0.1232], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0119, 0.0128, 0.0141, 0.0140, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:48:49,433 INFO [finetune.py:976] (0/7) Epoch 29, batch 650, loss[loss=0.2192, simple_loss=0.2821, pruned_loss=0.07813, over 4826.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.243, pruned_loss=0.04855, over 919134.22 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:48:50,188 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:48:57,488 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 10:49:22,503 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.587e+02 1.904e+02 2.267e+02 4.706e+02, threshold=3.807e+02, percent-clipped=2.0 +2023-03-27 10:49:23,100 INFO [finetune.py:976] (0/7) Epoch 29, batch 700, loss[loss=0.1947, simple_loss=0.2731, pruned_loss=0.05818, over 4886.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2444, pruned_loss=0.04905, over 926890.78 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:50:03,475 INFO [finetune.py:976] (0/7) Epoch 29, batch 750, loss[loss=0.2071, simple_loss=0.2764, pruned_loss=0.06893, over 4901.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2463, pruned_loss=0.04999, over 933688.37 frames. ], batch size: 37, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:50:24,589 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 10:50:36,839 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-27 10:50:46,355 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.613e+01 1.645e+02 1.965e+02 2.346e+02 5.118e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-27 10:50:46,988 INFO [finetune.py:976] (0/7) Epoch 29, batch 800, loss[loss=0.166, simple_loss=0.2532, pruned_loss=0.03934, over 4815.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2454, pruned_loss=0.04946, over 937647.98 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:51:15,273 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:51:16,468 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6506, 1.1854, 0.8703, 1.6534, 2.0998, 1.3816, 1.3620, 1.6794], + device='cuda:0'), covar=tensor([0.1466, 0.1993, 0.1843, 0.1124, 0.1846, 0.1924, 0.1491, 0.1794], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 10:51:17,120 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9048, 1.2793, 1.9314, 1.8938, 1.7310, 1.6632, 1.8378, 1.8539], + device='cuda:0'), covar=tensor([0.3523, 0.3808, 0.3078, 0.3415, 0.4511, 0.3717, 0.4075, 0.2708], + device='cuda:0'), in_proj_covar=tensor([0.0269, 0.0249, 0.0269, 0.0298, 0.0297, 0.0275, 0.0304, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:51:20,578 INFO [finetune.py:976] (0/7) Epoch 29, batch 850, loss[loss=0.1173, simple_loss=0.1936, pruned_loss=0.02049, over 4721.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2423, pruned_loss=0.04845, over 941342.17 frames. ], batch size: 59, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:51:24,346 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161231.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:51:48,477 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161257.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:04,264 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.396e+02 1.676e+02 2.040e+02 3.102e+02, threshold=3.353e+02, percent-clipped=0.0 +2023-03-27 10:52:04,925 INFO [finetune.py:976] (0/7) Epoch 29, batch 900, loss[loss=0.1711, simple_loss=0.2334, pruned_loss=0.05437, over 4857.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2388, pruned_loss=0.04676, over 946605.59 frames. ], batch size: 47, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:52:06,234 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:11,093 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9857, 1.7575, 2.1934, 1.4222, 1.9767, 2.2760, 1.6481, 2.4457], + device='cuda:0'), covar=tensor([0.1321, 0.2060, 0.1646, 0.2058, 0.1018, 0.1309, 0.2919, 0.0791], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0204, 0.0192, 0.0189, 0.0173, 0.0213, 0.0217, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:52:15,349 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161292.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:35,799 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:38,179 INFO [finetune.py:976] (0/7) Epoch 29, batch 950, loss[loss=0.1614, simple_loss=0.2334, pruned_loss=0.0447, over 4886.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2368, pruned_loss=0.04624, over 948239.18 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:53:28,289 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.815e+01 1.511e+02 1.765e+02 2.170e+02 3.612e+02, threshold=3.529e+02, percent-clipped=1.0 +2023-03-27 10:53:28,920 INFO [finetune.py:976] (0/7) Epoch 29, batch 1000, loss[loss=0.2122, simple_loss=0.2874, pruned_loss=0.06846, over 4822.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2395, pruned_loss=0.0473, over 949022.81 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:53:48,923 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 10:53:49,428 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161401.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:06,939 INFO [finetune.py:976] (0/7) Epoch 29, batch 1050, loss[loss=0.21, simple_loss=0.2937, pruned_loss=0.06315, over 4844.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2417, pruned_loss=0.04789, over 948528.80 frames. ], batch size: 49, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:54:13,081 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5465, 1.4156, 1.3141, 1.3331, 1.8127, 1.7453, 1.6373, 1.3284], + device='cuda:0'), covar=tensor([0.0322, 0.0344, 0.0645, 0.0343, 0.0210, 0.0465, 0.0290, 0.0369], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0113, 0.0102, 0.0118, 0.0105, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.9509e-05, 8.1921e-05, 1.1587e-04, 8.5688e-05, 7.9166e-05, 8.6512e-05, + 7.8162e-05, 8.7466e-05], device='cuda:0') +2023-03-27 10:54:28,540 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:30,827 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:39,302 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.696e+01 1.443e+02 1.825e+02 2.108e+02 3.139e+02, threshold=3.650e+02, percent-clipped=0.0 +2023-03-27 10:54:39,918 INFO [finetune.py:976] (0/7) Epoch 29, batch 1100, loss[loss=0.2023, simple_loss=0.2799, pruned_loss=0.06231, over 4816.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2433, pruned_loss=0.04876, over 951276.41 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:55:11,594 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161521.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:14,357 INFO [finetune.py:976] (0/7) Epoch 29, batch 1150, loss[loss=0.1696, simple_loss=0.2485, pruned_loss=0.04538, over 4743.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2449, pruned_loss=0.04924, over 953297.70 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:55:15,164 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-27 10:55:33,332 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-27 10:55:39,773 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161557.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:50,719 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:51,850 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.769e+01 1.420e+02 1.779e+02 2.194e+02 3.095e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-03-27 10:55:52,951 INFO [finetune.py:976] (0/7) Epoch 29, batch 1200, loss[loss=0.1498, simple_loss=0.224, pruned_loss=0.03779, over 4706.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2427, pruned_loss=0.04861, over 951960.35 frames. ], batch size: 23, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:56:03,018 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161587.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:21,387 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 10:56:21,948 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 10:56:22,439 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:33,591 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161621.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:36,405 INFO [finetune.py:976] (0/7) Epoch 29, batch 1250, loss[loss=0.1627, simple_loss=0.2389, pruned_loss=0.04321, over 4825.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2402, pruned_loss=0.04758, over 952489.16 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:56:40,126 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 10:56:51,959 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8254, 1.2232, 1.8599, 1.8634, 1.6590, 1.6179, 1.7759, 1.8046], + device='cuda:0'), covar=tensor([0.3728, 0.4204, 0.3286, 0.3434, 0.4669, 0.3703, 0.4308, 0.3031], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0251, 0.0271, 0.0300, 0.0300, 0.0277, 0.0307, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:57:07,718 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161669.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:57:07,786 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4394, 2.3111, 1.7922, 0.8776, 1.9677, 1.9225, 1.7617, 2.1410], + device='cuda:0'), covar=tensor([0.0885, 0.0642, 0.1569, 0.1904, 0.1301, 0.2399, 0.2257, 0.0771], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0188, 0.0200, 0.0179, 0.0208, 0.0209, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 10:57:07,797 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.2649, 1.4351, 1.5292, 0.8046, 1.5523, 1.7862, 1.7950, 1.4046], + device='cuda:0'), covar=tensor([0.0977, 0.0664, 0.0467, 0.0510, 0.0445, 0.0511, 0.0292, 0.0717], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0148, 0.0131, 0.0122, 0.0132, 0.0130, 0.0143, 0.0152], + device='cuda:0'), out_proj_covar=tensor([8.8020e-05, 1.0589e-04, 9.3013e-05, 8.5697e-05, 9.2468e-05, 9.2013e-05, + 1.0156e-04, 1.0853e-04], device='cuda:0') +2023-03-27 10:57:10,127 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8162, 1.7969, 1.6638, 1.6224, 2.3009, 2.3770, 2.0765, 1.8967], + device='cuda:0'), covar=tensor([0.0469, 0.0413, 0.0700, 0.0385, 0.0289, 0.0529, 0.0342, 0.0417], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0112, 0.0102, 0.0118, 0.0105, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.9511e-05, 8.1784e-05, 1.1567e-04, 8.5501e-05, 7.9208e-05, 8.6742e-05, + 7.8247e-05, 8.7668e-05], device='cuda:0') +2023-03-27 10:57:11,726 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.453e+02 1.676e+02 2.058e+02 3.499e+02, threshold=3.351e+02, percent-clipped=0.0 +2023-03-27 10:57:12,869 INFO [finetune.py:976] (0/7) Epoch 29, batch 1300, loss[loss=0.1348, simple_loss=0.2083, pruned_loss=0.0306, over 4771.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2377, pruned_loss=0.04664, over 954581.27 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:57:47,270 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 10:57:54,739 INFO [finetune.py:976] (0/7) Epoch 29, batch 1350, loss[loss=0.1758, simple_loss=0.2539, pruned_loss=0.04891, over 4911.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2387, pruned_loss=0.04699, over 954988.16 frames. ], batch size: 36, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:58:29,756 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161757.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:58:49,497 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.561e+02 1.829e+02 2.338e+02 6.236e+02, threshold=3.658e+02, percent-clipped=4.0 +2023-03-27 10:58:50,571 INFO [finetune.py:976] (0/7) Epoch 29, batch 1400, loss[loss=0.1509, simple_loss=0.2316, pruned_loss=0.03511, over 4763.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2424, pruned_loss=0.04838, over 956835.16 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:17,539 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:18,085 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161816.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:23,989 INFO [finetune.py:976] (0/7) Epoch 29, batch 1450, loss[loss=0.1704, simple_loss=0.2452, pruned_loss=0.04783, over 4857.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2443, pruned_loss=0.04886, over 957404.15 frames. ], batch size: 31, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:33,512 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-03-27 10:59:55,400 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:56,503 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.509e+02 1.810e+02 2.171e+02 4.074e+02, threshold=3.620e+02, percent-clipped=1.0 +2023-03-27 10:59:57,115 INFO [finetune.py:976] (0/7) Epoch 29, batch 1500, loss[loss=0.1886, simple_loss=0.2576, pruned_loss=0.05976, over 4750.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2448, pruned_loss=0.04909, over 956892.39 frames. ], batch size: 59, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:58,326 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161876.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:05,888 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:27,811 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161920.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:32,887 INFO [finetune.py:976] (0/7) Epoch 29, batch 1550, loss[loss=0.1736, simple_loss=0.2421, pruned_loss=0.05254, over 4736.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2439, pruned_loss=0.04776, over 956469.86 frames. ], batch size: 23, lr: 2.85e-03, grad_scale: 32.0 +2023-03-27 11:00:44,439 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:01:16,695 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.413e+02 1.667e+02 1.984e+02 3.261e+02, threshold=3.334e+02, percent-clipped=0.0 +2023-03-27 11:01:17,325 INFO [finetune.py:976] (0/7) Epoch 29, batch 1600, loss[loss=0.1842, simple_loss=0.2562, pruned_loss=0.05616, over 4802.00 frames. ], tot_loss[loss=0.169, simple_loss=0.243, pruned_loss=0.04752, over 957470.30 frames. ], batch size: 51, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:01:43,126 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-162000.pt +2023-03-27 11:01:53,672 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 11:01:57,786 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5767, 2.4950, 2.0246, 2.7606, 2.5181, 2.2253, 3.0570, 2.6182], + device='cuda:0'), covar=tensor([0.1220, 0.2067, 0.2861, 0.2266, 0.2378, 0.1525, 0.2577, 0.1548], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0252, 0.0249, 0.0208, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:01:59,455 INFO [finetune.py:976] (0/7) Epoch 29, batch 1650, loss[loss=0.1309, simple_loss=0.2008, pruned_loss=0.03048, over 4826.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2397, pruned_loss=0.04681, over 958611.64 frames. ], batch size: 39, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:02:22,365 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162057.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:02:34,947 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.499e+01 1.468e+02 1.758e+02 2.171e+02 4.899e+02, threshold=3.516e+02, percent-clipped=3.0 +2023-03-27 11:02:35,583 INFO [finetune.py:976] (0/7) Epoch 29, batch 1700, loss[loss=0.1959, simple_loss=0.2731, pruned_loss=0.05933, over 4822.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2378, pruned_loss=0.04647, over 957764.79 frames. ], batch size: 40, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:02:36,905 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:01,317 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:08,004 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162116.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:09,239 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7322, 1.7455, 1.6042, 1.7289, 1.3009, 4.1303, 1.7370, 2.0127], + device='cuda:0'), covar=tensor([0.3189, 0.2424, 0.2106, 0.2256, 0.1688, 0.0142, 0.2512, 0.1208], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0093, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 11:03:16,125 INFO [finetune.py:976] (0/7) Epoch 29, batch 1750, loss[loss=0.1769, simple_loss=0.2624, pruned_loss=0.04573, over 4807.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2404, pruned_loss=0.04775, over 956267.95 frames. ], batch size: 51, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:03:18,004 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9073, 1.3559, 0.9278, 1.6998, 2.2056, 1.4032, 1.5362, 1.6142], + device='cuda:0'), covar=tensor([0.1360, 0.1950, 0.1709, 0.1098, 0.1750, 0.1838, 0.1344, 0.1998], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0093, 0.0108, 0.0093, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:03:24,591 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162138.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:59,946 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:04:04,278 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:04:10,043 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.681e+01 1.626e+02 1.856e+02 2.284e+02 4.131e+02, threshold=3.712e+02, percent-clipped=2.0 +2023-03-27 11:04:10,644 INFO [finetune.py:976] (0/7) Epoch 29, batch 1800, loss[loss=0.1979, simple_loss=0.2649, pruned_loss=0.06549, over 4857.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.243, pruned_loss=0.04826, over 957183.54 frames. ], batch size: 49, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:04:16,231 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4216, 1.3827, 1.8498, 1.6724, 1.6260, 3.4392, 1.4837, 1.6017], + device='cuda:0'), covar=tensor([0.1050, 0.1924, 0.1169, 0.0996, 0.1626, 0.0219, 0.1464, 0.1817], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:04:30,452 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3341, 1.6016, 0.8963, 2.0488, 2.5448, 1.7209, 1.8155, 1.9106], + device='cuda:0'), covar=tensor([0.1281, 0.1939, 0.1887, 0.1105, 0.1653, 0.1743, 0.1343, 0.2012], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0093, 0.0108, 0.0093, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:04:40,283 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.2278, 2.6261, 2.4798, 1.2699, 2.9236, 2.1522, 1.8253, 2.4233], + device='cuda:0'), covar=tensor([0.0940, 0.1287, 0.2547, 0.2926, 0.1615, 0.2676, 0.3226, 0.1428], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0188, 0.0201, 0.0180, 0.0208, 0.0209, 0.0222, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:04:44,452 INFO [finetune.py:976] (0/7) Epoch 29, batch 1850, loss[loss=0.163, simple_loss=0.244, pruned_loss=0.04105, over 4745.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2445, pruned_loss=0.04898, over 957211.53 frames. ], batch size: 54, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:04:48,175 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162231.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:05:17,383 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.604e+02 1.802e+02 2.267e+02 3.875e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-27 11:05:17,692 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 11:05:17,994 INFO [finetune.py:976] (0/7) Epoch 29, batch 1900, loss[loss=0.1455, simple_loss=0.228, pruned_loss=0.0315, over 4791.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2457, pruned_loss=0.04955, over 956429.72 frames. ], batch size: 25, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:05:28,431 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162292.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:05:39,689 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 11:05:51,672 INFO [finetune.py:976] (0/7) Epoch 29, batch 1950, loss[loss=0.1489, simple_loss=0.2148, pruned_loss=0.04149, over 4711.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2425, pruned_loss=0.04794, over 956081.24 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:06:28,897 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8173, 2.9020, 2.7680, 1.8742, 2.7464, 3.1788, 3.1621, 2.5150], + device='cuda:0'), covar=tensor([0.0505, 0.0515, 0.0599, 0.0797, 0.0709, 0.0548, 0.0478, 0.0935], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0136, 0.0138, 0.0117, 0.0126, 0.0138, 0.0138, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:06:36,624 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.724e+01 1.430e+02 1.855e+02 2.130e+02 3.474e+02, threshold=3.709e+02, percent-clipped=0.0 +2023-03-27 11:06:37,251 INFO [finetune.py:976] (0/7) Epoch 29, batch 2000, loss[loss=0.1602, simple_loss=0.2236, pruned_loss=0.04835, over 4893.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2401, pruned_loss=0.04768, over 954645.61 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:07:03,744 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-27 11:07:14,705 INFO [finetune.py:976] (0/7) Epoch 29, batch 2050, loss[loss=0.1338, simple_loss=0.2012, pruned_loss=0.03325, over 4814.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.237, pruned_loss=0.04637, over 954688.46 frames. ], batch size: 25, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:07:19,649 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162433.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:07:45,861 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:07:47,599 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.116e+01 1.601e+02 1.964e+02 2.330e+02 4.874e+02, threshold=3.928e+02, percent-clipped=3.0 +2023-03-27 11:07:48,225 INFO [finetune.py:976] (0/7) Epoch 29, batch 2100, loss[loss=0.1661, simple_loss=0.2498, pruned_loss=0.04122, over 4787.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2381, pruned_loss=0.04714, over 953540.86 frames. ], batch size: 25, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:08:22,562 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:08:26,827 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.50 vs. limit=2.0 +2023-03-27 11:08:28,423 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162519.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:08:34,950 INFO [finetune.py:976] (0/7) Epoch 29, batch 2150, loss[loss=0.1382, simple_loss=0.2132, pruned_loss=0.03156, over 4761.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2401, pruned_loss=0.04738, over 954656.32 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:08:36,155 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5595, 1.3641, 1.9375, 1.8489, 1.5179, 3.4224, 1.3860, 1.6082], + device='cuda:0'), covar=tensor([0.0987, 0.1868, 0.1067, 0.0928, 0.1648, 0.0211, 0.1486, 0.1828], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:09:14,960 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162572.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 11:09:18,163 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.475e+02 1.779e+02 2.116e+02 3.329e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-03-27 11:09:18,787 INFO [finetune.py:976] (0/7) Epoch 29, batch 2200, loss[loss=0.15, simple_loss=0.2215, pruned_loss=0.03926, over 4797.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04856, over 955493.47 frames. ], batch size: 25, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:09:27,181 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162587.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:09:29,107 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 11:10:00,795 INFO [finetune.py:976] (0/7) Epoch 29, batch 2250, loss[loss=0.1885, simple_loss=0.258, pruned_loss=0.05948, over 4826.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2456, pruned_loss=0.04911, over 955931.72 frames. ], batch size: 47, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:10:33,519 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.222e+01 1.558e+02 1.861e+02 2.054e+02 5.864e+02, threshold=3.723e+02, percent-clipped=1.0 +2023-03-27 11:10:34,114 INFO [finetune.py:976] (0/7) Epoch 29, batch 2300, loss[loss=0.1742, simple_loss=0.2376, pruned_loss=0.0554, over 4700.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2453, pruned_loss=0.04873, over 954616.82 frames. ], batch size: 59, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:10:36,996 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4960, 1.2998, 1.3614, 1.4091, 1.6739, 1.6229, 1.4172, 1.2798], + device='cuda:0'), covar=tensor([0.0353, 0.0322, 0.0638, 0.0329, 0.0243, 0.0470, 0.0364, 0.0437], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0106, 0.0149, 0.0112, 0.0102, 0.0117, 0.0105, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.9099e-05, 8.1069e-05, 1.1560e-04, 8.4941e-05, 7.8937e-05, 8.6108e-05, + 7.7721e-05, 8.7255e-05], device='cuda:0') +2023-03-27 11:10:42,248 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5807, 1.4761, 1.4488, 1.5182, 1.1336, 2.9546, 1.1454, 1.4336], + device='cuda:0'), covar=tensor([0.3375, 0.2656, 0.2189, 0.2530, 0.1770, 0.0284, 0.2736, 0.1392], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0112, 0.0095, 0.0093, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 11:10:48,265 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8929, 1.8046, 1.8374, 1.1378, 1.8651, 1.9111, 1.8983, 1.5699], + device='cuda:0'), covar=tensor([0.0556, 0.0660, 0.0625, 0.0927, 0.0767, 0.0641, 0.0584, 0.1175], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0140, 0.0119, 0.0128, 0.0140, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:11:07,975 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 11:11:09,593 INFO [finetune.py:976] (0/7) Epoch 29, batch 2350, loss[loss=0.1351, simple_loss=0.2096, pruned_loss=0.03029, over 4770.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2428, pruned_loss=0.04805, over 955429.87 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:11:09,860 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 11:11:20,046 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162733.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:11:39,985 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.5443, 1.4100, 1.4092, 1.4121, 1.7808, 1.7259, 1.5566, 1.3268], + device='cuda:0'), covar=tensor([0.0340, 0.0312, 0.0602, 0.0315, 0.0212, 0.0470, 0.0338, 0.0447], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0106, 0.0149, 0.0112, 0.0102, 0.0117, 0.0104, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.9269e-05, 8.1218e-05, 1.1563e-04, 8.5071e-05, 7.8985e-05, 8.6458e-05, + 7.7634e-05, 8.7232e-05], device='cuda:0') +2023-03-27 11:11:50,493 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.467e+01 1.478e+02 1.866e+02 2.325e+02 4.091e+02, threshold=3.732e+02, percent-clipped=2.0 +2023-03-27 11:11:51,107 INFO [finetune.py:976] (0/7) Epoch 29, batch 2400, loss[loss=0.1582, simple_loss=0.2277, pruned_loss=0.04432, over 4905.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2396, pruned_loss=0.04705, over 957145.19 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:11:56,101 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0771, 1.8768, 1.6135, 1.6091, 1.7804, 1.7587, 1.7824, 2.5153], + device='cuda:0'), covar=tensor([0.3109, 0.3141, 0.2942, 0.3228, 0.3657, 0.2169, 0.3248, 0.1407], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0264, 0.0239, 0.0275, 0.0262, 0.0232, 0.0260, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:11:56,643 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:11:59,014 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4354, 2.2778, 1.9854, 0.9944, 2.1459, 1.9321, 1.7135, 2.1993], + device='cuda:0'), covar=tensor([0.1033, 0.0756, 0.1725, 0.2097, 0.1276, 0.2368, 0.2245, 0.0973], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0188, 0.0201, 0.0180, 0.0209, 0.0209, 0.0223, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:12:33,619 INFO [finetune.py:976] (0/7) Epoch 29, batch 2450, loss[loss=0.1456, simple_loss=0.2313, pruned_loss=0.02993, over 4817.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2371, pruned_loss=0.0462, over 957015.60 frames. ], batch size: 45, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:12:35,497 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:13:02,268 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162867.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:13:09,317 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.738e+01 1.481e+02 1.739e+02 2.016e+02 4.521e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 11:13:09,952 INFO [finetune.py:976] (0/7) Epoch 29, batch 2500, loss[loss=0.1675, simple_loss=0.2497, pruned_loss=0.04267, over 4930.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.239, pruned_loss=0.04715, over 958018.96 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:13:23,117 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:13:28,093 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162888.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:13:52,413 INFO [finetune.py:976] (0/7) Epoch 29, batch 2550, loss[loss=0.1842, simple_loss=0.2498, pruned_loss=0.0593, over 4842.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2418, pruned_loss=0.04794, over 954441.52 frames. ], batch size: 49, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:13:57,994 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6138, 1.4982, 1.9037, 1.9156, 1.7738, 2.9638, 1.5465, 1.6561], + device='cuda:0'), covar=tensor([0.0919, 0.1595, 0.1363, 0.0775, 0.1348, 0.0298, 0.1251, 0.1509], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:14:01,594 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:14:34,254 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0318, 2.0099, 1.4845, 2.0442, 2.0308, 1.7077, 2.6478, 2.0319], + device='cuda:0'), covar=tensor([0.1449, 0.2026, 0.3337, 0.2787, 0.2648, 0.1798, 0.2438, 0.1908], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0191, 0.0236, 0.0253, 0.0250, 0.0209, 0.0216, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:14:36,961 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.986e+01 1.516e+02 1.791e+02 2.248e+02 3.208e+02, threshold=3.582e+02, percent-clipped=0.0 +2023-03-27 11:14:37,596 INFO [finetune.py:976] (0/7) Epoch 29, batch 2600, loss[loss=0.1593, simple_loss=0.2365, pruned_loss=0.04105, over 4733.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.244, pruned_loss=0.04857, over 954360.10 frames. ], batch size: 54, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:15:17,741 INFO [finetune.py:976] (0/7) Epoch 29, batch 2650, loss[loss=0.1535, simple_loss=0.2389, pruned_loss=0.03403, over 4761.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.04818, over 952583.50 frames. ], batch size: 51, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:15:36,368 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-03-27 11:15:51,093 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.228e+01 1.484e+02 1.662e+02 2.061e+02 3.704e+02, threshold=3.324e+02, percent-clipped=1.0 +2023-03-27 11:15:51,717 INFO [finetune.py:976] (0/7) Epoch 29, batch 2700, loss[loss=0.1837, simple_loss=0.2416, pruned_loss=0.06292, over 4760.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2421, pruned_loss=0.04766, over 955245.77 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:16:02,231 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:16:25,140 INFO [finetune.py:976] (0/7) Epoch 29, batch 2750, loss[loss=0.1419, simple_loss=0.2191, pruned_loss=0.0324, over 4904.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2387, pruned_loss=0.04676, over 955326.15 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:16:48,575 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8842, 1.5995, 2.0786, 1.3969, 1.8007, 2.0432, 1.5324, 2.2165], + device='cuda:0'), covar=tensor([0.1088, 0.1908, 0.1208, 0.1679, 0.0864, 0.1237, 0.2555, 0.0749], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0208, 0.0195, 0.0191, 0.0176, 0.0214, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:16:52,301 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163151.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:16:52,355 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 11:17:03,860 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163167.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:17:10,462 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.483e+02 1.714e+02 1.995e+02 3.279e+02, threshold=3.429e+02, percent-clipped=0.0 +2023-03-27 11:17:10,478 INFO [finetune.py:976] (0/7) Epoch 29, batch 2800, loss[loss=0.1657, simple_loss=0.2335, pruned_loss=0.04898, over 4894.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2355, pruned_loss=0.04571, over 950879.17 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:17:15,462 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163183.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:37,949 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163215.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:44,450 INFO [finetune.py:976] (0/7) Epoch 29, batch 2850, loss[loss=0.2147, simple_loss=0.2819, pruned_loss=0.07374, over 4815.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2351, pruned_loss=0.0455, over 948992.67 frames. ], batch size: 40, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:17:59,028 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163240.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:59,814 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-27 11:18:05,214 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-27 11:18:23,894 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-27 11:18:31,673 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.921e+01 1.528e+02 1.890e+02 2.266e+02 3.566e+02, threshold=3.779e+02, percent-clipped=1.0 +2023-03-27 11:18:31,689 INFO [finetune.py:976] (0/7) Epoch 29, batch 2900, loss[loss=0.1719, simple_loss=0.2473, pruned_loss=0.04824, over 4867.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2369, pruned_loss=0.04653, over 949330.79 frames. ], batch size: 34, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:18:52,056 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163301.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:19:08,253 INFO [finetune.py:976] (0/7) Epoch 29, batch 2950, loss[loss=0.1735, simple_loss=0.2497, pruned_loss=0.04863, over 4697.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2405, pruned_loss=0.0477, over 947276.55 frames. ], batch size: 59, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:19:49,295 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.530e+02 1.927e+02 2.309e+02 4.929e+02, threshold=3.855e+02, percent-clipped=3.0 +2023-03-27 11:19:49,311 INFO [finetune.py:976] (0/7) Epoch 29, batch 3000, loss[loss=0.1826, simple_loss=0.2639, pruned_loss=0.05068, over 4834.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2434, pruned_loss=0.04846, over 949009.90 frames. ], batch size: 49, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:19:49,312 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 11:19:52,498 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1632, 2.0366, 1.6121, 0.6634, 1.8661, 1.8718, 1.7030, 1.9451], + device='cuda:0'), covar=tensor([0.1045, 0.0764, 0.1560, 0.1922, 0.1294, 0.2528, 0.2296, 0.0900], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0188, 0.0201, 0.0181, 0.0209, 0.0210, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:19:53,092 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3675, 1.4670, 1.7134, 1.6181, 1.6581, 3.0113, 1.5417, 1.6159], + device='cuda:0'), covar=tensor([0.0949, 0.1823, 0.0952, 0.0862, 0.1479, 0.0260, 0.1352, 0.1696], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0077, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:20:02,086 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1284, 1.2022, 2.1985, 2.0354, 1.9610, 1.8548, 1.9285, 2.0822], + device='cuda:0'), covar=tensor([0.3713, 0.3774, 0.3545, 0.3541, 0.4547, 0.3546, 0.4355, 0.2993], + device='cuda:0'), in_proj_covar=tensor([0.0269, 0.0249, 0.0270, 0.0300, 0.0298, 0.0276, 0.0304, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:20:05,062 INFO [finetune.py:1010] (0/7) Epoch 29, validation: loss=0.158, simple_loss=0.2251, pruned_loss=0.04545, over 2265189.00 frames. +2023-03-27 11:20:05,062 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 11:20:28,529 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1310, 2.0008, 1.7495, 2.2185, 2.5681, 2.2159, 1.8912, 1.6290], + device='cuda:0'), covar=tensor([0.1990, 0.1824, 0.1825, 0.1497, 0.1542, 0.1079, 0.2028, 0.1830], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0211, 0.0215, 0.0198, 0.0244, 0.0190, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:20:42,541 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7449, 1.1701, 0.9001, 1.5546, 2.1380, 1.4577, 1.5377, 1.6191], + device='cuda:0'), covar=tensor([0.1550, 0.2126, 0.1867, 0.1249, 0.1925, 0.1876, 0.1370, 0.1920], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0093, 0.0108, 0.0093, 0.0120, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 11:20:43,033 INFO [finetune.py:976] (0/7) Epoch 29, batch 3050, loss[loss=0.176, simple_loss=0.2602, pruned_loss=0.04595, over 4897.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2436, pruned_loss=0.04806, over 951423.67 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:20:57,918 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:21:07,489 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3192, 1.6737, 2.6530, 1.7842, 2.2587, 2.5717, 1.6973, 2.5766], + device='cuda:0'), covar=tensor([0.1117, 0.2465, 0.1094, 0.1781, 0.0938, 0.1115, 0.2877, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0175, 0.0213, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:21:16,336 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.385e+02 1.622e+02 1.937e+02 3.745e+02, threshold=3.244e+02, percent-clipped=0.0 +2023-03-27 11:21:16,352 INFO [finetune.py:976] (0/7) Epoch 29, batch 3100, loss[loss=0.1647, simple_loss=0.2356, pruned_loss=0.04687, over 4895.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2418, pruned_loss=0.04725, over 954183.21 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:21:20,794 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 11:21:22,367 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:21:51,332 INFO [finetune.py:976] (0/7) Epoch 29, batch 3150, loss[loss=0.1335, simple_loss=0.2148, pruned_loss=0.02613, over 4785.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2387, pruned_loss=0.04627, over 954780.14 frames. ], batch size: 29, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:21:55,027 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163531.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:22:19,158 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8464, 2.5732, 2.2009, 2.9032, 2.7105, 2.4239, 3.1159, 2.7870], + device='cuda:0'), covar=tensor([0.1223, 0.2097, 0.2961, 0.2453, 0.2629, 0.1663, 0.3461, 0.1637], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0189, 0.0235, 0.0252, 0.0248, 0.0207, 0.0215, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:22:25,913 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-27 11:22:31,222 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3324, 1.3218, 1.5049, 1.0750, 1.3564, 1.4932, 1.3262, 1.6842], + device='cuda:0'), covar=tensor([0.1163, 0.2179, 0.1207, 0.1535, 0.0860, 0.1175, 0.2804, 0.0767], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0175, 0.0213, 0.0219, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:22:33,519 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.497e+02 1.704e+02 2.227e+02 4.213e+02, threshold=3.407e+02, percent-clipped=5.0 +2023-03-27 11:22:33,535 INFO [finetune.py:976] (0/7) Epoch 29, batch 3200, loss[loss=0.1864, simple_loss=0.2524, pruned_loss=0.06024, over 4901.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2357, pruned_loss=0.04581, over 955532.93 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:22:49,213 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163596.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:23:07,473 INFO [finetune.py:976] (0/7) Epoch 29, batch 3250, loss[loss=0.2442, simple_loss=0.3065, pruned_loss=0.09094, over 4899.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2379, pruned_loss=0.04714, over 954126.93 frames. ], batch size: 43, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:23:49,082 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6826, 1.4718, 2.3172, 3.3572, 2.2121, 2.3653, 1.4371, 2.8660], + device='cuda:0'), covar=tensor([0.1719, 0.1447, 0.1100, 0.0562, 0.0815, 0.1625, 0.1516, 0.0492], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0165, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:23:52,619 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.629e+02 1.981e+02 2.506e+02 4.570e+02, threshold=3.962e+02, percent-clipped=6.0 +2023-03-27 11:23:52,635 INFO [finetune.py:976] (0/7) Epoch 29, batch 3300, loss[loss=0.1973, simple_loss=0.2792, pruned_loss=0.05774, over 4850.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2419, pruned_loss=0.04838, over 956624.14 frames. ], batch size: 44, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:23:55,264 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 11:24:11,016 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3047, 1.2635, 1.4772, 0.9904, 1.3033, 1.3749, 1.2746, 1.5599], + device='cuda:0'), covar=tensor([0.1304, 0.2572, 0.1543, 0.1772, 0.1082, 0.1514, 0.3537, 0.0999], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0208, 0.0195, 0.0191, 0.0176, 0.0214, 0.0220, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:24:31,730 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.6840, 1.5512, 1.5446, 0.9915, 1.7694, 1.9765, 1.8412, 1.4718], + device='cuda:0'), covar=tensor([0.0955, 0.0768, 0.0644, 0.0612, 0.0485, 0.0648, 0.0425, 0.0712], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0147, 0.0130, 0.0121, 0.0130, 0.0129, 0.0141, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.7352e-05, 1.0525e-04, 9.2508e-05, 8.4936e-05, 9.1208e-05, 9.1392e-05, + 1.0023e-04, 1.0734e-04], device='cuda:0') +2023-03-27 11:24:37,009 INFO [finetune.py:976] (0/7) Epoch 29, batch 3350, loss[loss=0.1911, simple_loss=0.2579, pruned_loss=0.06216, over 4877.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2447, pruned_loss=0.05001, over 955552.41 frames. ], batch size: 32, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:24:47,797 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.3062, 2.1606, 1.8646, 2.1239, 2.0813, 2.0504, 2.1350, 2.7937], + device='cuda:0'), covar=tensor([0.3424, 0.4128, 0.3144, 0.3664, 0.4026, 0.2432, 0.3618, 0.1634], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0260, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:24:55,338 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:25:00,019 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 11:25:21,353 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.486e+02 1.801e+02 2.174e+02 4.171e+02, threshold=3.603e+02, percent-clipped=1.0 +2023-03-27 11:25:21,369 INFO [finetune.py:976] (0/7) Epoch 29, batch 3400, loss[loss=0.1476, simple_loss=0.2321, pruned_loss=0.03158, over 4873.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2457, pruned_loss=0.05016, over 955993.16 frames. ], batch size: 34, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:25:37,709 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:25:49,653 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 11:25:58,975 INFO [finetune.py:976] (0/7) Epoch 29, batch 3450, loss[loss=0.1606, simple_loss=0.2389, pruned_loss=0.04116, over 4932.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2446, pruned_loss=0.04941, over 956084.82 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:26:41,241 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.103e+01 1.539e+02 1.877e+02 2.200e+02 3.171e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-27 11:26:41,257 INFO [finetune.py:976] (0/7) Epoch 29, batch 3500, loss[loss=0.1643, simple_loss=0.2357, pruned_loss=0.04651, over 4816.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.242, pruned_loss=0.04819, over 957395.05 frames. ], batch size: 41, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:26:55,481 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163896.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:27:17,137 INFO [finetune.py:976] (0/7) Epoch 29, batch 3550, loss[loss=0.1504, simple_loss=0.2249, pruned_loss=0.03792, over 4913.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2396, pruned_loss=0.04768, over 957879.24 frames. ], batch size: 37, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:27:19,716 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9641, 1.8702, 1.6350, 2.0855, 2.4252, 2.1019, 1.8634, 1.5689], + device='cuda:0'), covar=tensor([0.1937, 0.1787, 0.1807, 0.1530, 0.1452, 0.1145, 0.2063, 0.1870], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0212, 0.0217, 0.0199, 0.0246, 0.0191, 0.0218, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:27:26,555 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-27 11:27:38,070 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163944.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:27:59,310 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.875e+01 1.446e+02 1.734e+02 2.162e+02 4.667e+02, threshold=3.468e+02, percent-clipped=1.0 +2023-03-27 11:27:59,326 INFO [finetune.py:976] (0/7) Epoch 29, batch 3600, loss[loss=0.1521, simple_loss=0.2298, pruned_loss=0.03717, over 4899.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2378, pruned_loss=0.04712, over 956140.90 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:28:01,230 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0616, 1.2781, 0.6714, 1.8854, 2.3200, 1.8331, 1.5409, 1.7973], + device='cuda:0'), covar=tensor([0.1294, 0.2017, 0.2027, 0.1027, 0.1830, 0.1937, 0.1316, 0.1893], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0108, 0.0093, 0.0119, 0.0091, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:28:15,611 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-164000.pt +2023-03-27 11:28:21,501 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6519, 1.3649, 2.1961, 3.1914, 2.1005, 2.3607, 1.0628, 2.7063], + device='cuda:0'), covar=tensor([0.1726, 0.1517, 0.1182, 0.0601, 0.0856, 0.2054, 0.1759, 0.0465], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0125, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:28:32,048 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0605, 2.0426, 1.6420, 1.9490, 1.8918, 1.8439, 1.9626, 2.5540], + device='cuda:0'), covar=tensor([0.3405, 0.3728, 0.3221, 0.3369, 0.3852, 0.2335, 0.3474, 0.1647], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0240, 0.0275, 0.0261, 0.0232, 0.0260, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:28:36,273 INFO [finetune.py:976] (0/7) Epoch 29, batch 3650, loss[loss=0.1773, simple_loss=0.2324, pruned_loss=0.06111, over 4307.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2397, pruned_loss=0.04765, over 955407.43 frames. ], batch size: 18, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:29:10,775 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 11:29:19,111 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.932e+01 1.541e+02 1.828e+02 2.331e+02 7.133e+02, threshold=3.656e+02, percent-clipped=4.0 +2023-03-27 11:29:19,127 INFO [finetune.py:976] (0/7) Epoch 29, batch 3700, loss[loss=0.1368, simple_loss=0.2116, pruned_loss=0.031, over 4779.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2426, pruned_loss=0.04839, over 954052.15 frames. ], batch size: 26, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:29:21,172 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 11:29:36,841 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:29:56,496 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 11:30:00,904 INFO [finetune.py:976] (0/7) Epoch 29, batch 3750, loss[loss=0.1909, simple_loss=0.2678, pruned_loss=0.057, over 4896.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2438, pruned_loss=0.0486, over 953934.32 frames. ], batch size: 43, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:30:17,247 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164151.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:30:20,022 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164155.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:30:37,150 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.292e+01 1.655e+02 1.832e+02 2.179e+02 3.362e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 11:30:37,166 INFO [finetune.py:976] (0/7) Epoch 29, batch 3800, loss[loss=0.2078, simple_loss=0.2959, pruned_loss=0.05986, over 4911.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2457, pruned_loss=0.04933, over 954321.86 frames. ], batch size: 46, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:31:03,902 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:31:09,785 INFO [finetune.py:976] (0/7) Epoch 29, batch 3850, loss[loss=0.1443, simple_loss=0.2242, pruned_loss=0.03216, over 4748.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2445, pruned_loss=0.04857, over 953604.62 frames. ], batch size: 26, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:31:11,108 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-03-27 11:31:19,346 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6895, 1.4266, 2.3688, 1.9825, 1.7924, 4.1696, 1.4229, 1.6438], + device='cuda:0'), covar=tensor([0.0988, 0.2060, 0.1087, 0.1018, 0.1727, 0.0200, 0.1700, 0.1978], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0084, 0.0074, 0.0077, 0.0092, 0.0082, 0.0087, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:31:25,438 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9279, 1.4597, 0.9300, 1.7058, 2.2275, 1.6366, 1.8009, 1.6520], + device='cuda:0'), covar=tensor([0.1302, 0.1879, 0.1838, 0.1164, 0.1770, 0.1859, 0.1255, 0.1953], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0093, 0.0108, 0.0093, 0.0120, 0.0092, 0.0097, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 11:31:45,642 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.748e+01 1.394e+02 1.747e+02 2.221e+02 3.425e+02, threshold=3.494e+02, percent-clipped=0.0 +2023-03-27 11:31:45,658 INFO [finetune.py:976] (0/7) Epoch 29, batch 3900, loss[loss=0.1455, simple_loss=0.2271, pruned_loss=0.03199, over 4901.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.242, pruned_loss=0.04779, over 955580.06 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:32:27,499 INFO [finetune.py:976] (0/7) Epoch 29, batch 3950, loss[loss=0.1444, simple_loss=0.207, pruned_loss=0.04084, over 4740.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2396, pruned_loss=0.04721, over 956048.91 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:32:58,620 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164360.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:32:58,790 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.37 vs. limit=5.0 +2023-03-27 11:33:03,100 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6685, 1.6079, 1.4428, 1.8243, 1.9082, 1.7815, 1.3006, 1.4282], + device='cuda:0'), covar=tensor([0.2311, 0.2108, 0.2075, 0.1664, 0.1708, 0.1326, 0.2701, 0.2084], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0212, 0.0217, 0.0199, 0.0246, 0.0191, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:33:11,862 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.819e+01 1.491e+02 1.748e+02 1.970e+02 3.605e+02, threshold=3.496e+02, percent-clipped=1.0 +2023-03-27 11:33:11,878 INFO [finetune.py:976] (0/7) Epoch 29, batch 4000, loss[loss=0.142, simple_loss=0.2268, pruned_loss=0.02856, over 4835.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2376, pruned_loss=0.0464, over 957692.87 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:33:42,976 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:33:45,304 INFO [finetune.py:976] (0/7) Epoch 29, batch 4050, loss[loss=0.1966, simple_loss=0.2786, pruned_loss=0.05726, over 4829.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2403, pruned_loss=0.0474, over 957316.87 frames. ], batch size: 40, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:34:06,763 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:34:29,030 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.788e+01 1.600e+02 1.814e+02 2.182e+02 3.811e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-03-27 11:34:29,046 INFO [finetune.py:976] (0/7) Epoch 29, batch 4100, loss[loss=0.2334, simple_loss=0.2975, pruned_loss=0.08463, over 4821.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2429, pruned_loss=0.04816, over 957035.86 frames. ], batch size: 40, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:04,640 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:35:13,894 INFO [finetune.py:976] (0/7) Epoch 29, batch 4150, loss[loss=0.1514, simple_loss=0.217, pruned_loss=0.04286, over 4644.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2444, pruned_loss=0.04899, over 954902.85 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:35,843 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.0318, 1.6117, 2.1842, 1.5334, 2.0117, 2.2158, 1.5654, 2.2907], + device='cuda:0'), covar=tensor([0.1213, 0.2153, 0.1511, 0.1931, 0.0941, 0.1371, 0.2886, 0.0775], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0206, 0.0194, 0.0189, 0.0174, 0.0212, 0.0218, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:35:42,354 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164568.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:35:46,473 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.582e+02 1.835e+02 2.360e+02 4.097e+02, threshold=3.670e+02, percent-clipped=1.0 +2023-03-27 11:35:46,489 INFO [finetune.py:976] (0/7) Epoch 29, batch 4200, loss[loss=0.1583, simple_loss=0.2294, pruned_loss=0.04361, over 4858.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2452, pruned_loss=0.04907, over 954904.98 frames. ], batch size: 31, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:48,905 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164578.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:36:00,595 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 11:36:09,804 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-27 11:36:20,292 INFO [finetune.py:976] (0/7) Epoch 29, batch 4250, loss[loss=0.1789, simple_loss=0.2386, pruned_loss=0.0596, over 4711.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2421, pruned_loss=0.04798, over 954925.37 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:36:23,256 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:36:29,257 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164639.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 11:36:43,154 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164658.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:36:53,782 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.450e+02 1.663e+02 2.112e+02 3.483e+02, threshold=3.326e+02, percent-clipped=0.0 +2023-03-27 11:36:53,798 INFO [finetune.py:976] (0/7) Epoch 29, batch 4300, loss[loss=0.1519, simple_loss=0.2288, pruned_loss=0.03748, over 4913.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2391, pruned_loss=0.04719, over 953610.13 frames. ], batch size: 36, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:37:39,419 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:41,321 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164719.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:44,824 INFO [finetune.py:976] (0/7) Epoch 29, batch 4350, loss[loss=0.193, simple_loss=0.2569, pruned_loss=0.06449, over 4829.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2365, pruned_loss=0.04646, over 952618.21 frames. ], batch size: 39, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:37:46,788 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 11:37:58,700 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:58,714 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:20,795 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.196e+01 1.395e+02 1.786e+02 2.060e+02 3.738e+02, threshold=3.572e+02, percent-clipped=2.0 +2023-03-27 11:38:20,811 INFO [finetune.py:976] (0/7) Epoch 29, batch 4400, loss[loss=0.1353, simple_loss=0.2086, pruned_loss=0.03097, over 4734.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2377, pruned_loss=0.04679, over 951936.28 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:38:21,546 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7559, 1.6463, 1.5602, 1.6497, 1.3005, 4.0972, 1.5320, 1.7322], + device='cuda:0'), covar=tensor([0.3364, 0.2442, 0.2147, 0.2358, 0.1692, 0.0144, 0.2565, 0.1282], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 11:38:22,929 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 11:38:33,468 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:43,361 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:46,285 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164811.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:46,315 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8813, 1.6968, 1.9826, 1.1472, 1.8425, 1.9898, 1.9333, 1.5764], + device='cuda:0'), covar=tensor([0.0661, 0.0805, 0.0654, 0.0980, 0.0874, 0.0674, 0.0664, 0.1287], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0119, 0.0129, 0.0140, 0.0140, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:38:54,767 INFO [finetune.py:976] (0/7) Epoch 29, batch 4450, loss[loss=0.1407, simple_loss=0.2136, pruned_loss=0.03391, over 4768.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2409, pruned_loss=0.04812, over 953685.37 frames. ], batch size: 26, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:39:23,088 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164859.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:39:37,134 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.552e+02 1.889e+02 2.217e+02 4.905e+02, threshold=3.778e+02, percent-clipped=1.0 +2023-03-27 11:39:37,150 INFO [finetune.py:976] (0/7) Epoch 29, batch 4500, loss[loss=0.1948, simple_loss=0.2666, pruned_loss=0.06151, over 4907.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2428, pruned_loss=0.04876, over 953054.98 frames. ], batch size: 36, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:40:22,132 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164924.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:40:22,678 INFO [finetune.py:976] (0/7) Epoch 29, batch 4550, loss[loss=0.1845, simple_loss=0.2582, pruned_loss=0.05541, over 4860.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2442, pruned_loss=0.04926, over 951056.03 frames. ], batch size: 44, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:40:28,154 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164934.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:40:55,993 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.494e+02 1.764e+02 2.048e+02 3.220e+02, threshold=3.528e+02, percent-clipped=0.0 +2023-03-27 11:40:56,009 INFO [finetune.py:976] (0/7) Epoch 29, batch 4600, loss[loss=0.1315, simple_loss=0.207, pruned_loss=0.02798, over 4770.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.244, pruned_loss=0.04853, over 953980.90 frames. ], batch size: 28, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:40:58,467 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.2399, 3.6404, 3.8862, 4.0424, 3.9971, 3.6791, 4.3028, 1.4910], + device='cuda:0'), covar=tensor([0.0828, 0.0928, 0.0927, 0.1033, 0.1222, 0.1827, 0.0711, 0.5554], + device='cuda:0'), in_proj_covar=tensor([0.0357, 0.0249, 0.0289, 0.0299, 0.0341, 0.0289, 0.0310, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:41:22,216 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:41:23,387 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165016.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:41:28,739 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4051, 1.4767, 1.7239, 1.6870, 1.6591, 2.7189, 1.4453, 1.5886], + device='cuda:0'), covar=tensor([0.0937, 0.1537, 0.1148, 0.0825, 0.1314, 0.0339, 0.1241, 0.1418], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0077, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-27 11:41:29,265 INFO [finetune.py:976] (0/7) Epoch 29, batch 4650, loss[loss=0.1907, simple_loss=0.2555, pruned_loss=0.06294, over 4150.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2417, pruned_loss=0.0479, over 953726.25 frames. ], batch size: 65, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:41:53,747 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165064.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:01,328 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.478e+02 1.798e+02 2.114e+02 1.110e+03, threshold=3.597e+02, percent-clipped=3.0 +2023-03-27 11:42:01,343 INFO [finetune.py:976] (0/7) Epoch 29, batch 4700, loss[loss=0.1618, simple_loss=0.228, pruned_loss=0.04777, over 4247.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2381, pruned_loss=0.04656, over 952685.84 frames. ], batch size: 65, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:42:27,130 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:43,731 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:44,885 INFO [finetune.py:976] (0/7) Epoch 29, batch 4750, loss[loss=0.1594, simple_loss=0.2274, pruned_loss=0.04567, over 4735.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2356, pruned_loss=0.04566, over 953138.11 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:42:55,605 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9220, 1.7861, 1.5570, 1.3947, 1.7371, 1.7724, 1.7309, 2.2458], + device='cuda:0'), covar=tensor([0.3495, 0.3790, 0.3148, 0.3519, 0.3826, 0.2318, 0.3287, 0.1774], + device='cuda:0'), in_proj_covar=tensor([0.0288, 0.0263, 0.0239, 0.0273, 0.0260, 0.0232, 0.0258, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:43:21,070 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9328, 1.8432, 1.5318, 1.6758, 1.7417, 1.6989, 1.7944, 2.4063], + device='cuda:0'), covar=tensor([0.3567, 0.3626, 0.3172, 0.3408, 0.3807, 0.2385, 0.3424, 0.1664], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0259, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:43:21,526 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.579e+02 1.786e+02 2.031e+02 3.721e+02, threshold=3.572e+02, percent-clipped=1.0 +2023-03-27 11:43:21,542 INFO [finetune.py:976] (0/7) Epoch 29, batch 4800, loss[loss=0.1951, simple_loss=0.2849, pruned_loss=0.05267, over 4893.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2399, pruned_loss=0.0473, over 955356.95 frames. ], batch size: 43, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:43:27,583 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165184.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:43:34,860 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8120, 1.4079, 1.9293, 1.8629, 1.6738, 1.6289, 1.8142, 1.8014], + device='cuda:0'), covar=tensor([0.3712, 0.3534, 0.2734, 0.3370, 0.4368, 0.3511, 0.3751, 0.2673], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0250, 0.0270, 0.0301, 0.0300, 0.0277, 0.0306, 0.0256], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:43:53,932 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:43:54,433 INFO [finetune.py:976] (0/7) Epoch 29, batch 4850, loss[loss=0.2133, simple_loss=0.2783, pruned_loss=0.07414, over 4911.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2432, pruned_loss=0.04845, over 953185.55 frames. ], batch size: 36, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:44:00,488 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165234.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 11:44:08,503 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-03-27 11:44:24,478 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:44:26,737 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.742e+02 1.982e+02 2.317e+02 4.079e+02, threshold=3.965e+02, percent-clipped=3.0 +2023-03-27 11:44:26,753 INFO [finetune.py:976] (0/7) Epoch 29, batch 4900, loss[loss=0.151, simple_loss=0.2313, pruned_loss=0.03534, over 4770.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2442, pruned_loss=0.04832, over 954896.17 frames. ], batch size: 29, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:44:41,266 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:44:57,581 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-27 11:45:03,720 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:45:16,147 INFO [finetune.py:976] (0/7) Epoch 29, batch 4950, loss[loss=0.1337, simple_loss=0.192, pruned_loss=0.03774, over 4164.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2464, pruned_loss=0.04942, over 954231.78 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:45:48,522 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165362.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:45:51,018 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5659, 2.2632, 1.7904, 0.7290, 2.0012, 2.0920, 1.9201, 2.1256], + device='cuda:0'), covar=tensor([0.0918, 0.0771, 0.1681, 0.2202, 0.1366, 0.2176, 0.2107, 0.0865], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0189, 0.0203, 0.0180, 0.0209, 0.0210, 0.0222, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:45:56,866 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.088e+01 1.449e+02 1.743e+02 2.087e+02 3.629e+02, threshold=3.486e+02, percent-clipped=0.0 +2023-03-27 11:45:56,882 INFO [finetune.py:976] (0/7) Epoch 29, batch 5000, loss[loss=0.1301, simple_loss=0.2032, pruned_loss=0.02856, over 4911.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.244, pruned_loss=0.04871, over 951865.09 frames. ], batch size: 43, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:45:59,539 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 11:46:15,409 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165402.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:46:30,169 INFO [finetune.py:976] (0/7) Epoch 29, batch 5050, loss[loss=0.1568, simple_loss=0.2315, pruned_loss=0.04105, over 4891.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2408, pruned_loss=0.04751, over 951449.55 frames. ], batch size: 32, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:46:47,814 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:47:03,404 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.397e+02 1.700e+02 2.095e+02 3.650e+02, threshold=3.400e+02, percent-clipped=1.0 +2023-03-27 11:47:03,420 INFO [finetune.py:976] (0/7) Epoch 29, batch 5100, loss[loss=0.1347, simple_loss=0.2008, pruned_loss=0.03424, over 4712.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2366, pruned_loss=0.04598, over 951387.39 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:47:06,334 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165479.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:47:16,944 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 11:47:46,456 INFO [finetune.py:976] (0/7) Epoch 29, batch 5150, loss[loss=0.1809, simple_loss=0.2468, pruned_loss=0.0575, over 4928.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2381, pruned_loss=0.04687, over 951951.92 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:47:54,884 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9786, 1.6626, 2.3880, 1.5244, 2.1408, 2.2529, 1.5131, 2.3526], + device='cuda:0'), covar=tensor([0.1422, 0.2192, 0.1414, 0.1908, 0.1125, 0.1509, 0.2853, 0.1098], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0207, 0.0195, 0.0190, 0.0175, 0.0213, 0.0220, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:48:04,791 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:48:10,289 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 11:48:12,735 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 11:48:20,246 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.818e+01 1.606e+02 1.856e+02 2.249e+02 3.990e+02, threshold=3.713e+02, percent-clipped=3.0 +2023-03-27 11:48:20,262 INFO [finetune.py:976] (0/7) Epoch 29, batch 5200, loss[loss=0.1729, simple_loss=0.2508, pruned_loss=0.04747, over 4918.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2409, pruned_loss=0.04769, over 953230.22 frames. ], batch size: 36, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:48:32,658 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2955, 2.1592, 1.8619, 2.3427, 2.8133, 2.3210, 2.2154, 1.7378], + device='cuda:0'), covar=tensor([0.2225, 0.1883, 0.1873, 0.1592, 0.1686, 0.1171, 0.1919, 0.2001], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0213, 0.0217, 0.0200, 0.0247, 0.0192, 0.0218, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:48:45,780 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165612.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:48:53,940 INFO [finetune.py:976] (0/7) Epoch 29, batch 5250, loss[loss=0.1962, simple_loss=0.267, pruned_loss=0.06273, over 4746.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.243, pruned_loss=0.04845, over 952413.64 frames. ], batch size: 59, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:48:57,073 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8050, 1.8058, 1.5182, 1.9937, 2.5218, 2.0041, 1.9087, 1.4909], + device='cuda:0'), covar=tensor([0.2193, 0.1897, 0.1871, 0.1555, 0.1508, 0.1200, 0.2037, 0.1934], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0213, 0.0217, 0.0200, 0.0247, 0.0192, 0.0218, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:49:08,013 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.4342, 1.5362, 1.5312, 0.8694, 1.6245, 1.8800, 1.7832, 1.4039], + device='cuda:0'), covar=tensor([0.0990, 0.0633, 0.0568, 0.0530, 0.0521, 0.0556, 0.0357, 0.0771], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0145, 0.0130, 0.0120, 0.0131, 0.0129, 0.0140, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.7672e-05, 1.0407e-04, 9.2378e-05, 8.4235e-05, 9.1421e-05, 9.1531e-05, + 9.9776e-05, 1.0732e-04], device='cuda:0') +2023-03-27 11:49:26,776 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.319e+01 1.511e+02 1.772e+02 2.250e+02 3.554e+02, threshold=3.544e+02, percent-clipped=0.0 +2023-03-27 11:49:26,792 INFO [finetune.py:976] (0/7) Epoch 29, batch 5300, loss[loss=0.2301, simple_loss=0.2877, pruned_loss=0.08621, over 4861.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2447, pruned_loss=0.04944, over 953143.38 frames. ], batch size: 34, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:50:10,068 INFO [finetune.py:976] (0/7) Epoch 29, batch 5350, loss[loss=0.2251, simple_loss=0.2827, pruned_loss=0.08378, over 4270.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2445, pruned_loss=0.04858, over 951450.45 frames. ], batch size: 66, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:50:33,768 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165750.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:50:52,699 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6847, 1.6105, 1.5775, 1.6941, 1.3184, 3.6945, 1.4984, 1.8128], + device='cuda:0'), covar=tensor([0.3068, 0.2289, 0.2095, 0.2250, 0.1597, 0.0173, 0.2536, 0.1242], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0117, 0.0121, 0.0125, 0.0114, 0.0096, 0.0094, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 11:50:59,727 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.485e+02 1.788e+02 2.126e+02 3.231e+02, threshold=3.576e+02, percent-clipped=0.0 +2023-03-27 11:50:59,743 INFO [finetune.py:976] (0/7) Epoch 29, batch 5400, loss[loss=0.1438, simple_loss=0.2169, pruned_loss=0.03538, over 4820.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2421, pruned_loss=0.04778, over 953360.24 frames. ], batch size: 41, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:50:59,853 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6628, 1.5709, 1.5585, 1.5708, 1.2810, 3.0681, 1.1788, 1.6482], + device='cuda:0'), covar=tensor([0.3512, 0.2655, 0.2217, 0.2465, 0.1774, 0.0312, 0.2782, 0.1329], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0117, 0.0121, 0.0125, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 11:51:02,237 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:09,560 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-27 11:51:17,326 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1630, 2.0342, 1.7752, 1.9538, 1.9644, 1.9320, 1.9666, 2.6850], + device='cuda:0'), covar=tensor([0.3563, 0.3990, 0.3223, 0.3617, 0.3783, 0.2478, 0.3693, 0.1682], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0259, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:51:17,897 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165803.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:24,163 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165811.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:33,003 INFO [finetune.py:976] (0/7) Epoch 29, batch 5450, loss[loss=0.1355, simple_loss=0.2112, pruned_loss=0.02992, over 4883.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2386, pruned_loss=0.04647, over 954792.57 frames. ], batch size: 31, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:51:34,293 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:59,791 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165864.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:52:06,352 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.128e+01 1.456e+02 1.721e+02 2.001e+02 5.868e+02, threshold=3.442e+02, percent-clipped=2.0 +2023-03-27 11:52:06,368 INFO [finetune.py:976] (0/7) Epoch 29, batch 5500, loss[loss=0.1347, simple_loss=0.2013, pruned_loss=0.03404, over 4831.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2358, pruned_loss=0.04573, over 957135.97 frames. ], batch size: 30, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:52:27,319 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165907.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:52:40,105 INFO [finetune.py:976] (0/7) Epoch 29, batch 5550, loss[loss=0.1746, simple_loss=0.244, pruned_loss=0.05258, over 4829.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.239, pruned_loss=0.0472, over 958168.83 frames. ], batch size: 30, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:53:13,744 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-03-27 11:53:21,198 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 11:53:22,714 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.570e+02 1.926e+02 2.203e+02 4.633e+02, threshold=3.853e+02, percent-clipped=1.0 +2023-03-27 11:53:22,730 INFO [finetune.py:976] (0/7) Epoch 29, batch 5600, loss[loss=0.2272, simple_loss=0.2845, pruned_loss=0.08499, over 4834.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2424, pruned_loss=0.04796, over 956807.89 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:53:37,443 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-166000.pt +2023-03-27 11:53:54,008 INFO [finetune.py:976] (0/7) Epoch 29, batch 5650, loss[loss=0.2133, simple_loss=0.2752, pruned_loss=0.07571, over 4893.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2445, pruned_loss=0.04829, over 959079.73 frames. ], batch size: 32, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:54:11,184 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166054.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:54:23,584 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.094e+01 1.435e+02 1.719e+02 2.053e+02 3.744e+02, threshold=3.437e+02, percent-clipped=0.0 +2023-03-27 11:54:23,599 INFO [finetune.py:976] (0/7) Epoch 29, batch 5700, loss[loss=0.1261, simple_loss=0.1972, pruned_loss=0.02745, over 4091.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2411, pruned_loss=0.04792, over 941033.72 frames. ], batch size: 17, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:54:40,444 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/epoch-29.pt +2023-03-27 11:54:50,338 INFO [finetune.py:976] (0/7) Epoch 30, batch 0, loss[loss=0.168, simple_loss=0.2451, pruned_loss=0.04542, over 4857.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2451, pruned_loss=0.04542, over 4857.00 frames. ], batch size: 34, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:54:50,339 INFO [finetune.py:1001] (0/7) Computing validation loss +2023-03-27 11:54:52,161 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8689, 1.1360, 1.9549, 1.9037, 1.7428, 1.6729, 1.7797, 1.8946], + device='cuda:0'), covar=tensor([0.4188, 0.4090, 0.3423, 0.3799, 0.4952, 0.4132, 0.4485, 0.3068], + device='cuda:0'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0301, 0.0300, 0.0277, 0.0306, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:54:58,096 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1809, 1.9200, 1.8522, 1.8320, 1.8843, 1.8758, 1.8864, 2.5633], + device='cuda:0'), covar=tensor([0.3581, 0.4444, 0.3212, 0.3406, 0.3980, 0.2551, 0.3734, 0.1694], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0259, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:55:06,737 INFO [finetune.py:1010] (0/7) Epoch 30, validation: loss=0.1598, simple_loss=0.2264, pruned_loss=0.04658, over 2265189.00 frames. +2023-03-27 11:55:06,738 INFO [finetune.py:1011] (0/7) Maximum memory allocated so far is 6456MB +2023-03-27 11:55:11,871 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166106.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:55:21,141 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166115.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:55:41,975 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 11:55:43,737 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:55:52,370 INFO [finetune.py:976] (0/7) Epoch 30, batch 50, loss[loss=0.1742, simple_loss=0.25, pruned_loss=0.04915, over 4817.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2478, pruned_loss=0.0516, over 217456.56 frames. ], batch size: 38, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:55:53,240 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 11:55:54,324 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-27 11:56:02,558 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166159.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:56:16,051 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.060e+01 1.403e+02 1.686e+02 1.983e+02 3.736e+02, threshold=3.372e+02, percent-clipped=1.0 +2023-03-27 11:56:32,523 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9840, 1.8754, 2.0437, 1.3502, 1.9545, 2.0848, 2.0278, 1.6069], + device='cuda:0'), covar=tensor([0.0595, 0.0708, 0.0653, 0.0852, 0.0843, 0.0645, 0.0585, 0.1165], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0139, 0.0142, 0.0119, 0.0130, 0.0140, 0.0141, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 11:56:34,675 INFO [finetune.py:976] (0/7) Epoch 30, batch 100, loss[loss=0.1617, simple_loss=0.2302, pruned_loss=0.04658, over 4781.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2373, pruned_loss=0.04699, over 379523.39 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:56:38,209 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166207.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:56:39,347 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166208.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:57:00,953 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 11:57:07,205 INFO [finetune.py:976] (0/7) Epoch 30, batch 150, loss[loss=0.1757, simple_loss=0.2308, pruned_loss=0.06028, over 4134.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2335, pruned_loss=0.04566, over 507689.35 frames. ], batch size: 18, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:57:08,472 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166255.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:57:21,826 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.442e+02 1.804e+02 2.095e+02 4.016e+02, threshold=3.609e+02, percent-clipped=1.0 +2023-03-27 11:57:39,809 INFO [finetune.py:976] (0/7) Epoch 30, batch 200, loss[loss=0.1361, simple_loss=0.2146, pruned_loss=0.02878, over 4771.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2331, pruned_loss=0.04591, over 606689.74 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:57:42,332 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:00,761 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 11:58:14,802 INFO [finetune.py:976] (0/7) Epoch 30, batch 250, loss[loss=0.1908, simple_loss=0.2723, pruned_loss=0.05464, over 4829.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.235, pruned_loss=0.04594, over 685202.72 frames. ], batch size: 40, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:58:19,111 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.3043, 1.4947, 1.5762, 0.8028, 1.5803, 1.7948, 1.8189, 1.4063], + device='cuda:0'), covar=tensor([0.0907, 0.0668, 0.0543, 0.0515, 0.0465, 0.0633, 0.0381, 0.0712], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0147, 0.0131, 0.0121, 0.0131, 0.0130, 0.0141, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.8058e-05, 1.0484e-04, 9.2994e-05, 8.4829e-05, 9.1985e-05, 9.2202e-05, + 1.0038e-04, 1.0789e-04], device='cuda:0') +2023-03-27 11:58:26,039 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166368.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:26,668 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6894, 1.6530, 1.4731, 1.6503, 1.9742, 1.9105, 1.6757, 1.4891], + device='cuda:0'), covar=tensor([0.0384, 0.0349, 0.0604, 0.0374, 0.0248, 0.0509, 0.0399, 0.0469], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0107, 0.0149, 0.0112, 0.0104, 0.0119, 0.0105, 0.0117], + device='cuda:0'), out_proj_covar=tensor([8.0245e-05, 8.2049e-05, 1.1591e-04, 8.4978e-05, 8.0031e-05, 8.7710e-05, + 7.7878e-05, 8.8393e-05], device='cuda:0') +2023-03-27 11:58:30,118 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.508e+02 1.830e+02 2.341e+02 4.348e+02, threshold=3.661e+02, percent-clipped=2.0 +2023-03-27 11:58:48,138 INFO [finetune.py:976] (0/7) Epoch 30, batch 300, loss[loss=0.1798, simple_loss=0.2432, pruned_loss=0.05824, over 4884.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2381, pruned_loss=0.0465, over 744969.32 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:58:50,085 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166406.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:52,513 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166410.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 11:59:21,479 INFO [finetune.py:976] (0/7) Epoch 30, batch 350, loss[loss=0.1783, simple_loss=0.2479, pruned_loss=0.0543, over 4861.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2408, pruned_loss=0.04798, over 790303.31 frames. ], batch size: 31, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:59:22,176 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166454.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:25,721 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166459.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:37,685 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.476e+02 1.813e+02 2.161e+02 3.890e+02, threshold=3.626e+02, percent-clipped=2.0 +2023-03-27 11:59:41,428 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166481.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:55,115 INFO [finetune.py:976] (0/7) Epoch 30, batch 400, loss[loss=0.2059, simple_loss=0.2708, pruned_loss=0.07054, over 4892.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2435, pruned_loss=0.04864, over 828754.90 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:59:55,185 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166503.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:58,013 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166507.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:59,267 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:59,346 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 12:00:20,214 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166531.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:00:35,352 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166542.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:00:45,994 INFO [finetune.py:976] (0/7) Epoch 30, batch 450, loss[loss=0.1948, simple_loss=0.2699, pruned_loss=0.05981, over 4907.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2424, pruned_loss=0.04838, over 858100.47 frames. ], batch size: 36, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:00:56,912 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166570.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:01:00,792 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.283e+01 1.415e+02 1.694e+02 2.083e+02 4.695e+02, threshold=3.388e+02, percent-clipped=2.0 +2023-03-27 12:01:03,263 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1829, 2.1178, 1.7099, 2.1632, 2.1172, 1.8786, 2.5126, 2.1949], + device='cuda:0'), covar=tensor([0.1315, 0.1998, 0.2873, 0.2504, 0.2476, 0.1741, 0.2895, 0.1570], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0254, 0.0251, 0.0210, 0.0216, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:01:22,336 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166592.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:01:32,866 INFO [finetune.py:976] (0/7) Epoch 30, batch 500, loss[loss=0.1524, simple_loss=0.2233, pruned_loss=0.04072, over 4764.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2399, pruned_loss=0.04768, over 880511.40 frames. ], batch size: 28, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:01:46,179 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.1012, 3.6063, 3.7697, 3.9587, 3.8884, 3.5776, 4.1658, 1.3235], + device='cuda:0'), covar=tensor([0.0835, 0.0917, 0.1094, 0.1034, 0.1227, 0.1832, 0.0842, 0.6178], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0249, 0.0288, 0.0298, 0.0339, 0.0288, 0.0306, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:02:06,080 INFO [finetune.py:976] (0/7) Epoch 30, batch 550, loss[loss=0.1725, simple_loss=0.2416, pruned_loss=0.05173, over 4829.00 frames. ], tot_loss[loss=0.165, simple_loss=0.237, pruned_loss=0.04654, over 897216.46 frames. ], batch size: 30, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:02:12,183 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166663.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:02:20,414 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 7.871e+01 1.519e+02 1.786e+02 2.255e+02 3.834e+02, threshold=3.573e+02, percent-clipped=3.0 +2023-03-27 12:02:39,362 INFO [finetune.py:976] (0/7) Epoch 30, batch 600, loss[loss=0.1579, simple_loss=0.2371, pruned_loss=0.03939, over 4783.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2372, pruned_loss=0.046, over 912428.00 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:02:43,707 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166710.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 12:02:45,050 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 12:03:00,923 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:12,732 INFO [finetune.py:976] (0/7) Epoch 30, batch 650, loss[loss=0.1468, simple_loss=0.2156, pruned_loss=0.03901, over 4902.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2405, pruned_loss=0.04671, over 921652.16 frames. ], batch size: 35, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:03:15,787 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166758.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 12:03:19,465 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166764.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:21,883 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8504, 1.4110, 1.9389, 1.9333, 1.7117, 1.6956, 1.8921, 1.8471], + device='cuda:0'), covar=tensor([0.4145, 0.3909, 0.3242, 0.3811, 0.4883, 0.4101, 0.4375, 0.2998], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0251, 0.0270, 0.0301, 0.0301, 0.0279, 0.0306, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:03:26,540 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.552e+02 1.821e+02 2.218e+02 3.611e+02, threshold=3.642e+02, percent-clipped=1.0 +2023-03-27 12:03:41,554 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166796.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:45,812 INFO [finetune.py:976] (0/7) Epoch 30, batch 700, loss[loss=0.1865, simple_loss=0.2602, pruned_loss=0.05643, over 4832.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2415, pruned_loss=0.04696, over 926666.67 frames. ], batch size: 49, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:03:45,914 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166803.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:00,341 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166825.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:08,130 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166837.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:18,068 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166851.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:19,223 INFO [finetune.py:976] (0/7) Epoch 30, batch 750, loss[loss=0.139, simple_loss=0.2235, pruned_loss=0.02727, over 4857.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.242, pruned_loss=0.04686, over 934802.18 frames. ], batch size: 44, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:04:27,143 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:33,236 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.605e+01 1.409e+02 1.752e+02 2.054e+02 3.389e+02, threshold=3.504e+02, percent-clipped=0.0 +2023-03-27 12:04:41,696 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:52,668 INFO [finetune.py:976] (0/7) Epoch 30, batch 800, loss[loss=0.1851, simple_loss=0.2538, pruned_loss=0.05823, over 4775.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.243, pruned_loss=0.04711, over 939840.52 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:05:09,484 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:05:19,527 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6218, 1.5149, 1.0815, 0.3430, 1.2569, 1.4633, 1.4139, 1.4112], + device='cuda:0'), covar=tensor([0.0888, 0.0733, 0.1313, 0.1766, 0.1336, 0.2020, 0.1999, 0.0799], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0189, 0.0201, 0.0180, 0.0209, 0.0210, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:05:32,106 INFO [finetune.py:976] (0/7) Epoch 30, batch 850, loss[loss=0.1829, simple_loss=0.25, pruned_loss=0.05793, over 4749.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2399, pruned_loss=0.04612, over 944465.69 frames. ], batch size: 54, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:05:35,431 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-03-27 12:05:42,379 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166963.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:05:52,745 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1970, 1.3812, 0.7384, 2.0275, 2.6931, 1.9443, 1.7463, 1.8407], + device='cuda:0'), covar=tensor([0.1375, 0.2298, 0.2217, 0.1206, 0.1627, 0.1728, 0.1525, 0.2065], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0094, 0.0109, 0.0093, 0.0119, 0.0092, 0.0097, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 12:05:53,841 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.507e+02 1.769e+02 2.105e+02 4.355e+02, threshold=3.539e+02, percent-clipped=2.0 +2023-03-27 12:06:07,540 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166990.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:06:11,718 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.1084, 0.9262, 0.8879, 0.3398, 0.8523, 1.0653, 1.0875, 0.9178], + device='cuda:0'), covar=tensor([0.0811, 0.0539, 0.0520, 0.0506, 0.0579, 0.0554, 0.0338, 0.0584], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0146, 0.0131, 0.0121, 0.0131, 0.0130, 0.0141, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.8062e-05, 1.0470e-04, 9.2522e-05, 8.4448e-05, 9.1761e-05, 9.2035e-05, + 9.9985e-05, 1.0798e-04], device='cuda:0') +2023-03-27 12:06:16,010 INFO [finetune.py:976] (0/7) Epoch 30, batch 900, loss[loss=0.1657, simple_loss=0.2338, pruned_loss=0.04883, over 4900.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2375, pruned_loss=0.0459, over 944055.34 frames. ], batch size: 43, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:06:23,171 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167011.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:06:59,307 INFO [finetune.py:976] (0/7) Epoch 30, batch 950, loss[loss=0.1766, simple_loss=0.2439, pruned_loss=0.05465, over 4922.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2373, pruned_loss=0.0464, over 946981.22 frames. ], batch size: 37, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:07:11,953 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.1919, 1.8643, 2.3212, 1.5921, 2.0835, 2.2752, 1.7335, 2.3493], + device='cuda:0'), covar=tensor([0.1128, 0.1935, 0.1382, 0.1864, 0.0989, 0.1375, 0.2849, 0.0944], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0207, 0.0194, 0.0190, 0.0174, 0.0213, 0.0219, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:07:13,661 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.522e+02 1.848e+02 2.259e+02 3.601e+02, threshold=3.696e+02, percent-clipped=1.0 +2023-03-27 12:07:24,437 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167091.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:31,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167100.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:33,104 INFO [finetune.py:976] (0/7) Epoch 30, batch 1000, loss[loss=0.2081, simple_loss=0.2858, pruned_loss=0.06526, over 4827.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2392, pruned_loss=0.04692, over 948769.90 frames. ], batch size: 39, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:07:33,823 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167104.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:44,461 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167120.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:55,372 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167137.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:06,884 INFO [finetune.py:976] (0/7) Epoch 30, batch 1050, loss[loss=0.1727, simple_loss=0.2455, pruned_loss=0.04991, over 4819.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2417, pruned_loss=0.04746, over 949419.80 frames. ], batch size: 39, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:08:07,640 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5373, 2.3622, 2.1053, 1.0911, 2.3024, 1.9200, 1.7794, 2.2330], + device='cuda:0'), covar=tensor([0.0956, 0.0720, 0.1633, 0.1917, 0.1297, 0.2228, 0.2098, 0.0912], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0188, 0.0201, 0.0180, 0.0209, 0.0210, 0.0223, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:08:11,864 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:14,778 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:14,820 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:21,246 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.490e+02 1.818e+02 2.230e+02 3.401e+02, threshold=3.636e+02, percent-clipped=0.0 +2023-03-27 12:08:22,038 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.31 vs. limit=5.0 +2023-03-27 12:08:27,350 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167185.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:28,585 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167187.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:34,056 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 12:08:39,896 INFO [finetune.py:976] (0/7) Epoch 30, batch 1100, loss[loss=0.151, simple_loss=0.2375, pruned_loss=0.03221, over 4814.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2429, pruned_loss=0.04768, over 950553.00 frames. ], batch size: 51, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:08:46,970 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:09:01,419 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167235.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:09:13,764 INFO [finetune.py:976] (0/7) Epoch 30, batch 1150, loss[loss=0.1719, simple_loss=0.248, pruned_loss=0.04789, over 4732.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2434, pruned_loss=0.04822, over 950004.82 frames. ], batch size: 54, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:09:13,942 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 12:09:28,396 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.511e+02 1.772e+02 2.131e+02 4.281e+02, threshold=3.544e+02, percent-clipped=3.0 +2023-03-27 12:09:34,983 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167285.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:09:38,202 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-27 12:09:47,295 INFO [finetune.py:976] (0/7) Epoch 30, batch 1200, loss[loss=0.1569, simple_loss=0.2348, pruned_loss=0.03946, over 4246.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2436, pruned_loss=0.0485, over 951793.89 frames. ], batch size: 66, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:10:04,012 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9330, 1.5907, 2.0150, 2.0037, 1.7950, 1.7749, 1.9577, 1.9519], + device='cuda:0'), covar=tensor([0.4425, 0.3992, 0.3384, 0.3792, 0.4752, 0.3964, 0.4663, 0.3073], + device='cuda:0'), in_proj_covar=tensor([0.0270, 0.0250, 0.0270, 0.0301, 0.0301, 0.0279, 0.0306, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:10:20,449 INFO [finetune.py:976] (0/7) Epoch 30, batch 1250, loss[loss=0.1516, simple_loss=0.2293, pruned_loss=0.03692, over 4766.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2417, pruned_loss=0.04782, over 954063.03 frames. ], batch size: 27, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:10:36,983 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.462e+02 1.711e+02 2.091e+02 4.240e+02, threshold=3.422e+02, percent-clipped=1.0 +2023-03-27 12:10:56,504 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167391.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:09,177 INFO [finetune.py:976] (0/7) Epoch 30, batch 1300, loss[loss=0.1676, simple_loss=0.2428, pruned_loss=0.04622, over 4899.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2381, pruned_loss=0.04623, over 955727.27 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:11:24,564 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:32,405 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:38,987 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167439.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:41,460 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([0.0937, 0.9815, 0.9730, 0.4218, 0.8761, 1.1338, 1.1852, 0.9997], + device='cuda:0'), covar=tensor([0.0879, 0.0584, 0.0571, 0.0516, 0.0573, 0.0566, 0.0368, 0.0708], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0145, 0.0130, 0.0120, 0.0131, 0.0129, 0.0140, 0.0150], + device='cuda:0'), out_proj_covar=tensor([8.7866e-05, 1.0402e-04, 9.2133e-05, 8.3933e-05, 9.1372e-05, 9.1583e-05, + 9.9259e-05, 1.0724e-04], device='cuda:0') +2023-03-27 12:11:55,876 INFO [finetune.py:976] (0/7) Epoch 30, batch 1350, loss[loss=0.243, simple_loss=0.302, pruned_loss=0.09196, over 4052.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.239, pruned_loss=0.04663, over 955530.03 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:11:58,248 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167456.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:00,698 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:07,068 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:11,226 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.213e+01 1.429e+02 1.665e+02 1.960e+02 3.889e+02, threshold=3.329e+02, percent-clipped=2.0 +2023-03-27 12:12:23,216 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:29,688 INFO [finetune.py:976] (0/7) Epoch 30, batch 1400, loss[loss=0.1341, simple_loss=0.2122, pruned_loss=0.02799, over 4790.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2412, pruned_loss=0.04655, over 955768.29 frames. ], batch size: 25, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:12:43,988 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4655, 2.6362, 2.5526, 1.8994, 2.4288, 2.7966, 2.8817, 2.3160], + device='cuda:0'), covar=tensor([0.0673, 0.0595, 0.0753, 0.0858, 0.0873, 0.0660, 0.0543, 0.1049], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0118, 0.0128, 0.0139, 0.0139, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:12:45,921 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 12:12:51,033 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.9947, 2.0596, 1.6821, 1.7317, 2.4876, 2.5888, 2.0862, 1.9664], + device='cuda:0'), covar=tensor([0.0402, 0.0383, 0.0697, 0.0391, 0.0235, 0.0499, 0.0326, 0.0468], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0112, 0.0103, 0.0119, 0.0105, 0.0116], + device='cuda:0'), out_proj_covar=tensor([7.9911e-05, 8.1945e-05, 1.1616e-04, 8.4756e-05, 7.9671e-05, 8.7415e-05, + 7.7588e-05, 8.8018e-05], device='cuda:0') +2023-03-27 12:13:02,946 INFO [finetune.py:976] (0/7) Epoch 30, batch 1450, loss[loss=0.1821, simple_loss=0.2653, pruned_loss=0.04944, over 4846.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2419, pruned_loss=0.04637, over 956761.65 frames. ], batch size: 44, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:13:19,227 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.535e+02 1.844e+02 2.174e+02 4.375e+02, threshold=3.689e+02, percent-clipped=4.0 +2023-03-27 12:13:21,226 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2000, 2.0640, 1.7889, 2.1717, 2.7297, 2.2429, 2.1855, 1.5519], + device='cuda:0'), covar=tensor([0.2140, 0.2019, 0.2119, 0.1736, 0.1728, 0.1113, 0.1933, 0.1911], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0212, 0.0216, 0.0199, 0.0245, 0.0191, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:13:25,345 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167585.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:13:25,467 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 12:13:36,787 INFO [finetune.py:976] (0/7) Epoch 30, batch 1500, loss[loss=0.1655, simple_loss=0.2545, pruned_loss=0.03827, over 4833.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2425, pruned_loss=0.04672, over 957072.22 frames. ], batch size: 47, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:13:52,583 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8072, 1.7784, 1.6605, 1.8273, 1.5957, 4.5037, 1.6769, 2.0121], + device='cuda:0'), covar=tensor([0.3056, 0.2401, 0.2050, 0.2157, 0.1464, 0.0093, 0.2310, 0.1181], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0094, 0.0093, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 12:13:57,371 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167633.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:14:07,467 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.1075, 4.8422, 4.5017, 2.9651, 4.9365, 3.8465, 1.2116, 3.5652], + device='cuda:0'), covar=tensor([0.2171, 0.1677, 0.1587, 0.2608, 0.0722, 0.0817, 0.4177, 0.1251], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0182, 0.0162, 0.0132, 0.0165, 0.0125, 0.0150, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 12:14:10,429 INFO [finetune.py:976] (0/7) Epoch 30, batch 1550, loss[loss=0.1603, simple_loss=0.2355, pruned_loss=0.04261, over 4902.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2415, pruned_loss=0.04603, over 958577.61 frames. ], batch size: 43, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:14:18,208 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.4030, 2.2603, 2.3739, 1.6731, 2.3051, 2.4869, 2.3769, 1.9788], + device='cuda:0'), covar=tensor([0.0580, 0.0614, 0.0623, 0.0835, 0.0742, 0.0615, 0.0636, 0.1078], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0118, 0.0129, 0.0139, 0.0139, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:14:26,674 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.931e+01 1.416e+02 1.751e+02 2.105e+02 4.024e+02, threshold=3.503e+02, percent-clipped=1.0 +2023-03-27 12:14:44,009 INFO [finetune.py:976] (0/7) Epoch 30, batch 1600, loss[loss=0.152, simple_loss=0.2158, pruned_loss=0.04414, over 4190.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2406, pruned_loss=0.04615, over 958339.48 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:14:46,418 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.7215, 4.0063, 3.7883, 1.9609, 4.0774, 3.0893, 0.7879, 2.9113], + device='cuda:0'), covar=tensor([0.2454, 0.1633, 0.1689, 0.3497, 0.1025, 0.1043, 0.4737, 0.1539], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0182, 0.0162, 0.0132, 0.0165, 0.0126, 0.0150, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 12:14:54,735 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167719.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:11,337 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 12:15:17,589 INFO [finetune.py:976] (0/7) Epoch 30, batch 1650, loss[loss=0.1738, simple_loss=0.2452, pruned_loss=0.05123, over 4836.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2379, pruned_loss=0.04559, over 957628.54 frames. ], batch size: 30, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:15:19,506 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167756.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:21,896 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167760.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:26,195 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167767.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:32,526 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.722e+01 1.428e+02 1.633e+02 1.926e+02 4.440e+02, threshold=3.266e+02, percent-clipped=1.0 +2023-03-27 12:15:36,047 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167780.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:41,284 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:42,010 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-27 12:15:55,580 INFO [finetune.py:976] (0/7) Epoch 30, batch 1700, loss[loss=0.1684, simple_loss=0.2339, pruned_loss=0.05151, over 4734.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2354, pruned_loss=0.04513, over 956263.23 frames. ], batch size: 59, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:15:56,727 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167804.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:03,659 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167808.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:25,155 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167828.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:42,121 INFO [finetune.py:976] (0/7) Epoch 30, batch 1750, loss[loss=0.1379, simple_loss=0.2008, pruned_loss=0.03755, over 4212.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2377, pruned_loss=0.04573, over 956379.37 frames. ], batch size: 18, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:16:46,633 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-03-27 12:16:56,606 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2159, 2.2195, 1.8377, 2.2769, 2.1401, 2.0879, 2.0629, 3.0384], + device='cuda:0'), covar=tensor([0.3905, 0.4542, 0.3609, 0.4134, 0.4280, 0.2606, 0.4445, 0.1655], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0265, 0.0240, 0.0276, 0.0263, 0.0233, 0.0260, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:16:57,034 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.486e+02 1.829e+02 2.140e+02 4.770e+02, threshold=3.658e+02, percent-clipped=2.0 +2023-03-27 12:17:23,319 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8335, 1.0739, 1.8504, 1.8036, 1.6783, 1.6669, 1.7471, 1.8219], + device='cuda:0'), covar=tensor([0.3936, 0.3852, 0.3157, 0.3569, 0.4800, 0.3721, 0.4606, 0.3170], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0251, 0.0270, 0.0301, 0.0302, 0.0279, 0.0307, 0.0256], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:17:25,521 INFO [finetune.py:976] (0/7) Epoch 30, batch 1800, loss[loss=0.1538, simple_loss=0.2234, pruned_loss=0.04211, over 4233.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2412, pruned_loss=0.04668, over 955627.75 frames. ], batch size: 18, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:17:30,313 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8355, 3.3352, 3.5590, 3.7163, 3.6007, 3.3600, 3.8840, 1.1175], + device='cuda:0'), covar=tensor([0.0908, 0.0925, 0.0931, 0.1029, 0.1313, 0.1786, 0.0824, 0.5889], + device='cuda:0'), in_proj_covar=tensor([0.0357, 0.0249, 0.0290, 0.0301, 0.0340, 0.0290, 0.0309, 0.0306], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:17:58,716 INFO [finetune.py:976] (0/7) Epoch 30, batch 1850, loss[loss=0.1605, simple_loss=0.2373, pruned_loss=0.04182, over 4906.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2435, pruned_loss=0.04812, over 955272.34 frames. ], batch size: 36, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:18:13,648 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.458e+02 1.787e+02 2.144e+02 3.700e+02, threshold=3.573e+02, percent-clipped=1.0 +2023-03-27 12:18:30,634 INFO [checkpoint.py:75] (0/7) Saving checkpoint to pruned_transducer_stateless7_streaming/exp1/checkpoint-168000.pt +2023-03-27 12:18:33,563 INFO [finetune.py:976] (0/7) Epoch 30, batch 1900, loss[loss=0.1758, simple_loss=0.2419, pruned_loss=0.0548, over 4861.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2441, pruned_loss=0.04803, over 956050.04 frames. ], batch size: 31, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:18:53,313 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8579, 1.2859, 1.8972, 1.8967, 1.7090, 1.6515, 1.8232, 1.8302], + device='cuda:0'), covar=tensor([0.4167, 0.3914, 0.3181, 0.3678, 0.4580, 0.3945, 0.4594, 0.2971], + device='cuda:0'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0300, 0.0301, 0.0278, 0.0307, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:18:54,997 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168035.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:03,184 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2341, 2.0915, 1.7079, 2.1450, 2.6771, 2.2282, 2.2328, 1.6427], + device='cuda:0'), covar=tensor([0.2083, 0.1729, 0.1880, 0.1616, 0.1630, 0.1079, 0.1843, 0.1832], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0214, 0.0218, 0.0201, 0.0248, 0.0193, 0.0221, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:19:07,329 INFO [finetune.py:976] (0/7) Epoch 30, batch 1950, loss[loss=0.1579, simple_loss=0.225, pruned_loss=0.04539, over 4779.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2431, pruned_loss=0.04768, over 958041.86 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:19:14,618 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 12:19:21,545 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:22,071 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.458e+02 1.758e+02 2.118e+02 3.555e+02, threshold=3.516e+02, percent-clipped=0.0 +2023-03-27 12:19:30,462 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:35,803 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168096.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 12:19:40,898 INFO [finetune.py:976] (0/7) Epoch 30, batch 2000, loss[loss=0.1424, simple_loss=0.2136, pruned_loss=0.03563, over 4928.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2402, pruned_loss=0.04686, over 957220.63 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:19:54,028 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:54,684 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.7550, 1.7199, 1.4758, 1.6744, 2.0677, 1.9897, 1.7309, 1.5346], + device='cuda:0'), covar=tensor([0.0337, 0.0397, 0.0636, 0.0334, 0.0222, 0.0520, 0.0402, 0.0524], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0111, 0.0102, 0.0118, 0.0104, 0.0115], + device='cuda:0'), out_proj_covar=tensor([7.9598e-05, 8.1364e-05, 1.1556e-04, 8.4598e-05, 7.9178e-05, 8.6461e-05, + 7.7416e-05, 8.7389e-05], device='cuda:0') +2023-03-27 12:19:55,912 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2938, 2.1585, 1.8130, 2.2155, 2.6696, 2.2756, 2.2458, 1.7210], + device='cuda:0'), covar=tensor([0.1942, 0.1889, 0.1938, 0.1572, 0.1718, 0.1057, 0.1880, 0.1856], + device='cuda:0'), in_proj_covar=tensor([0.0251, 0.0215, 0.0220, 0.0202, 0.0250, 0.0195, 0.0222, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:20:01,892 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:20:07,902 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 12:20:14,571 INFO [finetune.py:976] (0/7) Epoch 30, batch 2050, loss[loss=0.1773, simple_loss=0.2424, pruned_loss=0.05611, over 4926.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.237, pruned_loss=0.04616, over 955718.11 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:20:29,509 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.484e+02 1.729e+02 2.115e+02 4.273e+02, threshold=3.459e+02, percent-clipped=3.0 +2023-03-27 12:20:47,506 INFO [finetune.py:976] (0/7) Epoch 30, batch 2100, loss[loss=0.1407, simple_loss=0.2263, pruned_loss=0.02749, over 4795.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2367, pruned_loss=0.04659, over 955294.01 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:21:19,171 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:21:41,936 INFO [finetune.py:976] (0/7) Epoch 30, batch 2150, loss[loss=0.2103, simple_loss=0.2916, pruned_loss=0.0645, over 4898.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2395, pruned_loss=0.04707, over 957067.78 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:22:00,990 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.551e+02 1.892e+02 2.224e+02 4.404e+02, threshold=3.784e+02, percent-clipped=3.0 +2023-03-27 12:22:01,142 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6598, 1.6691, 1.3744, 1.8264, 1.9185, 1.8338, 1.3626, 1.3487], + device='cuda:0'), covar=tensor([0.2401, 0.1967, 0.2053, 0.1576, 0.1772, 0.1255, 0.2524, 0.2170], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0214, 0.0219, 0.0201, 0.0248, 0.0194, 0.0220, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:22:12,556 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168294.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:22:17,117 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8729, 1.9437, 1.5858, 2.0548, 2.3941, 2.0283, 1.8000, 1.4482], + device='cuda:0'), covar=tensor([0.2180, 0.1817, 0.1933, 0.1570, 0.1624, 0.1228, 0.2283, 0.1975], + device='cuda:0'), in_proj_covar=tensor([0.0250, 0.0214, 0.0219, 0.0201, 0.0248, 0.0194, 0.0220, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:22:17,239 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-27 12:22:18,785 INFO [finetune.py:976] (0/7) Epoch 30, batch 2200, loss[loss=0.1797, simple_loss=0.2516, pruned_loss=0.05385, over 4885.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2421, pruned_loss=0.04806, over 956555.86 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:23:01,529 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=3.85 vs. limit=5.0 +2023-03-27 12:23:02,569 INFO [finetune.py:976] (0/7) Epoch 30, batch 2250, loss[loss=0.2014, simple_loss=0.273, pruned_loss=0.06489, over 4821.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2433, pruned_loss=0.04853, over 956560.65 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:23:17,399 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168375.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:17,906 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.029e+01 1.513e+02 1.826e+02 2.132e+02 3.584e+02, threshold=3.652e+02, percent-clipped=0.0 +2023-03-27 12:23:28,075 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168391.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 12:23:32,331 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8279, 1.5836, 2.3259, 3.4808, 2.3620, 2.4024, 1.2766, 3.0502], + device='cuda:0'), covar=tensor([0.1627, 0.1303, 0.1191, 0.0472, 0.0718, 0.1515, 0.1606, 0.0379], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0164, 0.0100, 0.0136, 0.0125, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-27 12:23:36,282 INFO [finetune.py:976] (0/7) Epoch 30, batch 2300, loss[loss=0.1299, simple_loss=0.2078, pruned_loss=0.02597, over 4820.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2429, pruned_loss=0.04782, over 954875.89 frames. ], batch size: 25, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:23:49,955 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:49,980 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:57,184 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([4.5552, 3.9578, 4.1651, 4.4224, 4.3251, 3.9822, 4.6599, 1.4165], + device='cuda:0'), covar=tensor([0.0771, 0.0796, 0.0855, 0.0888, 0.1129, 0.1644, 0.0642, 0.5973], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0248, 0.0289, 0.0299, 0.0339, 0.0288, 0.0308, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:24:09,555 INFO [finetune.py:976] (0/7) Epoch 30, batch 2350, loss[loss=0.1543, simple_loss=0.2264, pruned_loss=0.04106, over 4864.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2412, pruned_loss=0.04741, over 955531.25 frames. ], batch size: 31, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:24:21,870 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:24:24,771 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.258e+01 1.458e+02 1.699e+02 2.116e+02 4.301e+02, threshold=3.398e+02, percent-clipped=1.0 +2023-03-27 12:24:42,043 INFO [finetune.py:976] (0/7) Epoch 30, batch 2400, loss[loss=0.1624, simple_loss=0.2399, pruned_loss=0.04244, over 4860.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2379, pruned_loss=0.04621, over 958004.65 frames. ], batch size: 49, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:24:45,660 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.6870, 1.6205, 1.4644, 1.5407, 1.9666, 1.9921, 1.6506, 1.4168], + device='cuda:0'), covar=tensor([0.0410, 0.0363, 0.0751, 0.0383, 0.0261, 0.0386, 0.0375, 0.0520], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0104, 0.0114], + device='cuda:0'), out_proj_covar=tensor([7.9130e-05, 8.0732e-05, 1.1452e-04, 8.4213e-05, 7.8375e-05, 8.5644e-05, + 7.6898e-05, 8.6716e-05], device='cuda:0') +2023-03-27 12:25:15,072 INFO [finetune.py:976] (0/7) Epoch 30, batch 2450, loss[loss=0.2018, simple_loss=0.2628, pruned_loss=0.07037, over 4845.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2349, pruned_loss=0.04548, over 955400.82 frames. ], batch size: 47, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:25:30,928 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.487e+02 1.838e+02 2.245e+02 3.083e+02, threshold=3.676e+02, percent-clipped=0.0 +2023-03-27 12:25:39,366 INFO [zipformer.py:1188] (0/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:25:48,868 INFO [finetune.py:976] (0/7) Epoch 30, batch 2500, loss[loss=0.1713, simple_loss=0.2701, pruned_loss=0.03619, over 4806.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2376, pruned_loss=0.0467, over 955325.02 frames. ], batch size: 45, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:26:05,033 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 12:26:16,151 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.4744, 1.3376, 1.3574, 1.3889, 1.0535, 2.9237, 1.0917, 1.4027], + device='cuda:0'), covar=tensor([0.3442, 0.2614, 0.2278, 0.2562, 0.1746, 0.0268, 0.2858, 0.1363], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0093, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-27 12:26:27,843 INFO [finetune.py:976] (0/7) Epoch 30, batch 2550, loss[loss=0.1449, simple_loss=0.2301, pruned_loss=0.02988, over 4815.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2409, pruned_loss=0.04756, over 954713.06 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:26:55,336 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.538e+02 1.821e+02 2.163e+02 4.307e+02, threshold=3.641e+02, percent-clipped=2.0 +2023-03-27 12:27:09,674 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:27:17,862 INFO [finetune.py:976] (0/7) Epoch 30, batch 2600, loss[loss=0.1471, simple_loss=0.2252, pruned_loss=0.03447, over 4810.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2425, pruned_loss=0.04768, over 953893.91 frames. ], batch size: 25, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:27:44,426 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168739.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:28:01,793 INFO [finetune.py:976] (0/7) Epoch 30, batch 2650, loss[loss=0.1952, simple_loss=0.2711, pruned_loss=0.05961, over 4819.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2437, pruned_loss=0.04819, over 953881.24 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:28:08,660 INFO [scaling.py:679] (0/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 12:28:12,108 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.5177, 2.3648, 1.9287, 2.7508, 2.5237, 2.1724, 2.9654, 2.5510], + device='cuda:0'), covar=tensor([0.1310, 0.2373, 0.3088, 0.2534, 0.2597, 0.1665, 0.3021, 0.1739], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0193, 0.0240, 0.0255, 0.0252, 0.0211, 0.0218, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:28:21,698 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.510e+02 1.724e+02 1.979e+02 3.263e+02, threshold=3.448e+02, percent-clipped=0.0 +2023-03-27 12:28:34,756 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([3.8214, 3.3798, 3.5497, 3.7255, 3.6101, 3.3719, 3.8863, 1.2326], + device='cuda:0'), covar=tensor([0.0854, 0.0859, 0.0988, 0.1024, 0.1189, 0.1599, 0.0926, 0.5494], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0246, 0.0286, 0.0296, 0.0335, 0.0286, 0.0305, 0.0301], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:28:37,873 INFO [scaling.py:679] (0/7) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-03-27 12:28:43,722 INFO [finetune.py:976] (0/7) Epoch 30, batch 2700, loss[loss=0.1513, simple_loss=0.2297, pruned_loss=0.0364, over 4800.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2427, pruned_loss=0.0475, over 954127.29 frames. ], batch size: 45, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:29:11,014 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.8184, 1.2815, 0.8754, 1.6777, 2.1577, 1.5224, 1.5557, 1.5469], + device='cuda:0'), covar=tensor([0.1400, 0.2004, 0.1879, 0.1140, 0.1832, 0.1873, 0.1387, 0.1934], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-27 12:29:17,011 INFO [finetune.py:976] (0/7) Epoch 30, batch 2750, loss[loss=0.1522, simple_loss=0.2218, pruned_loss=0.04125, over 4871.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2397, pruned_loss=0.04706, over 954377.53 frames. ], batch size: 31, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:29:24,723 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([1.3511, 1.2814, 1.4241, 1.0415, 1.3137, 1.3982, 1.2392, 1.5803], + device='cuda:0'), covar=tensor([0.0987, 0.1781, 0.1221, 0.1337, 0.0695, 0.1110, 0.2743, 0.0693], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0205, 0.0191, 0.0188, 0.0173, 0.0210, 0.0218, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:29:32,259 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.461e+01 1.436e+02 1.670e+02 1.989e+02 2.987e+02, threshold=3.340e+02, percent-clipped=0.0 +2023-03-27 12:29:41,555 INFO [zipformer.py:1188] (0/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168889.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:29:50,503 INFO [finetune.py:976] (0/7) Epoch 30, batch 2800, loss[loss=0.1579, simple_loss=0.2381, pruned_loss=0.03884, over 4937.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2367, pruned_loss=0.0458, over 954027.52 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:29:57,688 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.2288, 1.8154, 2.0962, 2.1925, 1.9329, 1.9661, 2.1726, 1.9861], + device='cuda:0'), covar=tensor([0.5600, 0.5054, 0.4386, 0.5194, 0.6352, 0.5529, 0.6036, 0.4040], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0252, 0.0271, 0.0302, 0.0303, 0.0281, 0.0309, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-27 12:30:13,012 INFO [zipformer.py:1188] (0/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168937.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:30:23,989 INFO [finetune.py:976] (0/7) Epoch 30, batch 2850, loss[loss=0.1093, simple_loss=0.1919, pruned_loss=0.0134, over 4788.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2348, pruned_loss=0.04504, over 955253.43 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:30:33,472 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168967.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:30:34,703 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168969.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 12:30:38,882 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 8.606e+01 1.456e+02 1.728e+02 2.104e+02 5.266e+02, threshold=3.457e+02, percent-clipped=2.0 +2023-03-27 12:30:57,839 INFO [finetune.py:976] (0/7) Epoch 30, batch 2900, loss[loss=0.1582, simple_loss=0.246, pruned_loss=0.03519, over 4837.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2379, pruned_loss=0.04589, over 955131.65 frames. ], batch size: 49, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:31:14,246 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169028.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:31:15,917 INFO [zipformer.py:1188] (0/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169030.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 12:31:24,763 INFO [zipformer.py:2441] (0/7) attn_weights_entropy = tensor([2.8161, 4.6630, 4.4070, 2.6150, 4.7730, 3.6543, 0.9233, 3.4156], + device='cuda:0'), covar=tensor([0.2240, 0.1887, 0.1245, 0.2715, 0.0789, 0.0791, 0.4562, 0.1294], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0179, 0.0159, 0.0129, 0.0162, 0.0123, 0.0148, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-27 12:31:31,770 INFO [finetune.py:976] (0/7) Epoch 30, batch 2950, loss[loss=0.1809, simple_loss=0.2628, pruned_loss=0.0495, over 4828.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2418, pruned_loss=0.04721, over 954791.29 frames. ], batch size: 47, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:31:49,074 INFO [optim.py:369] (0/7) Clipping_scale=2.0, grad-norm quartiles 9.406e+01 1.615e+02 1.887e+02 2.255e+02 4.054e+02, threshold=3.773e+02, percent-clipped=1.0 +2023-03-27 12:31:53,306 INFO [zipformer.py:1188] (0/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169082.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:32:19,263 INFO [finetune.py:976] (0/7) Epoch 30, batch 3000, loss[loss=0.152, simple_loss=0.2347, pruned_loss=0.03462, over 4805.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2427, pruned_loss=0.04756, over 952033.38 frames. ], batch size: 45, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:32:19,265 INFO [finetune.py:1001] (0/7) Computing validation loss